When is the new grading list coming out?

[Ben Purton]

They premium you out

Like if your unrated its like a fortune. Even 2100+ is 200+ Euros. . 2200+ 100-150E etc etc etc

[Paul McKeown]

I fear in future that Organisers will have to put "minimum rating" requirements on their Opens as these problems begin to multiply with the bulk of entries increasingly in the bottom quartile.

Richard,

Deizisau has been doing this for years. Minima of 2000 FIDE/1800 DWZ apply to the A Open. Great tournament btw.

P.

[Jade Hoobe]

Please can someone answer the following query and put me right if I am wrong with my interpretation of the new grading system.

If all juniors this season are being treated as new starters, that means all their results last season were calculated without the 40 point rule being taken into consideration. Therefore, am I correct in thinking that if, last season, an former 80-graded junior player lost to a 140 graded adult player, the junior would still have received 90 points (ie 10 points MORE than their previous grade despite losing the game) ? If so, no wonder many juniors have seemingly inflated grades.

Alternatively, of course, a former 130-graded junior who defeated a 70-graded adult would only have received 120 points (ie 10 points LESS than his previous grade despite winning.)

Another point is that, according to the 'New Grades' section on the ECF website, juniors will be treated as new players EVERY year - therefore what is the point in juniors trying to improve their grade if they are going to be reverted to 'new starter' status every August ?

[Roger de Coverly]

If all juniors this season are being treated as new starters, that means all their results last season were calculated without the 40 point rule being taken into consideration. Therefore, am I correct in thinking that if, last season, an former 80-graded junior player lost to a 140 graded adult player, the junior would still have received 90 points (ie 10 points MORE than their previous grade despite losing the game) ? If so, no wonder many juniors have seemingly inflated grades.

We think so, but don't know so. There are at least two different statements floating about on how the "new starter" routines work. We've seen that whichever routine is used, it's more than capable of producing really high grades (and really low grades) for players scoring more than about 80% or less than 20%.

Another point is that, according to the 'New Grades' section on the ECF website, juniors will be treated as new players EVERY year - therefore what is the point in juniors trying to improve their grade if they are going to be reverted to 'new starter' status every August ?

Next to none - there's no point in their opponents keeping track of their personal performances either.

This really represents an immense lack of communication from the grading team. If they wanted to do something this radical, they could at least have leaked the proposal over the summer to test the waters as it were of its acceptability.

As far as juniors and other unrated players are concerned, they have almost invented a third type of grading system which is neither Clarke nor Elo. The underlying method is that you don't have a clue what your grade is likely to be until the end of the season when they will throw the results at a theoretically suspect computer program and then tell you your rating and ranking.

[Robert Dale]

I am increasingly confused by all this!

Here in Lincolnshire I run 3-4 U18 tournaments each year, 5 round Swiss rapidplay, with some training at lunchtime. We have about 20 juniors competing regularly, and we dutifully pay our game fees to have them graded. For most of these players, these are their only graded games, though they also play in the UK Chess Challenge and other local events. These are the "rank and file" of the junior chess world - they are never likely to become IMs, or even play in the British Championships, at any level, but they enjoy their chess, and I want to do my best to encourage them. There are one or two who are a bit stronger - our top player, Sam Milson, is up there at 177 Rapidplay, and plays in 4NCL and other congresses.

Now here is my question - does it skew the system having these events graded?

If indeed the 40 point rule vanishing for juniors, because they are being treated as new players each year. Doesn't that mean that Sam will lose points, even when he scores 5/5 (as he regularly does!) We value him being there, because it is an inspiration to the others, and hopefully he likes winning the trophies and small cash prizes we offer - but I don't want to destroy his grade!

[Roger de Coverly]

Doesn't that mean that Sam will lose points, even when he scores 5/5 (as he regularly does!)

If he scores that sort of percentage, the new player estimation routine might put him in the top 50. It's believed to have done that to two perhaps three of the "long play" top 50. There are at least two statements of how the routines work, one of them implies that recursion continues until convergence which might be at nearly 50 points greater than the highest graded opponent. One of the long play top 50 got to 219 with 9.5/10 against a field where the highest player was 171.

[Matthew Turner]

Robert,
I believe it will work something like this.
Sam will have his grading performance calculated against all his graded opponent (i.e. adults with an ECF grade). A similar process will take place for all juniors/foreign players. These grades will then be used in the grading calculations. If Sam performs at 170 against his graded adult opposition and Andy Fish performs at 50 against his graded adult opposition.
If Sam beat Andy
Sam gets 180 (you must get ten more than your 'grade')
Andy gets 40 (you must get ten less than your 'grade')

This is educated guesswork, because I think the iterative process can get in the way. For example if Sam grade against adult was based on 9 games his grade would come out as (9*170 + 180)/10 = 171. However, he would now need 181 points for beating Andy, so his grade would come out as (9*170 + 181)/10 = 171.1

In general, I think the conclusion is that good juniors benefit their grade by beating Fish.

[Richard Bates]

Just as a matter of information, has this entire new grading system been produced with the involvement of any proper mathematicians/statisticians, or just a group of enthusiastic amateurs who don't really understand what they are doing? It is incredible that having started with trying to find a solution to a claimed problem of deflation among adult players (which i'm told had some statistical, if not mathematical, justification), we have ended up with a complete revolution in the way juniors are graded, on little more justification (as far as can be ascertained) than a highly dubious assertion that "all juniors are significantly under-rated" (to the extent that 'increments' aren't sufficient and our grading systemis now like the stock market - past performance is no guide to future returns).

Not so long ago a bunch of people in FIDE decided to do something similar with rating calculations (interestingly on similar grounds - fast improving juniors). Fortunately in their case they actually decided to get some expert analysis at the last moment and the whole thing was put off. Whereas the ECF grading system could now be holed below the waterline, with the situation being unrecoverable in the near future. It certainly won't "sort itself out" in one year.

[Robert Dale]

Thanks Roger and Matthew.

So effectively, they calculate the grade in two stages - adults first, then juniors.
I will carry on grading and see what happens.
It only takes one blunder from Sam to catapult some fortunate junior into the grading stratosphere!

[Matthew Turner]

I think it is probably more like juniors first (juniors have to have a 'performance grade' to be able to calculate their adult opponents' grading perfomance on), then adults, then juniors, then juniors again and probably some more iterations as well.

[Robert Dale]

Let's hope ECF have lots of volunteers to sort out all the calculations - or a very powerful and foolproof computer programme. Is there not someone on the grading team who can give us an authoritative answer how it will work - and maybe even give us a miniature example, to prove that it works?

Having a spare 10 minutes, I calculated the performance ratings for our July tournament using the new grades and lots of "iterations" for the new players - the results were actually quite sensible apart from one player graded 46 last year, 101 on the new list, and 21 on his July performance. Now that's what I call volatility!

[Alex Holowczak]

highly dubious assertion that "all juniors are significantly under-rated"

It isn't highly dubious at all. If you ask most adult players in the country, they'll hate playing juniors because they know they're going to be stronger than their grade suggests. If you play 30 games per season, then you'll be fine, and the grade will be about right. Unfortunately, the players who Robert Dale points out who get a handful of graded games per season and little else are undergraded by more than the junior increment suggests. The vast majority of juniors fall into that category. There's a mountain of evidence to show that juniors who don't play 30 games per season are undergraded.

Most of the people who think this is "dubious" are very strong players. The juniors who don't play 30 games per season tend to be graded 50-100. As a result, the strong players never experience this problem in tournament play at all. Last season though, every tournament I played (U100/90 mainly) in seemed to have juniors at the top of the bottom section, even though their grade would suggest they should be nowhere near it. The problem does exist at the bottom of the scale, and if the grading system is going to cater for people with grades down to 0, then it should work properly for them, not dramatically undergrade them.

Take myself as an example. Graded 76, I finished 5th in the Terafinal Challengers in 2007. My results: Win v 48, Win v 116, Draw v 137, Lose v 142, Win v 38, Win v 103. That's a tournament performance of ~ 110. So, if I could have played 30 games that season at the same standard, or even slightly worse, I'd have been 100. Unfortunately, I couldn't, and they were the only games of the season. This means I had to count games from two years previously to get a grade. By comparison, I was rubbish. So my grade ended up at 90, even though my standard of play was 100+. This problem affects most juniors down at the bottom end of the scale. Indeed, since 2007, one of the players I played no longer has a grade. The UKCC Terafinal was probably their only grading tournament.

People have said that "you can't blame the grading system for that". Perhaps not, but you can't blame the players for not being able to play in that many graded tournaments. They're at the mercy of their parents, more often than not. As a result, the grading system should allow for that sort of problem when grading. As a result, I maintain that something had to be done to solve that problem. Regrettably, what has happened with these grades isn't the solution.

[Richard Bates]

highly dubious assertion that "all juniors are significantly under-rated"

It isn't highly dubious at all. If you ask most adult players in the country, they'll hate playing juniors because they know they're going to be stronger than their grade suggests. If you play 30 games per season, then you'll be fine, and the grade will be about right. Unfortunately, the players who Robert Dale points out who get a handful of graded games per season and little else are undergraded by more than the junior increment suggests. The vast majority of juniors fall into that category. There's a mountain of evidence to show that juniors who don't play 30 games per season are undergraded.

Most of the people who think this is "dubious" are very strong players. The juniors who don't play 30 games per season tend to be graded 50-100. As a result, the strong players never experience this problem in tournament play at all. Last season though, every tournament I played (U100/90 mainly) in seemed to have juniors at the top of the bottom section, even though their grade would suggest they should be nowhere near it. The problem does exist at the bottom of the scale, and if the grading system is going to cater for people with grades down to 0, then it should work properly for them, not dramatically undergrade them.

Take myself as an example. Graded 76, I finished 5th in the Terafinal Challengers in 2007. My results: Win v 48, Win v 116, Draw v 137, Lose v 142, Win v 38, Win v 103. That's a tournament performance of ~ 110. So, if I could have played 30 games that season at the same standard, or even slightly worse, I'd have been 100. Unfortunately, I couldn't, and they were the only games of the season. This means I had to count games from two years previously to get a grade. By comparison, I was rubbish. So my grade ended up at 90, even though my standard of play was 100+. This problem affects most juniors down at the bottom end of the scale. Indeed, since 2007, one of the players I played no longer has a grade. The UKCC Terafinal was probably their only grading tournament.

People have said that "you can't blame the grading system for that". Perhaps not, but you can't blame the players for not being able to play in that many graded tournaments. They're at the mercy of their parents, more often than not. As a result, the grading system should allow for that sort of problem when grading. As a result, I maintain that something had to be done to solve that problem. Regrettably, what has happened with these grades isn't the solution.

If you're going to quote me, don't selectively misquote me, please.

Anyway i just don't accept that people who don't play any games are improving at incredible rates. You get better by playing games. Juniors who play the most games will improve the most rapidly. I agree there is a possible issue with some juniors playing a large number of ungraded games, not picked up in the system. However, I think in general we are probably talking about juniors being undergraded on average by a number of points in single figures. More that sufficient to allow them to congregate at the top of graded sections, but not justifying across the board inflation of 20 or 30+ pts.

I accept that my views on this are partly informed by when i was a junior. The changes to the estimation process which Roger has claimed triggered all this I think came in subsequently. However one does not have to look very far among the current "top juniors" to find players who have had the same grade for a number of years (effectively indicating that they had reached a "stalling point") only to suddenly find themselves caterpaulted 40 or 50pts higher effectively as a result of the assumption that "they are a junior so they must be improving rapidly".

[Alex Holowczak]

Anyway i just don't accept that people who don't play any games are improving at incredible rates. You get better by playing games. Juniors who play the most games will improve the most rapidly. I agree there is a possible issue with some juniors playing a large number of ungraded games, not picked up in the system. However, I think in general we are probably talking about juniors being undergraded on average by a number of points in single figures. More that sufficient to allow them to congregate at the top of graded sections, but not justifying across the board inflation of 20 or 30+ pts.

You do get better by playing games, but they don't have to be graded games in tournament chess. They can be played anywhere. Over the Internet, at school at lunchtimes, in matches for your house/school in competitions. They won't be graded. Rarely will players break free from the ranks of lunchtime chess and into congress chess. The only exception is probably the UKCC Terafinal, or one very local tournament per season that maybe a few friends enter. I think big improvements are possible, particularly at that age and standard. An assumed junior increment is not necessarily accurate.

I agree that it doesn't justify across the board inflation, but I think that was more a product of the desire to combat the stretching. I don't think that was justified. All that was required was a tweak to junior grades to ensure that those who don't play many games are more accurately represented on the grading system. Even if this was applied manually, rather than through the analysis of game results. Although, perhaps the ECF could make all games played by juniors free to grade, i.e. no game fee? Then school tournaments and school leagues would be more inclined to have their tournaments graded. This would add games to the pool, and therefore improve the accuracy of junior grades.

[Roger de Coverly]

Grading systems should be in part self repairing. So assume a player with a grade of 100 based on performances over 3 seasons. They can now achieve 150 standard results but only play 5 games a season. Their own grade will lag their strength but it only affects 5 people a year so doesn't do much harm to the aggregate system. If they play 30 games, this "harms" 30 people but causes their personal grade to converge to its correct value within a year. From 100, it works better in the context of the 40 point rule to score 75% against 125 fields than to score 50% against 150 fields.

Juniors have always worked their way through rated restricted competitions, it's a learning experience. Start by hacking up 100 players, then move onto the 125s , repeat until IM.

I might almost prefer to go back to the pre 1990 system. The published grade is the actual performance adjusted for minimum number of games. The input to next year's calculation is the published grade plus something - which is zero for adults but some positive (possibly age-related) number for juniors. The point of the increment being not so much an assumption of improvement by the junior but a hack to give back to the adults, the points that improving players would otherwise remove.