How Competitive Do You Want It To Be? - OzChess

**MOZ** · 10-28-2007, 09:16 PM

The link shows the pairings for the forthcoming 2007 World Cup http://www.fide.com/news.asp?id=1491

For the 64 pairings the calculated mean absolute deviation of rating differences is 133.

MOZ

**Slav** · 10-29-2007, 10:26 AM

So??

**MOZ** · 10-29-2007, 12:07 PM

Originally Posted by Slav

So??

So, there is a well-established thread on chesschat that gives context to my thread-starter here.

On that thread I regularly calculated data-points for different events, local and international. While we should not expect to prove a casual relationship between attendance at events and their 'competitive index', we can still speculate and examine co-incidental data; that surely is better than just anecdotal opinion as a methodology.

regards
MOZ

**Iconoclast** · 10-30-2007, 04:42 PM

Originally Posted by MOZ

... I regularly calculated data-points for different events, local and international. While we should not expect to prove a casual relationship between attendance at events and their 'competitive index', we can still speculate and examine co-incidental data; that surely is better than just anecdotal opinion as a methodology.

MOZ,
1. you have well and truely proved your theory that the first few rounds of a giant Swiss is full of strength mismatches.

The next step is to ask the punters
1. Do they want those mismatches and salp taking opportunties?
2. Do they want more tight matches.

Depending on the answers we might find that a solution is (and can!) be found to accomodate both camps.

**MOZ** · 10-30-2007, 11:48 PM

Originally Posted by Iconoclast

MOZ,
1. you have well and truely proved your theory that the first few rounds of a giant Swiss is full of strength mismatches.

The next step is to ask the punters
1. Do they want those mismatches and salp taking opportunties?
2. Do they want more tight matches.

Depending on the answers we might find that a solution is (and can!) be found to accomodate both camps.

For a tournament at the rate of 1 game per week we have invented the intermingling divisionalised SWISS. This is the solution that does accommodate both camps. (Except those eccentrics who want to steer clear of fast improving juniors).

MOZ

**Slav** · 10-31-2007, 01:38 PM

Is a 133 ratings gap a big difference though? It sounds more like a small difference to me!

**MOZ** · 10-31-2007, 02:11 PM

Originally Posted by Slav

Is a 133 ratings gap a big difference though? It sounds more like a small difference to me!

It is an extraordinarily small difference.

Most Australian SWISSES will have 440+ as the difference in the first round.

**Slav** · 10-31-2007, 02:51 PM

Originally Posted by MOZ

It is an extraordinarily small difference.

Most Australian SWISSES will have 440+ as the difference in the first round.

Okay I think I'm with you now!!

**MOZ** · 02-28-2008, 09:36 PM

The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS.
The first round was accelerated.

The competitive index is calculated by taking the average of each absolute difference of paired players ratings.

For rounds 1 and 2 the average differences in rating were 394, 403.

MOZ*

**KiD** · 02-29-2008, 01:54 PM

Originally Posted by MOZ

The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS.

Why? It can't be due to pressure put on by some eccentrics, surely?

**MOZ** · 02-29-2008, 02:20 PM

Originally Posted by KiD

Why? It can't be due to pressure put on by some eccentrics, surely?

Simply due to field size.
You will notice that splitting fields into divisions is common in the very large tournaments such as held in Canberra or Parramatta.

When the eccentrics played the Club had just taken the cap off memberships (capped at 100 for 12 months), and the fields surged to 106 (2000), 117 (2002) etc (memberships peaked at 186). The divisions were a solution to ensuring competitive games; the traditional approach for large Swisses. (What was unique at Box Hill was the intermingling of pairings between divisions. This intermingling would not be appropriate in a week-end Swiss, but is a good solution in a SWISS played at the rate of one game per week).

After the move to a new venue, and with the growth of four new Clubs (Dark Horse, Croydon, Noble Park, and Canterbury Juniors <Tuesday>), our field sizes have settled at 60+ on a Friday.

Simple two round acceleration is now the appropriate method to ensure competitive games.

MOZ*

**MOZ** · 03-06-2008, 09:07 AM

Originally Posted by MOZ

The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS.
The first2 rounds were accelerated. (editted)

The competitive index is calculated by taking the average of each absolute difference of paired players ratings.

For rounds 1 and 2 the average differences in rating were 394, 403.

MOZ*

For rounds 1,2,3 and 4 the average differences in rating were 394, 403, 553, 465.

**Iconoclast** · 03-06-2008, 03:51 PM

That trend is bazar. I conclude that the ratings of a particular subgroup(s) are not a good guide to how they will perform. Although, I cannot say I am totally surprised, after all, Glicko does assume that players will not continue to improve (or slide.) Glicko is an awful attempt at modelling, which is now championed by those who can be fooled in their own credulity in the face of contrary evidence.

**MOZ** · 03-06-2008, 04:15 PM

Originally Posted by Iconoclast

That trend is bazar. I conclude that the ratings of a particular subgroup(s) are not a good guide to how they will perform. Although, I cannot say I am totally surprised, after all, Glicko does assume that players will not continue to improve (or slide.) Glicko is an awful attempt at modelling, which is now championed by those who can be fooled in their own credulity in the face of contrary evidence.

The trend is probably explained by the removal of acceleration after round 2.
In other words, the usual junk round seen in unaccelerated SWISS has been transferred to round 3. Whether it is as bad a junk round at 3, as it usually is at 1 is debatable.

You linkage if GLICKO issues is supererogatory. (Thought I would never get to use that word.

)

MOZ*

**Iconoclast** · 03-06-2008, 05:56 PM

Originally Posted by MOZ

Your linkage if GLICKO issues is supererogatory.

Shouldn't you be saying that like superogativity was a good thing?

10-28-2007, 09:16 PM	#1 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	How Competitive Do You Want It To Be? The link shows the pairings for the forthcoming 2007 World Cup http://www.fide.com/news.asp?id=1491 For the 64 pairings the calculated mean absolute deviation of rating differences is 133. MOZ __________________ FReedom though Fischer-Random chess to enjoy the whole game.

10-29-2007, 10:26 AM	#2 (permalink)
Slav Junior Member Join Date: Oct 2007 Posts: 28	So?? __________________ Touch that Q-pawn at your own risk!

10-29-2007, 12:07 PM	#3 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by Slav So?? So, there is a well-established thread on chesschat that gives context to my thread-starter here. On that thread I regularly calculated data-points for different events, local and international. While we should not expect to prove a casual relationship between attendance at events and their 'competitive index', we can still speculate and examine co-incidental data; that surely is better than just anecdotal opinion as a methodology. regards MOZ __________________ FReedom though Fischer-Random chess to enjoy the whole game.

10-30-2007, 04:42 PM	#4 (permalink)
Iconoclast Immoderator Join Date: Jul 2007 Location: Wollongong NSW Posts: 2,302	Originally Posted by MOZ ... I regularly calculated data-points for different events, local and international. While we should not expect to prove a casual relationship between attendance at events and their 'competitive index', we can still speculate and examine co-incidental data; that surely is better than just anecdotal opinion as a methodology. MOZ, 1. you have well and truely proved your theory that the first few rounds of a giant Swiss is full of strength mismatches. The next step is to ask the punters 1. Do they want those mismatches and salp taking opportunties? 2. Do they want more tight matches. Depending on the answers we might find that a solution is (and can!) be found to accomodate both camps. __________________ The individual is hopeless without the group. The group is hopeless without its individuals.

10-30-2007, 11:48 PM	#5 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by Iconoclast MOZ, 1. you have well and truely proved your theory that the first few rounds of a giant Swiss is full of strength mismatches. The next step is to ask the punters 1. Do they want those mismatches and salp taking opportunties? 2. Do they want more tight matches. Depending on the answers we might find that a solution is (and can!) be found to accomodate both camps. For a tournament at the rate of 1 game per week we have invented the intermingling divisionalised SWISS. This is the solution that does accommodate both camps. (Except those eccentrics who want to steer clear of fast improving juniors). MOZ __________________ FReedom though Fischer-Random chess to enjoy the whole game.

10-31-2007, 01:38 PM	#6 (permalink)
Slav Junior Member Join Date: Oct 2007 Posts: 28	Is a 133 ratings gap a big difference though? It sounds more like a small difference to me! __________________ Touch that Q-pawn at your own risk!

10-31-2007, 02:11 PM	#7 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by Slav Is a 133 ratings gap a big difference though? It sounds more like a small difference to me! It is an extraordinarily small difference. Most Australian SWISSES will have 440+ as the difference in the first round. __________________ FReedom though Fischer-Random chess to enjoy the whole game.

10-31-2007, 02:51 PM	#8 (permalink)
Slav Junior Member Join Date: Oct 2007 Posts: 28	Originally Posted by MOZ It is an extraordinarily small difference. Most Australian SWISSES will have 440+ as the difference in the first round. Okay I think I'm with you now!! __________________ Touch that Q-pawn at your own risk!

02-28-2008, 09:36 PM	#9 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Competitive indicies for the 2008 BHCC Autumn Cup The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS. The first round was accelerated. The competitive index is calculated by taking the average of each absolute difference of paired players ratings. For rounds 1 and 2 the average differences in rating were 394, 403. MOZ* __________________ FReedom though Fischer-Random chess to enjoy the whole game. Last edited by MOZ : 02-28-2008 at 10:05 PM

02-29-2008, 01:54 PM	#10 (permalink)
KiD Member Join Date: Jul 2007 Posts: 84	Originally Posted by MOZ The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS. Why? It can't be due to pressure put on by some eccentrics, surely?

02-29-2008, 02:20 PM	#11 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by KiD Why? It can't be due to pressure put on by some eccentrics, surely? Simply due to field size. You will notice that splitting fields into divisions is common in the very large tournaments such as held in Canberra or Parramatta. When the eccentrics played the Club had just taken the cap off memberships (capped at 100 for 12 months), and the fields surged to 106 (2000), 117 (2002) etc (memberships peaked at 186). The divisions were a solution to ensuring competitive games; the traditional approach for large Swisses. (What was unique at Box Hill was the intermingling of pairings between divisions. This intermingling would not be appropriate in a week-end Swiss, but is a good solution in a SWISS played at the rate of one game per week). After the move to a new venue, and with the growth of four new Clubs (Dark Horse, Croydon, Noble Park, and Canterbury Juniors <Tuesday>), our field sizes have settled at 60+ on a Friday. Simple two round acceleration is now the appropriate method to ensure competitive games. MOZ* __________________ FReedom though Fischer-Random chess to enjoy the whole game.

03-06-2008, 09:07 AM	#12 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by MOZ The usual practice of breaking our tournaments into two divisions was waived this year. Instead we have one all-in SWISS. The first2 rounds were accelerated. (editted) The competitive index is calculated by taking the average of each absolute difference of paired players ratings. For rounds 1 and 2 the average differences in rating were 394, 403. MOZ* For rounds 1,2,3 and 4 the average differences in rating were 394, 403, 553, 465. __________________ FReedom though Fischer-Random chess to enjoy the whole game.

03-06-2008, 03:51 PM	#13 (permalink)
Iconoclast Immoderator Join Date: Jul 2007 Location: Wollongong NSW Posts: 2,302	That trend is bazar. I conclude that the ratings of a particular subgroup(s) are not a good guide to how they will perform. Although, I cannot say I am totally surprised, after all, Glicko does assume that players will not continue to improve (or slide.) Glicko is an awful attempt at modelling, which is now championed by those who can be fooled in their own credulity in the face of contrary evidence. __________________ The individual is hopeless without the group. The group is hopeless without its individuals.

03-06-2008, 04:15 PM	#14 (permalink)
MOZ Volunteer Join Date: Jul 2007 Location: Ozchess is the marginalised cyber-spot where cc-Mods choose to engage with cc-banned posters. Posts: 3,495	Originally Posted by Iconoclast That trend is bazar. I conclude that the ratings of a particular subgroup(s) are not a good guide to how they will perform. Although, I cannot say I am totally surprised, after all, Glicko does assume that players will not continue to improve (or slide.) Glicko is an awful attempt at modelling, which is now championed by those who can be fooled in their own credulity in the face of contrary evidence. The trend is probably explained by the removal of acceleration after round 2. In other words, the usual junk round seen in unaccelerated SWISS has been transferred to round 3. Whether it is as bad a junk round at 3, as it usually is at 1 is debatable. You linkage if GLICKO issues is supererogatory. (Thought I would never get to use that word. ) MOZ* __________________ FReedom though Fischer-Random chess to enjoy the whole game.

03-06-2008, 05:56 PM	#15 (permalink)
Iconoclast Immoderator Join Date: Jul 2007 Location: Wollongong NSW Posts: 2,302	Originally Posted by MOZ Your linkage if GLICKO issues is supererogatory. Shouldn't you be saying that like superogativity was a good thing? __________________ The individual is hopeless without the group. The group is hopeless without its individuals.

Thread Tools
Show Printable Version Email this Page
Rate This Thread
Rate This Thread: