ECF grades compared with FIDE ratings

General discussions about ratings.
Barry Sandercock
Posts: 1356
Joined: Tue Sep 11, 2012 10:52 am

Re: ECF grades compared with FIDE ratings

Post by Barry Sandercock » Fri Jan 24, 2014 11:45 am

I don't see how Elo grades and be compared with ECF grades because now Elo grades go down to 1000, many games are played against players without an Elo grade so only register as ECF and not Elo grades.

Jonathan Bryant
Posts: 3452
Joined: Sun May 11, 2008 3:54 pm

Re: ECF grades compared with FIDE ratings

Post by Jonathan Bryant » Fri Jan 24, 2014 7:18 pm

MartinCarpenter wrote:I can see how all of the missing games will produce randomness for FIDE vs ECF grades. Some of the low/medium level FIDE grades in 4NCLN look very random in both directions.

What I can't really think is how that might produce systemic deflation - surely you're as likely to have clustered losses (or at least bad results) in those non counted games as wins? The juniors do of course spread their low grades because they tend to be quite active with it.....

To be honest, below quite a high level, I'd just ignore UK FIDE grades.

Well I'd agree with your last sentence sadly. Certainly for the tournaments in which I play.


As for, "surely you're as likely to have clustered losses ... as wins", this is simply not my experience. My score from non-counted games since I first got a FIDE rating in 2011 is +13 =5 -1. I'm prepared to believe that for reasons of bad luck or whatever I could be at the extreme end of the scale, I struggle to accept - until somebody actually proves otherwise - that my experience is totally different from everybody else's.

It's logical if you think about it. Lesser experienced players are less likely to have FIDE ratings and also less likely to be good. Therefore I'm more likely to beat them.

My experience in rapidplay games is very similar, incidentally.


I don't buy the 'correction for past inflation argument' btw. My first rating was acquired almost 50% in Spain. I'm convinced that's partly why it came out as high as it did (i.e. within 1 point an exact correlation with my ECF grade under the *8 +650 formula). If UK elo ratings were some how correcting to accuracy I'd have struggled against the typically higher-rated people that I met in Benasque. But I didn't.



No, my feeling is that in the tournaments that I play the elo rating system has simply broken down. There's no randomness in both directions as Martin finds elsewhere. Nearly everybody I meet has an elo rating that indicates a playing strength significantly below that suggested by their ECF.
Last edited by Jonathan Bryant on Fri Jan 24, 2014 7:32 pm, edited 1 time in total.

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

Re: ECF grades compared with FIDE ratings

Post by MartinCarpenter » Fri Jan 24, 2014 7:30 pm

Not impossible that it has :) Some of the 4NCL Div3N people do definitely have higher Elos than you'd expect but they're pretty well insulated from juniors.

As for unrated games, yes you will expect to have more net wins as they'll likely tend to be weaker/less experienced players. Its grading performance vs your 'norm' that matters of course and what that'll do is much less obvious. Some (many?) players do do better grading performance wise vs mildly weaker folk so it might be systemically hurting them a bit. Vice versa for some others.
(And some people will be notably (un)lucky!)

Jonathan Bryant
Posts: 3452
Joined: Sun May 11, 2008 3:54 pm

Re: ECF grades compared with FIDE ratings

Post by Jonathan Bryant » Fri Jan 24, 2014 7:43 pm

MartinCarpenter wrote:Not impossible that it has :) Some of the 4NCL Div3N people do definitely have higher Elos than you'd expect but they're pretty well insulated from juniors.

As for unrated games, yes you will expect to have more net wins as they'll likely tend to be weaker/less experienced players. Its grading performance vs your 'norm' that matters of course and what that'll do is much less obvious. Some (many?) players do do better grading performance wise vs mildly weaker folk so it might be systemically hurting them a bit. Vice versa for some others.
(And some people will be notably (un)lucky!)
I smell a blog post here. Perhaps even a series. I shall have to go through my uncounted games and see what my performance actually was ECF wise (slightly complicated by the fact we're talking several grading periods). My sense is, however, that it will turn out to see me punching above my weight.

I find that you can divide my unrated opponents into three main camps.

(a) adults new to tournament play (n=2 or 3 probably)
(b) juniors under the age of 12
(c) experienced adults who have a stable ECF in the 140-160 range but who have, for whatever reason, not got a FIDE yet.


Like I say, I accept the theoretical possibility that I could have been really unlucky in rapid play and standard play for several years running, but I think I'd like to see the evidence first before I jump to that conclusion.

Mind you, I once had a tournament that I played when my rapidplay ECF was at its highest ever point and my rapidplay FIDE was close to its lowest ever point. My ECF TPR game out as an improvement and yet I lost Elo points - to a great extent because of uncounted games. That kind of experience does rather warp your perception, I imagine.

Jonathan Bryant
Posts: 3452
Joined: Sun May 11, 2008 3:54 pm

Re: ECF grades compared with FIDE ratings

Post by Jonathan Bryant » Fri Jan 24, 2014 7:55 pm

For balance I should probably acknowledged that while the elo rating system in and around London is totally FUBAR there can be some beneficial side effects.

e.g. 1 the fact that some strong players can get into u2200 events make those tournaments more attractive to me
e.g. 2 that only 8 of 12 games I played at e2e4 events in May/June 2011 were against rated opposition meant I still didn't have an elo rating when I went to Benasque in July. And I picked up 200 euros for best unrated score as a result.

User avatar
Joey Stewart
Posts: 1860
Joined: Wed Apr 11, 2007 2:35 pm
Location: All Of Them

Re: ECF grades compared with FIDE ratings

Post by Joey Stewart » Wed Jan 29, 2014 2:59 pm

It seems that around 1900 the ratings get really random - A scan of the 4ncl registrations shows players ranging between 140 and 190 (the higher ones being way outside of the converted formula).
If it continues then I can see a future where it will no longer be possible to gain master titles, apart from exceptional players, as the up and coming juniors will be drowned in a tide of under rated opposition and never come close to getting the rating requirement needed.
Lose one queen and it is a disaster, Lose 1000 queens and it is just a statistic.

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

Re: ECF grades compared with FIDE ratings

Post by MartinCarpenter » Wed Jan 29, 2014 3:43 pm

Well I can't really see how it'll fix itself :( The cohort with random grades aren't playing enough to reset their FIDE grades on a sane time scale and are nearly all playing each other rather than the people with 'good' FIDE grades anyway.

User avatar
Christopher Kreuzer
Posts: 8806
Joined: Fri Aug 06, 2010 2:34 am
Location: London

Re: ECF grades compared with FIDE ratings

Post by Christopher Kreuzer » Wed Jan 29, 2014 6:40 pm

MartinCarpenter wrote:Well I can't really see how it'll fix itself :( The cohort with random grades aren't playing enough to reset their FIDE grades on a sane time scale and are nearly all playing each other rather than the people with 'good' FIDE grades anyway.
Agreed. And even if you make the effort to play for a sustained period against people with higher ratings, and get consistent results, and your rating goes up, you run into the same problem if you return to playing chess in areas where there is possible (probable?) under-rating. As Martin says, one of the main problems is under-rated players of about the same strength playing each other and reinforcing the effect. Does this all go back to the lowering of the rating floor, or was there something else that happened?

Jonathan Bryant
Posts: 3452
Joined: Sun May 11, 2008 3:54 pm

Re: ECF grades compared with FIDE ratings

Post by Jonathan Bryant » Wed Jan 29, 2014 7:31 pm

Christopher Kreuzer wrote:... or was there something else that happened?
I suspect the increased frequency of rating lists hasn't helped. I also wonder - as per my 'non counted games' problem - whether the system might work much better in countries where most actives players have elo ratings (e.g. as seems to be the case in Europe) and struggles more in England because many do not (because club chess isn't elo rated here).



As for Joey's point: "If it continues then I can see a future where it will no longer be possible to gain master titles"

I seem to remember Richard Bates making a similar point on here in the past, but perhaps I'm wrong.

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

Re: ECF grades compared with FIDE ratings

Post by Matt Mackenzie » Wed Jan 29, 2014 7:41 pm

Though presumably before it reaches anywhere near that point, something will be done about it.....
"Set up your attacks so that when the fire is out, it isn't out!" (H N Pillsbury)

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

Re: ECF grades compared with FIDE ratings

Post by MartinCarpenter » Thu Jan 30, 2014 9:59 am

What though?

What you'd need to do would be something like scrapping all the extant FIDE grades under a certain level and start everyone off from fresh with their converted ECF grade as their initial FIDE - not just estimated. You'd still have some issues with fast improvers (juniors!) updating very slowly etc but it'd help an awful lot.

Can you really imagine FIDE allowing either of those steps?

Yes, it'll work much better in countries where 'everything' is FIDE rated - the formula/ system is very good of course, but it presumes a half decent activity level to stay current. Reasonable as people not playing don't move much.

However people playing lots of games not in the system? That's a big problem. For any rating system of course :)

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

Re: ECF grades compared with FIDE ratings

Post by Roger de Coverly » Thu Jan 30, 2014 10:13 am

MartinCarpenter wrote:What though?
We might be flattering ourselves to think that ECF levels are correct and the International ones wrong. The ECF did after all add between 20 to 25 ECF points to the average player, bumping the centre of the grade distribution from around the 115 level to around the 140 level. Most players below international level rarely play anyone from outside the UK, so there's little evidence as to whether they are under performing or outperforming against the rest of the world. International junior events may offer clues, little in the performances of English players in recent events does much to suggest that they are underrated against their peers in other countries.

Performances in Gibraltar may be worth a look as it features English club players, but not enough of them that they play amongst themselves for at least half the games as would be the case at Hastings and the London Class FIDE Open.

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

Re: ECF grades compared with FIDE ratings

Post by MartinCarpenter » Thu Jan 30, 2014 10:32 am

Oh, I don't know how the current ECF conversion etc works.

The one thing that is clear is that the current ECF grades are pretty good at sorting players compared to each other - even if the absolute levels of peoples grades have got slightly messed up by all those changes. The FIDE grades at a low level really aren't.

Barry Sandercock
Posts: 1356
Joined: Tue Sep 11, 2012 10:52 am

Re: ECF grades compared with FIDE ratings

Post by Barry Sandercock » Thu Jan 30, 2014 10:41 am

I quite agree. FIDE grades at the lower level are very misleading as to the players strength. Only games against FIDE rated players are counted and these can be a small percentage of the player's games.

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

Re: ECF grades compared with FIDE ratings

Post by Roger de Coverly » Thu Jan 30, 2014 10:52 am

Barry Sandercock wrote:I quite agree. FIDE grades at the lower level are very misleading as to the players strength.
That gives an interesting side-effect. At events such as e2e4 which use FIDE ratings for seeded pairings, the distortions to the ranking order can give the effect of an old-fashioned random pairing system, that could potentially pair the strongest players in an early round.