English Chess Forum

Does anybody have any thoughts about the proposed changes to the grading system in 2020 ? In my opinion, monthly grading just adds another layer of confusion. Who does it really benefit ? Most organisations use the July/August list for their competitions, and the January list for those without an official grade in the earlier list. Some Congresses will use the January list, but a monthly list ? Really ?

The move to a 4 digit ELO type grade from the 3 digit ECF grade has obvious merits and should probably be supported.

The ECF claim that they are responding to the expectations of existing and potential new members. Well I know hundreds of chess players and none of them have an expectation that includes monthly grading !!

Steve Emmerton wrote: ↑

Sun Feb 17, 2019 10:10 am

Who does it really benefit ?

The claim is that it benefits new players and those, like juniors, whose standard of play has changed. Congresses can if they wish, base pairings and eligibility on a more up to date rating. Those which are FIDE rated often do this anyway.

It's a change of method as well, using new rating = old rating +/- k * (difference from expected score) rather than the new grade = performance over measuring period.

Using four figure numbers and the same notional scale will make differences between the national and international ratings rather more obvious.

A deliberate imbalance is proposed for juniors. If for example a junior plays adults of the same rating and scores 50%, their rating will increase if the 50% contains both wins and losses.

The other claim is that people simply find it fun to see their grade move about on a regular basis, so might play more or something. Baffles me but maybes there's an effect

Personally I think that an ELO system, especially a monthly one, would undermine the practice of using a single month’s rating (eg. July/January) for annual competitions. With the ELO rating being more a reflection of current form (and for that matter arguably more open to manipulation) aroundthe eg. July cutoff than the current ECF methodology.

It is perhaps worth considering the reasons why the ECF system was designed (in preference to ELO) in the first place, and whether the assessed disadvantages (as they relate to the English chess setup) back then of ELO still apply (and/or whether they are mitigated by more frequent calculation)

Also it is interesting to note (if I interpreted the Board minutes correctly) that the Board seems to favour moving to a 4 digit grade, regardlessof the calculation methodology. Whereas I think that this is something that is actually undesirable given the confusion that it could cause between the two rating lists (Ecf/FIDE)

The very first line in the ECF statement on the website actually says 'The ECF Board is responding to changes in the expectations of both members and potential members'. This would imply that existing members have been contacting the Board to say 'I'd really like to see my grade updated every month' - I don't think so !! I have spoken to a couple of dozen of the many ECF Members that I know and I've yet to find a single one who has this expectation or desire - in fact totally the opposite. And, unless they have Derek Acora on the Board, how can they possibly know what the expectations of potential future Members might be ? No, this proposal is purely for the benefit of junior players, who form a minor part of the greater ECF Membership, and who can't wait a few months for a new grade. This proposal should be buried as soon as possible and never see the light of day again.

I have pupils at my (Secondary) school asking me every week about their new grades. Some are yet to get their first grade, who get the most frustrated, some have just got their first grade and others are more experienced. Monthly grades will be greatly appreciated by junior players who are used to their grades on line (both chess and video games) changing after every game. It would be something else to help encourage junior players.

Neill Cooper wrote: ↑

Mon Feb 18, 2019 9:18 am

Monthly grades will be greatly appreciated by junior players who are used to their grades on line (both chess and video games) changing after every game.

The ECF have ruled out going for continuous grades. Are they mistaken?

Roger de Coverly wrote: ↑

Mon Feb 18, 2019 10:46 am

The ECF have ruled out going for continuous grades. Are they mistaken?

Annual August grades for eligibility are the bedrock of the grading system. 6 monthly grades have been valuable, particularly for new and improving players. Monthly grades will be much appreciated. Continuous grades are much easier to calculate online than for Over The Board chess but can work, as the Yorkshire "Chessnuts" grading list showed. I think nationally it might be more of a challenge so the ECF are not mistaken to rule them out.

"The very first line in the ECF statement on the website actually says 'The ECF Board is responding to changes in the expectations of both members and potential members'."

So people who are not yet members are asking how quickly they will get a grade?

Most importantly, a new player should not have to wait more than a month to get their first grade. This should be the main priority. Whether the whole list is updated live is not so much of an issue, though it should be an ambition. Ideally the gradings should be updated on the first (maybe a first Wednesday) of the month, like FIDE ratings - and organisers might be obliged to use them for eligibility in tournaments.

Kevin Thurlow wrote: ↑

Mon Feb 18, 2019 6:56 pm

"The very first line in the ECF statement on the website actually says 'The ECF Board is responding to changes in the expectations of both members and potential members'."

So people who are not yet members are asking how quickly they will get a grade?

As well as people who are members asking when will they get their first grade.

I am aware that existing members might wonder about their first grade - I was wondering how many people who were not yet members would. In my experience, most new members at clubs don't know what grades are. It may be that new junior players are aware as they have spoken to their classmates, and of course monthly grades are primarily (if not exclusively) aimed at juniors. I'm not saying that's a bad thing.

Neill Cooper wrote: ↑

Mon Feb 18, 2019 12:00 pm

Annual August grades for eligibility are the bedrock of the grading system.

In an Elo system designed on FIDE lines, that goes out of the window as all rating lists should be equally valid. That said the summer ones will be the stable ones for club players as they may not be playing between the end of one season and the start of the next.

Congresses will however have to choose a rating list date for eligibility definitions. That's what FIDE rated ones now do anyway. One technical point though. It's argued that the system needs to be tolerant of late result submission and it's suggested as a consequence that ratings be changed retrospectively. I'm unconvinced this is a good idea, because of the Congress eligibility point. The FIDE system seemed to cope with having 4NCL results all processed for the end May list, even though some of the games had been played the previous October or November.

Roger de Coverly wrote: ↑

Tue Feb 19, 2019 9:13 pm

[One technical point though. It's argued that the system needs to be tolerant of late result submission and it's suggested as a consequence that ratings be changed retrospectively. I'm unconvinced this is a good idea, because of the Congress eligibility point.

That depends on what you publish. You could recalculate all ratings going as far back as you like, but only publish the current month's ratings, so there is never any change to historical published ratings. That would stop someone becoming ineligible to play in an event that had already taken place (although it wouldn't stop someone who was so inclined to do there own calculations to work out what a corrected historical rating would have been).

Roger de Coverly wrote: ↑

Tue Feb 19, 2019 9:13 pm

The FIDE system seemed to cope with having 4NCL results all processed for the end May list, even though some of the games had been played the previous October or November.

it copes, but it doesn't change historical ratings. Doesn't it just treat late results as if they were played in the current period, but with the ratings that applied when they were really played? That means that games played in periods between when the late results were really played and when they were rated won't be recalculated to give their corrected rating change.