FIDE Ratings - Part Results

General discussions about ratings.
DonaldMoir
Posts: 15
Joined: Tue Dec 20, 2016 1:43 pm

FIDE Ratings - Part Results

Post by DonaldMoir » Thu Mar 01, 2018 3:08 pm

I am fortunate in being able to spend part of my retirement in south India. A good friend is principal of a school there and I quite often accompany a party of schoolchildren to FIDE-Rating tournaments, of which there are many.

A problem has slowly come to my attention. Quite a few of the children have gone to rating tournaments and achieved one good result but a lot of bad ones. For example, one 13-year-old boy achieved one draw and 12 defeats against rated players. Now, almost a year later, his chess has improved and a recent rating tournament he beat four out of eight rated players, giving him a rating performance for that tournament of 1179. But this has not been enough to give him a rating. (If my own calculations are correct, he probably now has a 'rating' of about 960.)

Put in another way, it would have been better for him to have lost that initial match rather than drawn it. Then he would not have had a 'part result' and would now have a rating of 1179. At that same tournament, to my horror, I found myself advising another child, who had lost three games against rated players, not to try to win or draw his final game so as not to get a 'part result' which would handicap him in future tournaments.

Surely - at least for children - it would be better not to have this system. As it is, the logic would suggest that we should not take promising children to such tournaments until they are overwhelmingly certain to get a rating.

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

Re: FIDE Ratings - Part Results

Post by Roger de Coverly » Thu Mar 01, 2018 3:20 pm

DonaldMoir wrote:
Thu Mar 01, 2018 3:08 pm

Surely - at least for children - it would be better not to have this system. As it is, the logic would suggest that we should not take promising children to such tournaments until they are overwhelmingly certain to get a rating.
It's not usually a problem in the UK as we have a domestic system with its own rules. Events aimed at inexperienced players would not be FIDE rated. For junior players, the UK system just "forgets" results from earlier time periods.

It's actually quite old advice to keep your potential high flyers out of the FIDE rating system until they can come in at a respectable rating.

User avatar
Paolo Casaschi
Posts: 1186
Joined: Thu Jan 08, 2009 6:46 am

Re: FIDE Ratings - Part Results

Post by Paolo Casaschi » Thu Mar 01, 2018 7:50 pm

DonaldMoir wrote:
Thu Mar 01, 2018 3:08 pm
Surely - at least for children - it would be better not to have this system. As it is, the logic would suggest that we should not take promising children to such tournaments until they are overwhelmingly certain to get a rating.
This is true only if you care more about rating than about playing chess. What’s the problem with just playing chess, possibly against the strongest opponents you can find, and let the rating system do its own thing without manipulating it?

NickFaulks
Posts: 8452
Joined: Sat Jan 02, 2010 1:28 pm

Re: FIDE Ratings - Part Results

Post by NickFaulks » Fri Mar 02, 2018 12:13 am

Paolo Casaschi wrote:
Thu Mar 01, 2018 7:50 pm
DonaldMoir wrote:
Thu Mar 01, 2018 3:08 pm
Surely - at least for children - it would be better not to have this system. As it is, the logic would suggest that we should not take promising children to such tournaments until they are overwhelmingly certain to get a rating.
This is true only if you care more about rating than about playing chess. What’s the problem with just playing chess, possibly against the strongest opponents you can find, and let the rating system do its own thing without manipulating it?
Thanks Paolo, that needed saying. The view has been expressed on this forum, by a senior and highly respected coach, that English juniors should avoid testing themselves against their peers overseas because of the danger that they might drop rating points. It makes you want to cry.
If you want a picture of the future, imagine a QR code stamped on a human face — forever.

DonaldMoir
Posts: 15
Joined: Tue Dec 20, 2016 1:43 pm

Re: FIDE Ratings - Part Results

Post by DonaldMoir » Fri Mar 02, 2018 1:24 am

There is no problem with these chess-mad Indian youngsters not playing chess. About 250 participate in school chess electives weekly and more than 100 regularly go to tournaments, even though most of them never win even a small trophy.

But, in this delightfully old-fashioned part of rural India, chess is highly respected and popular and a FIDE-rating is a valued mark of progress. I am trying to grapple with the FIDE’s rationale for not counting a succession of defeats if a child fails to score a point, but holding all the defeats against him/her if even a single draw is achieved.

Kevin Thurlow
Posts: 5802
Joined: Wed Apr 30, 2008 12:28 pm

Re: FIDE Ratings - Part Results

Post by Kevin Thurlow » Fri Mar 02, 2018 9:12 am

" I am trying to grapple with the FIDE’s rationale for not counting a succession of defeats if a child fails to score a point, but holding all the defeats against him/her if even a single draw is achieved."

Rating systems in general don't work if someone scores 0 % or 100 %. So I would not be inclined to blame FIDE here.

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

Re: FIDE Ratings - Part Results

Post by Roger de Coverly » Fri Mar 02, 2018 10:54 am

Kevin Thurlow wrote:
Fri Mar 02, 2018 9:12 am
Rating systems in general don't work if someone scores 0 % or 100 %. So I would not be inclined to blame FIDE here.
The anomaly is that a result of 0 from 5 for a new player is forgotten whilst a result of 1/2 from 5 isn't. So if before the last round of a tournament, a player's score is 0/4 and the desire is to eventually obtain as high a FIDE rating as possible, the gaming advice would be to either lose or withdraw from the tournament. It is possible to rate players who score 0% or 100%, you use the approximation sometimes known as the 400 point rule. Others might know it as the BCF/ECF method.

I'm sure FIDE could tinker with the rules as to what happens when initial performances are below the 1000 minimum threshold, so that improving players who entered their first tournament too early don't get ratings well below their strength or be given incentives to lose or withdraw.

User avatar
Paolo Casaschi
Posts: 1186
Joined: Thu Jan 08, 2009 6:46 am

Re: FIDE Ratings - Part Results

Post by Paolo Casaschi » Fri Mar 02, 2018 3:20 pm

Roger de Coverly wrote:
Fri Mar 02, 2018 10:54 am
I'm sure FIDE could tinker with the rules as to what happens when initial performances are below the 1000 minimum threshold, so that improving players who entered their first tournament too early don't get ratings well below their strength or be given incentives to lose or withdraw.
The obvious solution is to make sure more and more tournaments are FIDE rated so that everyone playing any organized chess gets on the FIDE rating ladder as soon as possible, avoiding any manipulation of the initial rating that would follow any tinkering with artificial rules. Ultimately, the more games you throw into the rating system the more it will become accurate, no tinkering required.

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

Re: FIDE Ratings - Part Results

Post by Roger de Coverly » Fri Mar 02, 2018 5:18 pm

Paolo Casaschi wrote:
Fri Mar 02, 2018 3:20 pm
The obvious solution is to make sure more and more tournaments are FIDE rated so that everyone playing any organized chess gets on the FIDE rating ladder as soon as possible
That's what India are doing, which leads to the problems described in the first post. Elo's system of rating doesn't in my view cope terribly well where players have rapidly changing strengths. Perhaps no rating system does.

NickFaulks
Posts: 8452
Joined: Sat Jan 02, 2010 1:28 pm

Re: FIDE Ratings - Part Results

Post by NickFaulks » Fri Mar 02, 2018 5:26 pm

Roger de Coverly wrote:
Fri Mar 02, 2018 5:18 pm
Elo's system of rating doesn't in my view cope terribly well where players have rapidly changing strengths. Perhaps no rating system does.
Analysis of the data suggests otherwise, subject to games being submitted for rating.
If you want a picture of the future, imagine a QR code stamped on a human face — forever.

Kevin Thurlow
Posts: 5802
Joined: Wed Apr 30, 2008 12:28 pm

Re: FIDE Ratings - Part Results

Post by Kevin Thurlow » Fri Mar 02, 2018 6:06 pm

1000 is quite low of course. I believe India does not have its own rating system, so if you want any rating, you have to go for the FIDE one.

It does remind me that when I got mine - I think the floor was 2000 then, I just grovelled on with 5 games at 2060 and 4 at 2020 (from memory). At my next tournament I had 5 games at 2670. I spent a few years thinking I should have avoided the second tournament... Obviously, the first two were a bit nearer the truth!

I agree that people should not be encouraged to throw games. If you do that, why bother playing?

Nick Grey
Posts: 1838
Joined: Thu Dec 01, 2011 12:16 am

Re: FIDE Ratings - Part Results

Post by Nick Grey » Fri Mar 02, 2018 8:40 pm

On FIDE games against unrated they are not usually rated.
When difference is +/- 400 not rated. But you get +/- 10 on ECF.
In this country not everyone wants to play FIDE because there is a significant ECF annual fee difference.

Sometimes it is difficult to explain to adults wanting to improve their rating they cannot afford to lose or draw against a lower rated player, and pick up a percentage of positive results against higher rated players.

Rating is not everything.

DonaldMoir
Posts: 15
Joined: Tue Dec 20, 2016 1:43 pm

Re: FIDE Ratings - Part Results

Post by DonaldMoir » Wed May 02, 2018 2:25 pm

Sorry to raise this issue again. Would anyone be able to comment on this specific example - a 13-year-old Indian boy who is mad keen on chess.

He has just participated in another rating tournament where he did pretty well. But he still cannot overcome the legacy of his previous poor results last year.

His FIDE "part results" record is:
February 2017: 1/5 Rp 910
February 2017: 0/2 Rp 286
September 2017: 0/4 Rp 300
October 2017: 2.5/6 Rp 1074
December 2017: 0,5/4 Rp 807
February 2018: 4.5/9 Rp 1233
March 2018: 1.5/6 Rp 903
May 2018: 3/7 Rp 1102

Clearly he made a big error in winning one point in February 2017.

Is there any way of calculating, from the above results, how close he is to reaching the magic 1000 mark?

NickFaulks
Posts: 8452
Joined: Sat Jan 02, 2010 1:28 pm

Re: FIDE Ratings - Part Results

Post by NickFaulks » Wed May 02, 2018 2:38 pm

I'm always interested in looking at individual cases, but don't have time to do the preliminary detective work to find out who this is.
If you want a picture of the future, imagine a QR code stamped on a human face — forever.

LawrenceCooper
Posts: 7167
Joined: Tue Dec 20, 2011 8:13 am

Re: FIDE Ratings - Part Results

Post by LawrenceCooper » Wed May 02, 2018 2:53 pm

NickFaulks wrote:
Wed May 02, 2018 2:38 pm
I'm always interested in looking at individual cases, but don't have time to do the preliminary detective work to find out who this is.
There's only 28 pages of unrated Indians born in 2005 and 30 pages for those born in 2004 so it won't take you long :wink:

Post Reply