When is the new grading list coming out?

[Alex Holowczak]

We've all known about this problem for months, when last year's 'new' grades came out, but the ECF grading people, it seems, just don't get it.

I get the impression that they're hiding behind reams of statistical analysis. That's all well and good, but they're neglecting the practical implications of the changes.

[Jonathan Bryant]

When I looked at the top 50 ENG, I noted a relatively unfamiliar name. This player had played 10 games under Morley College in division 3 of the London League. The London League records detail results and "old" grades for all matches so I was able to look up his performance history. Actually I could only find 9 of the 10 games. His score was a rather impressive 8.5/9. Eight of the players had grades, the highest being 157 and I assumed 150 for the ungraded. On the 9 games using "old" grades his performance was 180. When I substituted a minimum grade of 140 for every opponent (40 point rule), the performance was 188. Impressive but not top 50.

This individual flogged a Streatham blogger in the match between Morley College and Streatham. Perhaps Jonathan Bryant might care to paste the game score here? That would provide an independent, albeit insufficient, basis for an estimation of playing strength.

Quite clearly the Morley College captain did not believe that the player in question was superior to players graded 160 last season.

It's true I was soundly beaten by this fellow. I won't post the game here as it will show how badly I played rather than how well he did (plus I'm saving it for a blog article). However,

1) I seem to remember him saying he was coming back after a long break so perhaps he is the same guy as the fellow with the same name who's grade stops some years ago.

2) I remember feeling at the time of the game that was good - better in fact than the standard I usually play. Since I normally face up to 160s and at most low 170s (old money), though, that doesn't say much about whether he's in the top 50 in the country.

3) I don't remember what board I was on but I'm pretty sure it wasn't top board. (Does somebody know where to look this up?) It seems unlikely that a top 50 player would be getting to face me on a middling board of a london league match.

4) The main thing I remember about the game was that the chap was really a thoroughly pleasant guy. Not releveant to the thread I know but worth mentioning.

J

[Roger de Coverly]

3) I don't remember what board I was on but I'm pretty sure it wasn't top board. (Does somebody know where to look this up?)

This is where I found it (google for London League Chess )

http://www.lcl.streamlinenettrial.co.uk/74.htm

So it was board 5.

[Richard James]

3) I don't remember what board I was on but I'm pretty sure it wasn't top board. (Does somebody know where to look this up?)

This is where I found it (google for London League Chess )

http://www.lcl.streamlinenettrial.co.uk/74.htm

So it was board 5.

GC Taylor scored 9½/10 in the London League:

1 v 137
1 v 135
1 v 157
1 v 110
1 v 107
1 v ug (e140)
1 v 150
1 v 128
½ v 135
1 v 124

which I make a grading performance of about 177.

His position in the board order suggested that he was thought to be about 150 strength.

His last published grade was 187 in 1998.

This would suggest that 177 plus a few for the new system would be about right.

Yes, I agree that you cannot grade someone scoring 95% with any degree of accuracy but something in the region of 180-185 seems a lot more sensible than 219.

[Neill Cooper]

It should also be noted that juniors are now treated as ungraded players. .

Howard.
Why? Who agreed that this should be how junior grades are calculated? For the vast majority of juniors (there are about 2500 in the grading list) it is not needed. They only play 10 to 20 graded games a season, and improve slowly.

Treating juniors as ungraded players might be useful for those who play 30+ games a season, with many against adults, but is a disaster for the Briant Poulter League. In each of div 3 and 4 we have 40 to 60 players (that is about 100 juniors of secondary school age), many of whom have existing gradings and, with a few exceptions, are not rapidly improving. 90% of players only play in the league (i.e. against other U18s). So by ignoring previous grades you abolish the main links with the rest of the grading community (because over the seasons different players who do have links have played in their divisions). You are left with few players who do play 30+ games in a season (e.g. Robert Maguire, Jasper Tambini) - who beat everyone they play. The result is that the grades are a complete farce - e.g. Alex Mitchell is graded 182 with RP grade of 92 (his real ability being somewhere in between). My son is the first person EVER to lose 50 grading points in a season, when his ability has not changed. Indeed even his results are not greatly different. This is ludicrous. The grading calculation MUST be changed. Here is my suggestion:

Some juniors do improve rapidly (I teach some of them at Wilsons). They play 30+ games a season. Those are the ones whose grades should be reset. A junior who does not play 30 games in a season should not have their grade reset. Or at least they should have the grades from previous seasons taken into account.

[Brian Valentine]

It should also be noted that juniors are now treated as ungraded players. This dashes any hopes that any adult has of securing a massive increase in grade by spotting an 'obviously overgraded junior' playing a 30 game against them. If both players score at least 3 points, and those are the only 30 games that both players play that season, the adult will see no change in grade whatsoever.

Howard,
Any increase in transparency on these thorny issues is welcomed - thank you.

Nevertheless I'm confused by the above quote. What happens if two juniors play each other? I think, based on my reading of this quote, that they each should get starting grades; how are these estimated if earlier games are disregarded?

[Roger de Coverly]

Some juniors do improve rapidly (I teach some of them at Wilsons). They play 30+ games a season. Those are the ones whose grades should be reset.

I think there should be an upper grading bound on any resetting (and you might apply resets to "inexperienced" adults as well). It appears to be indicated that the players in this year's British Championship will not know their grading performance for the event until this time next year if they played any of the several juniors. In fact none of the performance tables for any events at the British can be published as final if they involved junior players.

[Sebastian Stone]

Some juniors do improve rapidly (I teach some of them at Wilsons). They play 30+ games a season. Those are the ones whose grades should be reset.

I think there should be an upper grading bound on any resetting (and you might apply resets to "inexperienced" adults as well). It appears to be indicated that the players in this year's British Championship will not know their grading performance for the event until this time next year if they played any of the several juniors. In fact none of the performance tables for any events at the British can be published as final if they involved junior players.

Some juniors play more games in a season than I've played in my life.

[David Shepherd]

If Yan-Fan for example plays an adult with say a grade of 100 and makes a blunder and loses does this mean he gets a grading performance of 50 as he is a new player ie no 40 point protection?

[IM Jack Rudd]

If Yan-Fan for example plays an adult with say a grade of 100 and makes a blunder and loses does this mean he gets a grading performance of 50 as he is a new player ie no 40 point protection?

Presumably only in the first iteration. His other games in the season will probably give him a grading performance of 200 or so, in which case for the second iteration, the game will count as 110 because of the 40-point rule.

[Ian Thompson]

It should also be noted that juniors are now treated as ungraded players. This dashes any hopes that any adult has of securing a massive increase in grade by spotting an 'obviously overgraded junior' playing a 30 game against them. If both players score at least 3 points, and those are the only 30 games that both players play that season, the adult will see no change in grade whatsoever.

Howard,
Any increase in transparency on these thorny issues is welcomed - thank you.

Nevertheless I'm confused by the above quote. What happens if two juniors play each other? I think, based on my reading of this quote, that they each should get starting grades; how are these estimated if earlier games are disregarded?

What happens if two juniors play each other? A good question, to which I know the answer under the old grading system, but not under the new one. Under the old system the players' grades are calculated based on their results against graded players and games between two ungraded players are added into the calculations after this has been done. If there are no games against graded players, then the players involved will not get a grade. One consequence of this approach is that if a group of ungraded players played an APA tournament against each other all their grades could be determined by one of them playing a single game against a single graded opponent. Two extreme examples (where the grades have come from the ECF Result File Checker):

1. 10 player APA tournament where Player 1 scores 9/9, Player 2 8/9, Player 3 7/9, ... Player 10 0/9. Player 10 also plays and loses to Michael Adams (grade 272 last year). Player 1 gets a grade of 343, Player 2 322, down to Player 10 192.

2. 10 player APA tournament where Player 1 scores 9/9, Player 2 8/9, Player 3 7/9, ... Player 10 0/9. Player 10 also plays and loses to a player with a grade of 1 last year). Player 1 gets a grade of 172, Player 2 151, down to Player 10 25.

You might, quite reasonably, say that these scenarios are unlikely to happen in practice. What is much more likely is that a group of ungraded children play lots of games against each other in a schools league and just a few of them play a few games outside the league against graded opponents. All their grades will be determined by the small number of results against graded opponents outside the league.

If all juniors are now to be treated as ungraded players, then the chances of these situations arising must be much greater than they were before. I hope the grading team has realised this and allowed for it. I would be interested to know how they have allowed for it,.

[David Shepherd]

Thanks Jack

I fail to see why any juniors should have their grade reset each year - I feel this is not the answer. This seems to be a major change instigated at the last minute with no discussion. I would like to see what the grades were without this revision in the method - the shadow grades previously published looked more realistic to me, although I suspect still had a problem.

Yes we all know the junior grades were wrong and needed changing but a major change to the grading system such as this should have been discussed before being implemented.

Rather than resetting the junior grades I would prefer that the 40 point rule is altered for juniors to allow their opponents grades to be say for example up to 60 points above theirs but limited to 40 below. (when playing adults the grading cap from the adults point of view would still be 40:40). This would introduce inflation but if the levels were set correctly any the random annual junior bonus points could be reduced to reflect this.

I am disturbed also when looking at the current changes in juniors rapidplay grade compared with the changes in long play grade.

[Roger de Coverly]

If Yan-Fan for example plays an adult with say a grade of 100 and makes a blunder and loses does this mean he gets a grading performance of 50 as he is a new player ie no 40 point protection?

Presumably only in the first iteration. His other games in the season will probably give him a grading performance of 200 or so, in which case for the second iteration, the game will count as 110 because of the 40-point rule.

What if he retired from chess in disgust - Is it the average of one game at 50 and 8 games at 200 and something?

[Richard Bates]

Frankly i'm astonished that any of the instigators of these new junior grades can even begin to attempt to defend them. They are a nonsense plain and simple, formed on no reasonable statistical basis and unsurprisingly resulting in absurd outcomes.

If Juniors are generally undergraded due to their capacity for rapid improvement so be it. It is far better than the farce that has now ensued. The fact is that when juniors improve, they tend to improve rapidly. HOWEVER, they also tend to hit a barrier and stall completely.

I have no doubt that as a consequence of this absurd regrading, a significant number of juniors have now achieved their career high. Despite many of them not having the faintest idea how to really play the game.

[Rob Thompson]

I don't think the new junior grades are as bad as everyone is making out. Certainly, for me personally (as an U-17 this coming season) they seem to be pretty accurate, and looking at the players around me on the top 100 juniors they are all players who i think of as being of a pretty similar standard to me, especially those of near my own age group. However, it seems that as the players get younger the grades get more and more inaccurate, which is exactly the same situation that we had before the changes. The only difference seems to be that now young juniors are overrated rather than underrated.

My suggestion to stop these problems would be that there should be a minimum age limit on grades (14 maybe? that's a complete guess, so feel free to correct). This would mean that there were no more underrated or overrated juniors. It is also similar to the idea of resetting juniors' grades every year, except that the all play all situation isn't as likely as the junior leagues (I imagine, where i live (south devon) there are no such things, and junior tournaments are ungraded) would mainly be filled with younger juniors who wouldn't end up with a grade anyway.