Database Upgrade

Request amendments or pass comments on this free service.
Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 10:25 am

Whatever the pros and cons of the deflation issue, there is surely no doubt that the junior grades (for the younger age groups) cannot be published in their current form without inviting total ridicule. It is just not remotely credible that so many children are anywhere near as strong as the revised grades would have them. And this is only the ones who have grades!

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Database Upgrade

Post by Matthew Turner » Sat Mar 07, 2009 12:44 pm

Richard,
There are certainly some (young) juniors who will be overgraded after this revision. However, I think you have to be slightly careful, because it is easy to fall into the trap of saying This junior who was 110 last year is never 150 strength. However, what is 150 strength? After the new grades are introduced perhaps 150 is what you think of as a 130. When you put it into this context I think the junior grades are much more reasonable.

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Database Upgrade

Post by Roger de Coverly » Sat Mar 07, 2009 1:01 pm

However, what is 150 strength?
Someone who can score 25% against 175 ish opposition perhaps. At the 175 level, the new grades are within shouting distance of the old ones, typical changes of 5-10 points.
After the new grades are introduced perhaps 150 is what you think of as a 130
But it's a 130 who is expected to score 25% against players in the 170's. Open tournaments are harsh and unforgiving arenas for inexperienced players so it's only the most improving 130's that try their luck.

In the context of the juniors we're not talking about 130 -> 150. It's 80 -> 150 in some cases.

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Database Upgrade

Post by Matthew Turner » Sat Mar 07, 2009 3:50 pm

As far as I can see the juniors who have gone from 80 to 150 are very young, played a lot and are improving rapidly. If nothing had happened they could easily have risen from 80 to say 115 or even 120. Adults of 120 have typically risen to 140, so we are talking about a (very) few juniors being overgraded by 10, 15 or perhaps 20 points, which although this may or may not be a problem, it is not as severe as might at first appear.

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 3:53 pm

Matthew Turner wrote:Richard,
There are certainly some (young) juniors who will be overgraded after this revision. However, I think you have to be slightly careful, because it is easy to fall into the trap of saying This junior who was 110 last year is never 150 strength. However, what is 150 strength? After the new grades are introduced perhaps 150 is what you think of as a 130. When you put it into this context I think the junior grades are much more reasonable.
I accept that Matthew, but i am approaching this from a perspective of the claims that these grading revisions will correct historic anomalies*. So i am approaching it from the perspective that these grades are suggesting that there are dozens of children who are significantly stronger (in relative terms) than, say, i was at the same age (and this only effects children with published grades - it would be reasonable to deduce that there are many others with no published grade who are in a similar position). Now, who knows, it may well be that we are entering a new golden age for UK junior chess (an army of Magnus Carlsens, as somebody put it), but it all seems a bit incongruous with the Junior age group taken as a whole (U18) looks as if it is a bit weaker in grading terms than in my youth, certainly at the top end (David Howell excluded).

*I admit that it may well be that i have misunderstood the junior situation, and that it is argued that historical junior grades were significantly understated as well, although i find that difficult to square with my experience of taking 5 years from entering the system to get to 200 strength, whilst always being in the top few of my age group.

BTW who is this 8 year old Scot who is apparently 200 strength?

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 4:01 pm

Having said that, it is possible (my memory isn't that good) that i was only playing almost exclusively in an underrated junior pool for the first three of those years (at which point i was fortunate enough to have a grade of 159 enabling me to clean up in a couple of large prized major tournaments, before entering Opens from then on), so perhaps the advance to 200 was artificially prolonged- I got a FIDE rating slightly quicker from memory.

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 4:05 pm

Matthew Turner wrote:As far as I can see the juniors who have gone from 80 to 150 are very young, played a lot and are improving rapidly. If nothing had happened they could easily have risen from 80 to say 115 or even 120. Adults of 120 have typically risen to 140, so we are talking about a (very) few juniors being overgraded by 10, 15 or perhaps 20 points, which although this may or may not be a problem, it is not as severe as might at first appear.
I'm not sure it's correct to say they've gone from 80 to 150 (or 115-120) in one year. They've more likely gone from 60, say, last year to 150 this year. We haven't yet seen what they will advance to next year!

I'm more intrigued by the maths that sees people going from grades in single figures to the 70s or 80s! ;)

Matthew Turner
Posts: 3600
Joined: Fri May 16, 2008 11:54 am

Re: Database Upgrade

Post by Matthew Turner » Sat Mar 07, 2009 5:13 pm

Richard,
I am intrigued by this 200 strength, 8 year-old Scot, I would be fascinated if someone can provide a name.
How can you mathematically justify a player going from single figures to 70 or 80. Very simple - you can't, but I believe that you cannot mathematically justify a lot of the changes taking place and you shouldn't try. To get a grade of 1 you must have played for a year to get a grade. After two year's experience of graded competition you should have improved, you should certainly be a lot better than a new starter. Why not then be given a grade of 70 0r 80. It is recognition that you have continued to play. You can still play in Minors and your parents wil be delighted at the big rise in your grade. Everyone is happy, the system works and I didn't need my calculator once!

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Database Upgrade

Post by Roger de Coverly » Sat Mar 07, 2009 5:24 pm

I'm not sure it's correct to say they've gone from 80 to 150 (or 115-120) in one year.
They've gone from 80 to 150 in the same year. All the values quoted are reworks of the July 2008 actuals. So the 150 represents last year's reworked performance plus the age based enhancement. I think the enhancement at young ages used to be 10 and is now 15. So we are talking about moving from a performance of 70 plus age enhancement of 10 to a performance of 135 plus age enhancement of 15.

If the "new" performance of 135 is correct then it can be asserted that the "old" performance of 70 was really, really wrong. So what was the "old" system doing so badly wrong that the "new" approach has corrected? As far as I am aware, the "new" 2008 values are more or less calculated "normally" - by using the "new" 2007 values and adding in new players with the same estimation process as for the current system.

As far as I am aware, the estimation process runs in a loop. So first calculate an estimate based on games against players already in the system. Then put ungraded players in with initial estimates of the first pass. Repeat until convergence. Somewhere in the loop, the 40 point rule and the junior enhancements are presumably included. What strikes me is that how we know there is only one solution? There's a field of mathematics called "numeric methods" which studies these interative process things and one of the issues to test for is whether there is a unique solution.

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Database Upgrade

Post by Roger de Coverly » Sat Mar 07, 2009 5:56 pm

I am intrigued by this 200 strength, 8 year-old Scot, I would be fascinated if someone can provide a name.
Born in 1949 according to FIDE

http://ratings.fide.com/card.phtml?event=2401622

Also the secretary of the Inverness chess club. (Chess Scotland website)

Plays at Gibraltar in the U1800 section

http://www.gibraltarchesscongress.com/g ... teurs.html


So you can take a 140 ish middle-aged Scotsman, input an erroneous date of birth and watch his grade rocket to nearly 200. Do we really trust the "new" grades?

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 6:51 pm

Roger de Coverly wrote:
I am intrigued by this 200 strength, 8 year-old Scot, I would be fascinated if someone can provide a name.
Born in 1949 according to FIDE

http://ratings.fide.com/card.phtml?event=2401622

Also the secretary of the Inverness chess club. (Chess Scotland website)

Plays at Gibraltar in the U1800 section

http://www.gibraltarchesscongress.com/g ... teurs.html


So you can take a 140 ish middle-aged Scotsman, input an erroneous date of birth and watch his grade rocket to nearly 200. Do we really trust the "new" grades?
What a shame! I was on the verge of going down the bookies and putting some money on a Scottish World Champion within ten years! I reckon i'd have got a good price :(

I guess from this we can assume that there will be a small spike in the gradings of players in the Inverness area? :roll:

Roger de Coverly
Posts: 21291
Joined: Tue Apr 15, 2008 2:51 pm

Re: Database Upgrade

Post by Roger de Coverly » Sat Mar 07, 2009 7:21 pm

Plays at Gibraltar in the U1800 section

http://www.gibraltarchesscongress.com/g ... teurs.html
Actually doesn't this case demonstrate that the "new" junior grades are just plain wrong and may be double ,triple or even quadruple counting the junior increment? It looks to me as if an incorrect date of birth was entered into the ECF grading database, perhaps 1999 instead of 1949. As a consequence the "old" grade has gone up by 10 points (as one would expect). Also according to the grading database, there are only 15 games taken into account presumably all at Gibraltar. Version 3 of the download had the recalculated "new" grade as 153. As far as I can tell, the opposition in the U1800 at Gibraltar was mostly or exclusively adult opposition. Therefore one would expect the v5 recalculated grade (after the false age was input) to come out at 168 or thereabouts given that the new junior increment is supposed to be 15.

The grade is only 15 numbers plus a 16th for the junior increment. What 15 numbers do you add up, divide by 15 and add something to get 196?

Richard Bates
Posts: 3338
Joined: Fri Nov 14, 2008 8:27 pm

Re: Database Upgrade

Post by Richard Bates » Sat Mar 07, 2009 7:28 pm

Roger de Coverly wrote:
Plays at Gibraltar in the U1800 section

http://www.gibraltarchesscongress.com/g ... teurs.html
Actually doesn't this case demonstrate that the "new" junior grades are just plain wrong and may be double ,triple or even quadruple counting the junior increment? It looks to me as if an incorrect date of birth was entered into the ECF grading database, perhaps 1999 instead of 1949. As a consequence the "old" grade has gone up by 10 points (as one would expect). Also according to the grading database, there are only 15 games taken into account presumably all at Gibraltar. Version 3 of the download had the recalculated "new" grade as 153. As far as I can tell, the opposition in the U1800 at Gibraltar was mostly or exclusively adult opposition. Therefore one would expect the v5 recalculated grade (after the false age was input) to come out at 168 or thereabouts given that the new junior increment is supposed to be 15.

The grade is only 15 numbers plus a 16th for the junior increment. What 15 numbers do you add up, divide by 15 and add something to get 196?
The other bizarre thing is that it looks as if this was a recent input error. This chap was originally published at 136, and was altered to 146 (presumably after addition of the originally omitted junior increment) on the amended list!

Sean Hewitt

Re: Database Upgrade

Post by Sean Hewitt » Sun Mar 08, 2009 1:46 pm

The player concerned is Robert Rough.

He seems to have played 5 games graded by the ECF in each of the last three seasons, and my guess is the same as Rogers - ie this is the Gib Amateur each season. He certainly played this season as I met and had dinner with him there!

However, for some reason, when that tournament was graded an incorrect date of birth was submitted. I think I may know why. The tournament used a piece of software used for FIDE events to track and report the results and that software will not allow results to be reported without a date of birth. The organisers probably didnt know his date of birth, and stuck 01/01/2000 in. When the results were extrapolated and sent to the ECF the ficticious date of birth remained. If the ECF grading system didnt have his date of birth then it will believe the first event that provides it. Hence Robert Rough became 8 years old and his "old" grade went from 136 to 146 due to the 10 point junior increment.

However, I have no idea how the new junior grades have been calculated so can offer no explanation as to how the knowledge that a player is 8 years old can have a 43 point grading difference when the player has played exactly the same opposition.

I would have thought that Howard Grist is the person tp answer that.

I can say that I have had Robert's date of birth corrected, so his grade will be corrected on the next update. Viva la forum!

Howard Grist
Posts: 83
Joined: Tue Dec 12, 2006 1:14 pm
Location: Southend-on-Sea
Contact:

Re: Database Upgrade

Post by Howard Grist » Sun Mar 08, 2009 5:26 pm

Roger de Coverly wrote:
Howard Grist wrote: However the central principle of the ECF grading system is that the pool of adult players is constant.
Why?
As this the central principle, I'm not sure that it can be 'explained'. However if it wasn't the case, you could not sensibly carry forward games played by players played over different seasons.
Roger de Coverly wrote:Historically the health of the system was measured from the average grade of all the participants. Are you now saying that we should expect to see the average grade of all players increasing?
It is far easier to quote and calculate your statistic than the one that I have quoted. However if you subscribe to your philosophy then you would start shouting inflation when the mean grade rises when 500 GMs start playing in this country. I would expect the mean grade of all participants to rise in this instance and this would not indicate a breakdown of the grading system.
Roger de Coverly wrote:Is it difficult? You take the 2007 published grades You remove the junior increments which presumably were added at the end after the 2007 performances had been calculated. You either run 2008 with no junior increment at all or with whatever formula you wish to test. You compare the resulting 2008 grades with the "actuals". You've then got some benchmarks for the effect of different formulae. I doubt that the players above 200 would be affected by any change involving under 10 s but the under 100s might well be.
It's more diificult than you are saying here. This can only be done sensibly if you assume that the 2007 published grades are correct. The 2007 grades - particularly for juniors - include large chunks of junior increment from 2007, and smaller chunks of increment from previous years. If these increments are incorrect then you will not get useful answers from such a test.
Roger de Coverly wrote:That's an artificial construction because there is (or rather was) no such thing as a grade calculated against other adults. The way the junior increment was tested historically was to follow players through the age ranges and see how their grading improved with age and experience. This measures their performance both against other adults and other juniors. The strength of a junior increases not just against adults but also against their fellow and younger peers.
Historically we did not have the grading information that we do now. We know that juniors improve, but by how much? You can't measure this by comparing them with other juniors who are also improving, you can only do it by comparing against the general pool of adult players - which is assumed to be constant.
Roger de Coverly wrote:Reality check - do you really believe in under 10 players of a 150 standard?
Good question - on the old basis, no - but on the new basis 150 is the old 130, and juniors have been undergraded for some considerable time. So it's really a question of changing your ides on these things. Unfortunately, I still don't have quite enough data to do some of the checks that I'd like to do - such as whether the 10 year-olds this year are significantly better than those of last year as not all the grades are calculated using three years worth of data. I don't think that they are as far out as you seem to think though.
Former ECF Grading System Programmer

Locked