ECF Director wants to scrap Clarke grading system

General discussions about ratings.
Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Roger de Coverly » Tue Aug 25, 2015 11:56 am

Ian Jamieson wrote:I would have thought that having fewer distinct grades and multiple players with the same grade would increase the number of possible pairings.
Given the possibility of arbiter bias it is considered important that one and only one set of pairings is possible. The practical methods of doing this revolve around listing players in ranking order at the start of the tournament and using this ranking order. Whilst this could be done by lottery, the accepted method is to use the grading numbers or Elo ratings. Historically I believe it was the perceived need to show a ranking order for early pairings in events like the British Championships that lead to the previously "secret" grading numbers 200 etc., being disclosed as well as the grade 3a etc.
Last edited by Roger de Coverly on Tue Aug 25, 2015 12:03 pm, edited 1 time in total.

Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Roger de Coverly » Tue Aug 25, 2015 12:02 pm

Ian Jamieson wrote: 6B 153 - 160

It does beg the question what about players under 6B / 153? Or was the standard of chess in England that much higher in the 60s (?). Perhaps one of our older forum members could enlighten us.
Numbers had replaced grades for practical purposes by the end of the 1960s. From what can be established from old magazines, the grades eventually made their way down to 8 or 9. Anyone below that would have been considered "not strong enough to have a grade". With all the grading being done manually, graders were discriminatory over what events they chose to grade. Local league top divisions might be graded, but not the lowest division.

MartinCarpenter
Posts: 3044
Joined: Tue May 24, 2011 10:58 am

Re: ECF Director wants to scrap Clarke grading system

Post by MartinCarpenter » Tue Aug 25, 2015 12:08 pm

Mick Norris wrote:we have got used to the "accuracy" of the exact grades, and some players do think that at, say, 151 they are "better" than someone at 150, and will base board orders around this.
This is maybe actually one of the things that really rapidly updated grades help to put into perspective - my Yorkshire grade has routinely varied by up to +-5 after every game......

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Tue Aug 25, 2015 2:32 pm

@ Roger de Coverly

Thanks for the background information.

E Michael White
Posts: 1420
Joined: Fri Jun 01, 2007 6:31 pm

Re: ECF Director wants to scrap Clarke grading system

Post by E Michael White » Tue Aug 25, 2015 8:13 pm

Ian Jamieson wrote:The old Clarke system would only require 8 digits at the present time because of the way you, or the person you are following, has chosen to extend the system.
Clarke choose the numbering system. Raaphi Persitz is the only player I remember around 1958 to be graded in 1aa. His grade may have been a union grade or provisional grade on <30 games. The more important point is that Clarke initially seemed to have seen 1a as the top grade from year to year although the strength of players in it may increase.

The grade letter system 1a,1b etc was completely abandoned in the 1972 lists. The national lists then went down to 209. Below that there was a combined union list.

In 1968 I recall seeing a player with a union grade of 11b when the chess unions published separate lists.
Last edited by E Michael White on Tue Aug 25, 2015 11:39 pm, edited 1 time in total.

User avatar
Matt Mackenzie
Posts: 5205
Joined: Tue Mar 31, 2009 11:51 pm
Location: Millom, Cumbria

Re: ECF Director wants to scrap Clarke grading system

Post by Matt Mackenzie » Tue Aug 25, 2015 10:12 pm

11b would have been equal to what in numbers?
"Set up your attacks so that when the fire is out, it isn't out!" (H N Pillsbury)

E Michael White
Posts: 1420
Joined: Fri Jun 01, 2007 6:31 pm

Re: ECF Director wants to scrap Clarke grading system

Post by E Michael White » Tue Aug 25, 2015 11:42 pm

11b = 73-80
ie. 8 points per division

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Wed Aug 26, 2015 10:45 am

Roger de Coverly wrote:
Ian Jamieson wrote:I would have thought that having fewer distinct grades and multiple players with the same grade would increase the number of possible pairings.
Given the possibility of arbiter bias it is considered important that one and only one set of pairings is possible. The practical methods of doing this revolve around listing players in ranking order at the start of the tournament and using this ranking order. Whilst this could be done by lottery, the accepted method is to use the grading numbers or Elo ratings. Historically I believe it was the perceived need to show a ranking order for early pairings in events like the British Championships that lead to the previously "secret" grading numbers 200 etc., being disclosed as well as the grade 3a etc.
Having thought about it, personally I would have stuck with the old grade and allocated the ranking order randomly instead of disclosing the 3 digit numbers, particularly for a tournament with the number of rounds that the British Championship has. If there were too many players to identify a clear winner if all the games were decisive (there's not a lot you can do about draws) then that would have been an argument for restricting the number of starting players.

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Wed Aug 26, 2015 10:49 am

E Michael White wrote:
Ian Jamieson wrote:The old Clarke system would only require 8 digits at the present time because of the way you, or the person you are following, has chosen to extend the system.
Clarke choose the numbering system. Raaphi Persitz is the only player I remember around 1958 to be graded in 1aa. His grade may have been a union grade or provisional grade on <30 games. The more important point is that Clarke initially seemed to have seen 1a as the top grade from year to year although the strength of players in it may increase.

The grade letter system 1a,1b etc was completely abandoned in the 1972 lists. The national lists then went down to 209. Below that there was a combined union list.

In 1968 I recall seeing a player with a union grade of 11b when the chess unions published separate lists.
Sorry.

Extending the system by adding a letter may have made sense if you thought it was only going to be for the odd player but I don't think Clarke would have extended the system that way if he had envisaged the improvement in the strength of players over the years (even if a small part of the improvement may be down to inflation)

Thanks also for the background information.

Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Roger de Coverly » Wed Aug 26, 2015 10:52 am

Ian Jamieson wrote: Having thought about it, personally I would have stuck with the old grade and allocated the ranking order randomly instead of disclosing the 3 digit numbers, particularly for a tournament with the number of rounds that the British Championship has.
You would accept a pairing of the number 1 seed with the number 2 seed in the first round then?

User avatar
IM Jack Rudd
Posts: 4818
Joined: Tue Apr 17, 2007 1:13 am
Location: Bideford
Contact:

Re: ECF Director wants to scrap Clarke grading system

Post by IM Jack Rudd » Wed Aug 26, 2015 10:59 am

Roger de Coverly wrote:
Ian Jamieson wrote: Having thought about it, personally I would have stuck with the old grade and allocated the ranking order randomly instead of disclosing the 3 digit numbers, particularly for a tournament with the number of rounds that the British Championship has.
You would accept a pairing of the number 1 seed with the number 2 seed in the first round then?
If more than half the players in the tournament lie within the same eight-point grading band, the seeded Swiss system is going to go slightly haywire anyway.

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Wed Aug 26, 2015 11:07 am

Roger de Coverly wrote:
Ian Jamieson wrote: Having thought about it, personally I would have stuck with the old grade and allocated the ranking order randomly instead of disclosing the 3 digit numbers, particularly for a tournament with the number of rounds that the British Championship has.
You would accept a pairing of the number 1 seed with the number 2 seed in the first round then?
Yes.

Why not?

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Wed Aug 26, 2015 11:23 am

Roger de Coverly wrote:
Ian Jamieson wrote: Having thought about it, personally I would have stuck with the old grade and allocated the ranking order randomly instead of disclosing the 3 digit numbers, particularly for a tournament with the number of rounds that the British Championship has.
You would accept a pairing of the number 1 seed with the number 2 seed in the first round then?
How often is the number 1 seed going to play the number 2 seed in the first round if the ranking order is allocated randomly? Not very often.

I would be prepared to accept the small risk of this happening to avoid what I see as the boring top half vs bottom half pairings we see in a seeded swisses.

Request to moderators - this thread is in danger of being a thread for talking about seeded swisses rather than grading. If one of you has the time could one of you perhaps split the thread perhaps starting with Adam Raoof's post

Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Roger de Coverly » Wed Aug 26, 2015 11:26 am

Ian Jamieson wrote: Why not?

Isn't it obvious? If one of them wins, the loser gets pairings against players who have faired less well, thus easier. If they draw, both get easier pairings than their opposition.

A Swiss system is a hybrid between an all play all and a knockout.

Ian Jamieson
Posts: 203
Joined: Tue Apr 03, 2007 4:00 pm

Re: ECF Director wants to scrap Clarke grading system

Post by Ian Jamieson » Wed Aug 26, 2015 11:32 am

Roger de Coverly wrote:
Ian Jamieson wrote: Why not?

Isn't it obvious? If one of them wins, the loser gets pairings against players who have faired less well, thus easier. If they draw, both get easier pairings than their opposition.

A Swiss system is a hybrid between an all play all and a knockout.
That might be a fair point for your average 5 round weekend swiss but is it really going to make a major difference in the British Championship with 11 rounds?

Post Reply