Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Keith,

If I understand you correctly, rankings are used to determine the membership
of "primary seedable groups" (1-2, 3-4, 5-8, 9-16, 17-32) and games won are
used to determine intra-group seniority. An immediate comment is that games
won would need to be replaced by % wins in case one of the blocks becomes
smaller after a withdrawal (illness, injury etc) during the block stage. I
suspect that ties on games or % wins would be common and the decision
whether to break ties by ranking or randomly is significant.

In practice, the current system already uses "secondary seedable groups",
namely 1-8, 9-12, 13-16, 17-24 and 25-32 for the purposes of the KO draw and
what you are proposing is simply a different way of identifying the members
of the secondary seedable groups, using the primary seedable groups and
their internal ordering as an input rather than pure rankings. This means
that the order in the 9-16 and 17-32 primary seedable groups is significant
in that it is better to be in 9-12 than 13-16 and better to be in 17-24 than
in 25-32. One drawback is that the withdrawal of a player from a block
through injury could unfairly (?) benefit someone he has already beaten and
penalise someone he has already lost to.

Another drawback is that your idea would potentially offer the players who
were 3 and 4 the option of dropping games to manipulate the identity of
their semi-final opponent (i.e. suppose that A and B are well clear of the
field on rankings and hence are bound to be 1 or 2 but A has dropped 2 games
in his block and so seems destined to be no. 2; C and D are likewise
destined to be 3 and 4 but C has a poor personal record against A. Might C
not be tempted to lose his final game to fall behind D if by doing so he
could ensure that his potential semi-final opponent was B? ).

I have not found any of the anti-rankings arguments convincing and I don't
think the current system has much wrong with it when applied to AC.

Stephen

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of Keith Aiton
Sent: 25 October 2010 08:43
To: Nottingham Board
Subject: Re: [Croquet] Block Play or Ranking List

----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Keith,

If I understand you correctly, rankings are used to determine the membership
of "primary seedable groups" (1-2, 3-4, 5-8, 9-16, 17-32) and games won are
used to determine intra-group seniority. An immediate comment is that games
won would need to be replaced by % wins in case one of the blocks becomes
smaller after a withdrawal (illness, injury etc) during the block stage. I
suspect that ties on games or % wins would be common and the decision
whether to break ties by ranking or randomly is significant.

In practice, the current system already uses "secondary seedable groups",
namely 1-8, 9-12, 13-16, 17-24 and 25-32 for the purposes of the KO draw and
what you are proposing is simply a different way of identifying the members
of the secondary seedable groups, using the primary seedable groups and
their internal ordering as an input rather than pure rankings. This means
that the order in the 9-16 and 17-32 primary seedable groups is significant
in that it is better to be in 9-12 than 13-16 and better to be in 17-24 than
in 25-32. One drawback is that the withdrawal of a player from a block
through injury could unfairly (?) benefit someone he has already beaten and
penalise someone he has already lost to.

Another drawback is that your idea would potentially offer the players who
were 3 and 4 the option of dropping games to manipulate the identity of
their semi-final opponent (i.e. suppose that A and B are well clear of the
field on rankings and hence are bound to be 1 or 2 but A has dropped 2 games
in his block and so seems destined to be no. 2; C and D are likewise
destined to be 3 and 4 but C has a poor personal record against A. Might C
not be tempted to lose his final game to fall behind D if by doing so he
could ensure that his potential semi-final opponent was B? ).

I have not found any of the anti-rankings arguments convincing and I don't
think the current system has much wrong with it when applied to AC.

Stephen

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of Keith Aiton
Sent: 25 October 2010 08:43
To: Nottingham Board
Subject: Re: [Croquet] Block Play or Ranking List

----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Martyn,
One of the most important aspects is to understand all the available formats and try to find the one that best fits the criteria for your event. It is important to appreciate that very few croquet tournaments are all about finding the best player. There is usually a compromise between given people more games/more flexibility/getting a final/availability of time etc etc.

For example, a standard open weekend shouldn't be designed to maximise the chance that the best player (the player that plays best!) wins. It should appreciate the social dynamic of the event. Some players may want to;
1. play as many as games as possible
2. Finish play by 4pm each day
3. Have a 2 hour lunch break
4. Play people as close to their own ability level as possible
5. Go to watch a football match on Sat afternoon
6. Have a late start on Sunday to recover from the night before
7. etc etc

Blocks fail to achieve the majority of the above.

So, what are better formats. Let's start with the flexible swiss. Here, players are offered games almost as soon as they have finished. Over the duration of the event, they will generally play opponents who are doing as well/badly as they are and thereby are likely to have more interesting, competitive games. If they don't want to play for a couple of hours - no problem. The drawbacks are that there is no guaranteed final and it needs competant management.

However, the flexible swiss is too relaxed a format to use for serious events. A few years ago, Ian Burridge invented the "Burridge swiss" which has been used as the qualifying method for the Open Champs for the past couple (3?) of years. In this, you play a much stricter swiss of around 11 rounds. In order to qualify, you need to win 6 games. Therefore, as soon as 6 rounds have been played, all the players on either 6/6 or 0/6 are removed. This means that there are no dead games. These players are offered Z games if they wish.
After round 7, all the players on 6/7 or 1/7 are removed etc etc
Meaning that in round 11, the only players left are those on 5/10 and we have exciting play-offs for the final KO places.

An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.

Another format for more "social" type events is the Egyptian.

The double round robin is an example of a better block format - good in that it generates a good likelihood that the best player will win, but bad in that some players will lose the possibility to win as the event progresses, but will still have a significant impact on the result. With croquet being a very insular game, it is never good to have to play your friend/wife in a game where one of you wants to win and it doesn't matter to the other.

Quite a lot of 3 day events in the UK use a standard best of 3 knockout format with some form of swiss/Egyptian consolation event. These are also a good combination of two different formats to maximise what everyone wants from the event. I believe that these events have become overseeded and would be interested in seeing one of the regionals use a random draw.

That's just a quick overview - I hope it is helpful.

Regards,

Chris




From:
To:
CC: ; .uk
Date: Sun, 24 Oct 2010 09:59:58 -0700
Subject: Re: [Croquet] Block Play or Ranking List





Chris

You stated “Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used”

I will not ask about the maths involved – I am sure other will

But, would you please name a few better formats and why you like them?

US plays almost nothing but blocks and I would love to find some alternatives.

Regards
Martyn

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Keith,

If I understand you correctly, rankings are used to determine the membership
of "primary seedable groups" (1-2, 3-4, 5-8, 9-16, 17-32) and games won are
used to determine intra-group seniority. An immediate comment is that games
won would need to be replaced by % wins in case one of the blocks becomes
smaller after a withdrawal (illness, injury etc) during the block stage. I
suspect that ties on games or % wins would be common and the decision
whether to break ties by ranking or randomly is significant.

In practice, the current system already uses "secondary seedable groups",
namely 1-8, 9-12, 13-16, 17-24 and 25-32 for the purposes of the KO draw and
what you are proposing is simply a different way of identifying the members
of the secondary seedable groups, using the primary seedable groups and
their internal ordering as an input rather than pure rankings. This means
that the order in the 9-16 and 17-32 primary seedable groups is significant
in that it is better to be in 9-12 than 13-16 and better to be in 17-24 than
in 25-32. One drawback is that the withdrawal of a player from a block
through injury could unfairly (?) benefit someone he has already beaten and
penalise someone he has already lost to.

Another drawback is that your idea would potentially offer the players who
were 3 and 4 the option of dropping games to manipulate the identity of
their semi-final opponent (i.e. suppose that A and B are well clear of the
field on rankings and hence are bound to be 1 or 2 but A has dropped 2 games
in his block and so seems destined to be no. 2; C and D are likewise
destined to be 3 and 4 but C has a poor personal record against A. Might C
not be tempted to lose his final game to fall behind D if by doing so he
could ensure that his potential semi-final opponent was B? ).

I have not found any of the anti-rankings arguments convincing and I don't
think the current system has much wrong with it when applied to AC.

Stephen

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of Keith Aiton
Sent: 25 October 2010 08:43
To: Nottingham Board
Subject: Re: [Croquet] Block Play or Ranking List

----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Martyn,
One of the most important aspects is to understand all the available formats and try to find the one that best fits the criteria for your event. It is important to appreciate that very few croquet tournaments are all about finding the best player. There is usually a compromise between given people more games/more flexibility/getting a final/availability of time etc etc.

For example, a standard open weekend shouldn't be designed to maximise the chance that the best player (the player that plays best!) wins. It should appreciate the social dynamic of the event. Some players may want to;
1. play as many as games as possible
2. Finish play by 4pm each day
3. Have a 2 hour lunch break
4. Play people as close to their own ability level as possible
5. Go to watch a football match on Sat afternoon
6. Have a late start on Sunday to recover from the night before
7. etc etc

Blocks fail to achieve the majority of the above.

So, what are better formats. Let's start with the flexible swiss. Here, players are offered games almost as soon as they have finished. Over the duration of the event, they will generally play opponents who are doing as well/badly as they are and thereby are likely to have more interesting, competitive games. If they don't want to play for a couple of hours - no problem. The drawbacks are that there is no guaranteed final and it needs competant management.

However, the flexible swiss is too relaxed a format to use for serious events. A few years ago, Ian Burridge invented the "Burridge swiss" which has been used as the qualifying method for the Open Champs for the past couple (3?) of years. In this, you play a much stricter swiss of around 11 rounds. In order to qualify, you need to win 6 games. Therefore, as soon as 6 rounds have been played, all the players on either 6/6 or 0/6 are removed. This means that there are no dead games. These players are offered Z games if they wish.
After round 7, all the players on 6/7 or 1/7 are removed etc etc
Meaning that in round 11, the only players left are those on 5/10 and we have exciting play-offs for the final KO places.

An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.

Another format for more "social" type events is the Egyptian.

The double round robin is an example of a better block format - good in that it generates a good likelihood that the best player will win, but bad in that some players will lose the possibility to win as the event progresses, but will still have a significant impact on the result. With croquet being a very insular game, it is never good to have to play your friend/wife in a game where one of you wants to win and it doesn't matter to the other.

Quite a lot of 3 day events in the UK use a standard best of 3 knockout format with some form of swiss/Egyptian consolation event. These are also a good combination of two different formats to maximise what everyone wants from the event. I believe that these events have become overseeded and would be interested in seeing one of the regionals use a random draw.

That's just a quick overview - I hope it is helpful.

Regards,

Chris




From:
To:
CC: ; .uk
Date: Sun, 24 Oct 2010 09:59:58 -0700
Subject: Re: [Croquet] Block Play or Ranking List





Chris

You stated “Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used”

I will not ask about the maths involved – I am sure other will

But, would you please name a few better formats and why you like them?

US plays almost nothing but blocks and I would love to find some alternatives.

Regards
Martyn On Mon, Oct 25, 2010 at 10:38 AM, chris clarke
<> wrote:
>
> An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.


Actually the range of qualifiers in a Burridge Swiss is much smaller
than the potential range from blocks with a 50% qualification criteria
(which we used in the Opens for about 7 years).

In real terms it is within 2 or 3 players of the mean for the number
of rounds (this varies as fewer will qualify with 5 out of 9 than from
5 out of 10) .

Obviously you rarely get a power of 2 players, but, unless you are
oversubscribed, you rarely have a power of 2 blocks either, unless you
want to sacrifice the number of games played. (E.g. with an entry of
46, you can either have 4 blocks of 11/12 or 8 of 5/6 neither of which
have a particularly nice number of games for 3 days play.)

Dave

_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Keith,

If I understand you correctly, rankings are used to determine the membership
of "primary seedable groups" (1-2, 3-4, 5-8, 9-16, 17-32) and games won are
used to determine intra-group seniority. An immediate comment is that games
won would need to be replaced by % wins in case one of the blocks becomes
smaller after a withdrawal (illness, injury etc) during the block stage. I
suspect that ties on games or % wins would be common and the decision
whether to break ties by ranking or randomly is significant.

In practice, the current system already uses "secondary seedable groups",
namely 1-8, 9-12, 13-16, 17-24 and 25-32 for the purposes of the KO draw and
what you are proposing is simply a different way of identifying the members
of the secondary seedable groups, using the primary seedable groups and
their internal ordering as an input rather than pure rankings. This means
that the order in the 9-16 and 17-32 primary seedable groups is significant
in that it is better to be in 9-12 than 13-16 and better to be in 17-24 than
in 25-32. One drawback is that the withdrawal of a player from a block
through injury could unfairly (?) benefit someone he has already beaten and
penalise someone he has already lost to.

Another drawback is that your idea would potentially offer the players who
were 3 and 4 the option of dropping games to manipulate the identity of
their semi-final opponent (i.e. suppose that A and B are well clear of the
field on rankings and hence are bound to be 1 or 2 but A has dropped 2 games
in his block and so seems destined to be no. 2; C and D are likewise
destined to be 3 and 4 but C has a poor personal record against A. Might C
not be tempted to lose his final game to fall behind D if by doing so he
could ensure that his potential semi-final opponent was B? ).

I have not found any of the anti-rankings arguments convincing and I don't
think the current system has much wrong with it when applied to AC.

Stephen

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of Keith Aiton
Sent: 25 October 2010 08:43
To: Nottingham Board
Subject: Re: [Croquet] Block Play or Ranking List

----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Martyn,
One of the most important aspects is to understand all the available formats and try to find the one that best fits the criteria for your event. It is important to appreciate that very few croquet tournaments are all about finding the best player. There is usually a compromise between given people more games/more flexibility/getting a final/availability of time etc etc.

For example, a standard open weekend shouldn't be designed to maximise the chance that the best player (the player that plays best!) wins. It should appreciate the social dynamic of the event. Some players may want to;
1. play as many as games as possible
2. Finish play by 4pm each day
3. Have a 2 hour lunch break
4. Play people as close to their own ability level as possible
5. Go to watch a football match on Sat afternoon
6. Have a late start on Sunday to recover from the night before
7. etc etc

Blocks fail to achieve the majority of the above.

So, what are better formats. Let's start with the flexible swiss. Here, players are offered games almost as soon as they have finished. Over the duration of the event, they will generally play opponents who are doing as well/badly as they are and thereby are likely to have more interesting, competitive games. If they don't want to play for a couple of hours - no problem. The drawbacks are that there is no guaranteed final and it needs competant management.

However, the flexible swiss is too relaxed a format to use for serious events. A few years ago, Ian Burridge invented the "Burridge swiss" which has been used as the qualifying method for the Open Champs for the past couple (3?) of years. In this, you play a much stricter swiss of around 11 rounds. In order to qualify, you need to win 6 games. Therefore, as soon as 6 rounds have been played, all the players on either 6/6 or 0/6 are removed. This means that there are no dead games. These players are offered Z games if they wish.
After round 7, all the players on 6/7 or 1/7 are removed etc etc
Meaning that in round 11, the only players left are those on 5/10 and we have exciting play-offs for the final KO places.

An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.

Another format for more "social" type events is the Egyptian.

The double round robin is an example of a better block format - good in that it generates a good likelihood that the best player will win, but bad in that some players will lose the possibility to win as the event progresses, but will still have a significant impact on the result. With croquet being a very insular game, it is never good to have to play your friend/wife in a game where one of you wants to win and it doesn't matter to the other.

Quite a lot of 3 day events in the UK use a standard best of 3 knockout format with some form of swiss/Egyptian consolation event. These are also a good combination of two different formats to maximise what everyone wants from the event. I believe that these events have become overseeded and would be interested in seeing one of the regionals use a random draw.

That's just a quick overview - I hope it is helpful.

Regards,

Chris




From:
To:
CC: ; .uk
Date: Sun, 24 Oct 2010 09:59:58 -0700
Subject: Re: [Croquet] Block Play or Ranking List





Chris

You stated “Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used”

I will not ask about the maths involved – I am sure other will

But, would you please name a few better formats and why you like them?

US plays almost nothing but blocks and I would love to find some alternatives.

Regards
Martyn On Mon, Oct 25, 2010 at 10:38 AM, chris clarke
<> wrote:
>
> An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.


Actually the range of qualifiers in a Burridge Swiss is much smaller
than the potential range from blocks with a 50% qualification criteria
(which we used in the Opens for about 7 years).

In real terms it is within 2 or 3 players of the mean for the number
of rounds (this varies as fewer will qualify with 5 out of 9 than from
5 out of 10) .

Obviously you rarely get a power of 2 players, but, unless you are
oversubscribed, you rarely have a power of 2 blocks either, unless you
want to sacrifice the number of games played. (E.g. with an entry of
46, you can either have 4 blocks of 11/12 or 8 of 5/6 neither of which
have a particularly nice number of games for 3 days play.)

Dave

_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25 October 2010 10:38, chris clarke <> wrote:

> Ian Burridge invented the "Burridge swiss" ... a much stricter swiss of
> around 11 rounds. In order to qualify, you need to win 6 games. Therefore,
> as soon as 6 rounds have been played, all the players on either 6/6 or 0/6
> are removed. This means that there are no dead games.
>

It seemed that it's not quite a strict swiss - I got the impression that the
first four or so games were pre-allocated on a strict seeded basis. If I'm
right, a weaker player in the event quickly finds himself 4 (out of the 6)
games lost without blinking.

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
block in the top-half.
At the same time 4 qualifiers from another block all played in the bottom
half.

I see this situation as being far from desirable - as was the match up of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 23/10/2010 17:13, wrote:
> Failure to do so in the last WCF AC Championships resulted in a lop-sided
> KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from another
> block in the top-half.
> At the same time 4 qualifiers from another block all played in the bottom
> half.

This is a red herring. Only two matches in the first two rounds were between players from the same
block. That's 2 out 24 matches - not a big percentage. (They were both second round, too.)

And it's not absolutely clear that avoiding block clashes should be the highest priority.
(Personally, I think that it's worth aiming for, but not the most important thing in running a World
Championship.)

> I see this situation as being far from desirable - as was the match up of
> 2 players from the same block in Round 2.

If there is a desire for a tournament where the players play lots of different opponents, then this
possibly could be accommodated, but I don't see it as the real reason for holding a World Championship.

Dave




_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Hi Kevin,

Thanks.

In my scenario, the blocks are utterly pointless.

My scenario simply serves to illustrate a principle - test a theory at
the extremes and see what happens. I posed two questions to John Riches
- while you haven't answered either, I'd be most interested in how you
would answer the second one. What would you do if you were the manager?

IMHO the results from the blocks should serve absolutely no purpose
other than determining who qualifies and provide for a minimum number of
games for every competitor. This is particularly important/useful for
international events, where players can be out-of-season and have
travelled considerable distances.

Using your system, it is possible that in the first round of the KO the
Top ranked player in the world could play the second ranked (in a best
of 3, possibly on the Hurlingham cricket pitch?) and the 3rd ranked
player could play the 4th (again, a best of three). In a World
Championship, this is not just far from desirable, it is unacceptable.

I think that formats of the recent World Championships have been
excellent - a very decent compromise that still recognises the results
of the blocks, and IMHO see no reason to change it.

Regards,
Reg

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of

Sent: 23 October 2010 17:13
To: Reg Bamford; johnriches; Notts List
Subject: [Croquet] Betr: Block Play or Ranking List

Reg,

I appreciate your input but have the following question:

In your scenario, what is the point of the blocks?
Do the blocks provide any useful purpose?

In the current format of the WCF World Championships, 64 (or 80) players
take part in the blocks, with only 32 progressing to the KO stages.

Therefore, IMHO, the blocks provide a useful function, and the results
from
those blocks should be paramount in deciding the seeding in the KO.

Failure to do so in the last WCF AC Championships resulted in a
lop-sided
KO, with all 4 qualifiers from one block (pink?) and 3 qualifiers from
another
block in the top-half.
At the same time 4 qualifiers from another block all played in the
bottom
half.

I see this situation as being far from desirable - as was the match up
of
2 players from the same block in Round 2.

But that's just my opinion!

Kevin
>-- Oorspronkelijk bericht --
>Date: Sat, 23 Oct 2010 13:39:47 +0100
>From: "Reg Bamford" <>
>To: "johnriches" <>,
> "Notts List" <.uk>
>Subject: [Croquet] Block Play or Ranking List
>
>
>I make a contribution to this thread at my peril and probably against
my
>own better judgement. Here are two questions for John Riches...
>
>I use an effective test in my own business when someone comes up with a
>theory/idea/initiative - simply test it under extreme conditions. So
>here's the scenario I pose:
>
>Which group of 16 players (the 16 block winners or top 16 ranked
player)
>is more likely to win a competition under the following conditions:
>
>- it's the World Championships
>- 16 blocks of 2 players (32 players in total)
>- 1 game per block
>- All 32 players proceed to the KO
>- KO seeding done according to results in the Blocks (wins, then
points)
>- Games played on front lawns at Hurlingham (i.e. flat, green, slow
with
>easy hoops)
>- Each round of the KO is best of 17 (it's quite a long event this
time)
>
>Here are my questions, John. 1) Who is more likely to win this event -
>one of the 16 block winners or one of the top 16 ranked players? And 2)
>how would you choose to seed the KO if you were manager of this
>tournament if you were allowed a choice?
>
>If I tried answering my own questions, I would write as follows: a) I
>reckon that the probability of the winner of the event being a block
>winner is around 65% (Chris, does that sound about right?). But I would
>bet my house and the health of my kids that the winner of the event
will
>come from the top 4 in the World rankings (Chris, what probability is
>that - much in it if I said 99.5%?). And b) I would prefer (and I bet
so
>would most people, nationalities and friendships aside) the Final
>contested by the 2 best players, so I would use the Rankings to
>determine seedings in the KO.
>
>Cheers, Reg
>
>-----Original Message-----
>From: croquet-.uk
>[mailto:croquet-.uk] On Behalf Of
johnriches
>Sent: 23 October 2010 00:04
>To: Notts List
>Subject: Re: [Croquet] draw/process
>
>...
>
>as I have said above, even if the original premise is accepted, that
>would not in itself justify using rankings to seed the draw for the
>finals, rather than using the block results alone.
>
>....
>
>
>_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Reg,
You are banging your head against a brick wall. Riches would always use the block results. Kevin would probably do so as well.

JR wrote;
"Seeding is desirable before the event starts, but ONLY to ensure that each player gets a roughly equal set of opponents; NOT to increase the chance of a highly-ranked player winning the event regardless of how well he plays."

What people never remember is the position of the number 1 seed. All a number 1 seed cares about is avoiding the (let's say dozen) or so players he can lose to for as long as possible. If the event is seeded normally, he avoids these players until the quarter finals.
If we seed by block results, the number 1 seed can win all his block games, and then have to play anyone from then on. They might end up playing a Mark McInerney. And yet, the runner-up in his block might get an easier draw. It's all random - down to the clearly random results from a handful of games.
These people claim they want a "fair" draw and yet they want to use the minimum amount of data possible to create a fair draw. By definition, this will make it less fair.

If people can't see that using 1 game for seeding purposes isn't fair, it is impossible to progress the debate.

My answers to your questions would be c. 70% and c. 98.5%

Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used.

Let's be quite clear - the reason blocks were introduced into the worlds is to give players a warm-up and ensure the weaker players got more games. Now that we've sacrificed 4 days to achieve this, we should be allowed the optimal format for the KO of our most prestigious event.

Chris It seems like we are getting nowhere here!

Throughout my life, I have played in, been an administrator of, a fan of,
and spectator of numerous sports. I am a self-confessed sports-freak!

No other sport, to the best of my knowledge, in events where blocks/groups
precede a knock-out, would even consider re-seeding the KO and ignoring results
in the groups.

Try a little test. When you go to work, the bar, or just socialise, ask players
or fans of other sports what they think of the idea. I can almost guarantee
their reaction!
Better still, try explaining to a sports journalist. If we want croquet to
be treated better by the media, we need to make it seem much fairer, and
less elitist.

A knock-out competition, by it's very nature, is not "fair". It is almost
impossible for both finalists to have played opposition of equal strength.
Live with it!

If a top seed has to play somebody who also has a high ranking/index, but
who has played poorly so far in the competition, so be it.
Would you rather play somebody who is on form, or the player struggling?

As I have said before, the World Champion is not necessarily the best player,
or even the top-ranked player.
He is simply the player who has overcome (i.e. beaten) all the opponents
put before him. Regardless of whether they are weak players or strong players.
He has done all that is required of him.

I still fail to see any justification whatsoever for using historical results
to seed the KO, and effectively ignore the results in the blocks.

Kevin





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. ----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25/10/2010, at 6:12 PM, Keith Aiton wrote:

> Which is one reason why I posted the following a couple of days ago...................
>
> "I haven't been following this topic in minute detail, but how does this
> stack up as an idea:
>
> Once the 32 qualifiers have been found you put them by ranking in the
> following groups:
>
> 1-2
> 3-4
> 5-8
> 9-16
> 17-32
>
> Then, the actual places in the draw (which are determined on a strict 1v32,
> 2v31 and so on basis) are determined by the numbers of games won in the
> block stage, so, for example if the #1 ranked player won 8 games but the #2
> ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
> so on within each group.
>
> Ties to be separated either by random draw or by ranking."
>
> Keith

Keith,

Your idea is an interesting one. It is far better than full seeding and ignoring the block results completely.
However, it still takes account of past history in a way that can unfairly (in my opinion) disadvantage a player who has played well in the blocks and advantage a player who has not.

To take a rather extreme example (but not the most extreme), suppose the 4th seed (according to rankings) plays poorly and comes 4th in his block, while the 30th seed plays well and wins his block, possibly having beaten the 4th seed if they were in the same block..

Then you would have, in the first KO round, the 4th seed playing the 4th-from-bottom qualifier who must have been seeded lower than 16th - probably 29th - and the 30th seed, who won his block, probably playing the 16th seed.
The 30th seed is given a harder KO draw than the 4th seed for no other reason than what the 4th seed did months ago to achieve his high ranking!

The 4th seed gets the same 4th-weakest KO opponent regardless of how well or poorly he plays in the blocks (provided he qualifies), while the 30th seed gets a player from the top half of the seedings regardless of how well he plays in the blocks.

To me that does not seem fair, and it makes the playing of the blocks almost a waste of time.
Once the 4th seed has won sufficient games to qualify, why should he try to win any more games? He will get the same KO position and opponent regardless. And if he does not try to win his remaining games, that can unfairly influence who else in his block qualifies and who they will play in the finals.

JR.


NOTE: Views expressed in this email are my own personal opinions and do not necessarily represent those of the ACA.





_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Keith,

If I understand you correctly, rankings are used to determine the membership
of "primary seedable groups" (1-2, 3-4, 5-8, 9-16, 17-32) and games won are
used to determine intra-group seniority. An immediate comment is that games
won would need to be replaced by % wins in case one of the blocks becomes
smaller after a withdrawal (illness, injury etc) during the block stage. I
suspect that ties on games or % wins would be common and the decision
whether to break ties by ranking or randomly is significant.

In practice, the current system already uses "secondary seedable groups",
namely 1-8, 9-12, 13-16, 17-24 and 25-32 for the purposes of the KO draw and
what you are proposing is simply a different way of identifying the members
of the secondary seedable groups, using the primary seedable groups and
their internal ordering as an input rather than pure rankings. This means
that the order in the 9-16 and 17-32 primary seedable groups is significant
in that it is better to be in 9-12 than 13-16 and better to be in 17-24 than
in 25-32. One drawback is that the withdrawal of a player from a block
through injury could unfairly (?) benefit someone he has already beaten and
penalise someone he has already lost to.

Another drawback is that your idea would potentially offer the players who
were 3 and 4 the option of dropping games to manipulate the identity of
their semi-final opponent (i.e. suppose that A and B are well clear of the
field on rankings and hence are bound to be 1 or 2 but A has dropped 2 games
in his block and so seems destined to be no. 2; C and D are likewise
destined to be 3 and 4 but C has a poor personal record against A. Might C
not be tempted to lose his final game to fall behind D if by doing so he
could ensure that his potential semi-final opponent was B? ).

I have not found any of the anti-rankings arguments convincing and I don't
think the current system has much wrong with it when applied to AC.

Stephen

-----Original Message-----
From: croquet-.uk
[mailto:croquet-.uk] On Behalf Of Keith Aiton
Sent: 25 October 2010 08:43
To: Nottingham Board
Subject: Re: [Croquet] Block Play or Ranking List

----- Original Message -----
From: <>
To: "Nottingham Board" <.uk>
Sent: Monday, October 25, 2010 7:56 AM
Subject: [Croquet] Block Play or Ranking List


> It seems like we are getting nowhere here!
.
.
.
.
> Kevin

Which is one reason why I posted the following a couple of days
ago...................

"I haven't been following this topic in minute detail, but how does this
stack up as an idea:

Once the 32 qualifiers have been found you put them by ranking in the
following groups:

1-2
3-4
5-8
9-16
17-32

Then, the actual places in the draw (which are determined on a strict 1v32,
2v31 and so on basis) are determined by the numbers of games won in the
block stage, so, for example if the #1 ranked player won 8 games but the #2
ranked player won 9 games then the #2 ranked player becomes the #1 seed, and
so on within each group.

Ties to be separated either by random draw or by ranking."

Keith


_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. Martyn,
One of the most important aspects is to understand all the available formats and try to find the one that best fits the criteria for your event. It is important to appreciate that very few croquet tournaments are all about finding the best player. There is usually a compromise between given people more games/more flexibility/getting a final/availability of time etc etc.

For example, a standard open weekend shouldn't be designed to maximise the chance that the best player (the player that plays best!) wins. It should appreciate the social dynamic of the event. Some players may want to;
1. play as many as games as possible
2. Finish play by 4pm each day
3. Have a 2 hour lunch break
4. Play people as close to their own ability level as possible
5. Go to watch a football match on Sat afternoon
6. Have a late start on Sunday to recover from the night before
7. etc etc

Blocks fail to achieve the majority of the above.

So, what are better formats. Let's start with the flexible swiss. Here, players are offered games almost as soon as they have finished. Over the duration of the event, they will generally play opponents who are doing as well/badly as they are and thereby are likely to have more interesting, competitive games. If they don't want to play for a couple of hours - no problem. The drawbacks are that there is no guaranteed final and it needs competant management.

However, the flexible swiss is too relaxed a format to use for serious events. A few years ago, Ian Burridge invented the "Burridge swiss" which has been used as the qualifying method for the Open Champs for the past couple (3?) of years. In this, you play a much stricter swiss of around 11 rounds. In order to qualify, you need to win 6 games. Therefore, as soon as 6 rounds have been played, all the players on either 6/6 or 0/6 are removed. This means that there are no dead games. These players are offered Z games if they wish.
After round 7, all the players on 6/7 or 1/7 are removed etc etc
Meaning that in round 11, the only players left are those on 5/10 and we have exciting play-offs for the final KO places.

An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.

Another format for more "social" type events is the Egyptian.

The double round robin is an example of a better block format - good in that it generates a good likelihood that the best player will win, but bad in that some players will lose the possibility to win as the event progresses, but will still have a significant impact on the result. With croquet being a very insular game, it is never good to have to play your friend/wife in a game where one of you wants to win and it doesn't matter to the other.

Quite a lot of 3 day events in the UK use a standard best of 3 knockout format with some form of swiss/Egyptian consolation event. These are also a good combination of two different formats to maximise what everyone wants from the event. I believe that these events have become overseeded and would be interested in seeing one of the regionals use a random draw.

That's just a quick overview - I hope it is helpful.

Regards,

Chris




From:
To:
CC: ; .uk
Date: Sun, 24 Oct 2010 09:59:58 -0700
Subject: Re: [Croquet] Block Play or Ranking List





Chris

You stated “Blocks were the bane of my croquet career - a waste of time for me (not others, I do understand that). In NZ and Aus, this is the standard format. I hate it. Part of the problem is that most players don't understand the better formats and certainly don't understand the maths involved in what type of seeding should be used”

I will not ask about the maths involved – I am sure other will

But, would you please name a few better formats and why you like them?

US plays almost nothing but blocks and I would love to find some alternatives.

Regards
Martyn On Mon, Oct 25, 2010 at 10:38 AM, chris clarke
<> wrote:
>
> An ingenious format that offers plenty of games and removes the need to play meaningless block games. The negative is a high likelihood of an uneven number of qualifiers for the K.O.


Actually the range of qualifiers in a Burridge Swiss is much smaller
than the potential range from blocks with a 50% qualification criteria
(which we used in the Opens for about 7 years).

In real terms it is within 2 or 3 players of the mean for the number
of rounds (this varies as fewer will qualify with 5 out of 9 than from
5 out of 10) .

Obviously you rarely get a power of 2 players, but, unless you are
oversubscribed, you rarely have a power of 2 blocks either, unless you
want to sacrifice the number of games played. (E.g. with an entry of
46, you can either have 4 blocks of 11/12 or 8 of 5/6 neither of which
have a particularly nice number of games for 3 days play.)

Dave

_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe. On 25 October 2010 10:38, chris clarke <> wrote:

> Ian Burridge invented the "Burridge swiss" ... a much stricter swiss of
> around 11 rounds. In order to qualify, you need to win 6 games. Therefore,
> as soon as 6 rounds have been played, all the players on either 6/6 or 0/6
> are removed. This means that there are no dead games.
>

It seemed that it's not quite a strict swiss - I got the impression that the
first four or so games were pre-allocated on a strict seeded basis. If I'm
right, a weaker player in the event quickly finds himself 4 (out of the 6)
games lost without blinking. On Mon, Oct 25, 2010 at 1:43 PM, Dave Kibble <> wrote:
> It seemed that it's not quite a strict swiss - I got the impression that the
> first four or so games were pre-allocated on a strict seeded basis. If I'm
> right, a weaker player in the event quickly finds himself 4 (out of the 6)
> games lost without blinking.

Not really.

Yes, I did it fully seeded for 3 rounds (I don't think Richard did),
but I don't think that it's as bad as you make out.

If we have say 48 players, then there would be 6 "mini-swisses" of 8
players each. In the group with the top player, there would also be
players 12,13,24,25,36,37 & 48 (assuming that the seeding is striped).
They would play each other in the order 1v48, 24v25, 13v36 & 12v37.
Assuming that everything goes to seeding, then the next two rounds
will be 1v24, 12v13, 25v48 & 36v37 and 1v12, 24v13, 25v36 & 48v37. So
by the third round the 48th player ought to have someone with whom he
is reasonably competitive, and in the fourth round he will be in with
all the other people on zero wins (who by seeding ought to be the
players in the group 43-48), and may play any of them, as it's
difficult to pre-allocate that far ahead. He will not lose 4 games
"without blinking" unless he is particularly bad (and in that case he
is *always* be likely to be the one who gets the first NQ).

In an unseeded swiss, he may well get two top players in succession
(thus being on 0 out of 2 "without blinking"), and can then get a
reasonable player (say in the 10-20 range) who has himself had a bad
draw. This is much more likely to end up with a player in the fourth
round not having taken croquet in his previous 3 games. On the other
side of that coin, if the "weaker player" gets a couple of wins early
on, all that happens is that he gets then gets beaten by several good
players in a row, putting him in a worse situation than if he were
down with the other zero loss players straight away. Ultimately the
more losses you have, the weaker your opponents will be (this is the
theory behind the "swiss gambit").

Dave

_______________________________________________
___________________________________________________

Posted on the Croquet mailing list. Go to http://nottingham-lists.org.uk/mailman/listinfo/croquet to subscribe.