ECF to scrap its grading system

Debate directly related to English Chess Federation matters.
Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

ECF to scrap its grading system

Post by Roger de Coverly » Thu Dec 06, 2018 11:30 pm

From the December ECF newsletter
This method of calculation is statistically unsound once the frequency of calculation is such that the number of results in each batch for the majority of players is small. Moving to more frequent grading lists will require us to replace the ECF calculation method with a more suitable one.

John McKenna

Re: ECF to scrap its grading system

Post by John McKenna » Fri Dec 07, 2018 12:11 am

Scrap?

A souped-up version based on the old chassis is more likely, knowing the ECF.

NB: Roger's early allergic reaction is probably due to a syndrome that is rooted in a chronic distrust of statistical, as opposed to purely numerical, methods.

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

Re: ECF to scrap its grading system

Post by Roger de Coverly » Fri Dec 07, 2018 12:23 am

John McKenna wrote:
Fri Dec 07, 2018 12:11 am
NB: Roger's early allergic reaction is probably due to a syndrome that is rooted in a chronic distrust of statistical, as opposed to purely numerical, methods.
It's rather more a question of due process. Any such change should go to the ECF Council for approval on the grounds of the impact on every competitive chess player. If the Directors dictate the change, that poses the question as to what is the use of Council and why such decisions cannot be directly voted.

The now deceased Yorkshire system managed to produce continuous grades without having to use an Elo methodology.

Adopting an Elo system brings up the problem that has bedevilled all such methods, namely how to deal with rapidly improving players who don't play so much chess captured for a rating system.

Mick Norris
Posts: 10329
Joined: Tue Apr 17, 2007 10:12 am
Location: Bolton, Greater Manchester

Re: ECF to scrap its grading system

Post by Mick Norris » Fri Dec 07, 2018 8:03 am

The calculation of grades for players and the publication of grading lists is recognised as being a core function of the Federation. Annual grading lists have been published since time immemorial, and for the past few years the time interval has been six-monthly. The Board recognise that there is a clear demand from many sections of the chess community for more frequent grading lists. They have set a target of providing monthly grading capability for the 2019/2020 season. The present method of calculating grades was devised to handle batch processing of of results, with each batch containing a significant number of results for each active player. This method of calculation is statistically unsound once the frequency of calculation is such that the number of results in each batch for the majority of players is small. Moving to more frequent grading lists will require us to replace the ECF calculation method with a more suitable one. For this reason there is no merit in a half-way house of, for example, quarterly grading lists.

We assure organisers that, at least for a transitional period, there will be no requirement for leagues or club competitions to submit interim results while the event is still in progress, although as at present with six-monthly lists such interim submissions will be encouraged.
I think therefore Council gets to discuss this before it comes in?
Any postings on here represent my personal views

David Sedgwick
Posts: 5249
Joined: Mon Apr 09, 2007 5:56 pm
Location: Croydon

Re: ECF to scrap its grading system

Post by David Sedgwick » Fri Dec 07, 2018 8:40 am

The calculation of grades for players and the publication of grading lists is recognised as being a core function of the Federation. Annual grading lists have been published since time immemorial, and for the past few years the time interval has been six-monthly. The Board recognise that there is a clear demand from many sections of the chess community for more frequent grading lists. They have set a target of providing monthly grading capability for the 2019/2020 season. The present method of calculating grades was devised to handle batch processing of of results, with each batch containing a significant number of results for each active player. This method of calculation is statistically unsound once the frequency of calculation is such that the number of results in each batch for the majority of players is small. Moving to more frequent grading lists will require us to replace the ECF calculation method with a more suitable one. For this reason there is no merit in a half-way house of, for example, quarterly grading lists.

We assure organisers that, at least for a transitional period, there will be no requirement for leagues or club competitions to submit interim results while the event is still in progress, although as at present with six-monthly lists such interim submissions will be encouraged.
Mick Norris wrote:
Fri Dec 07, 2018 8:03 am
I think therefore Council gets to discuss this before it comes in?
Only if requisitionists force such a discussion, I suspect.

If the new system is going be introduced for 2019 - 2020, that would have to happen at the Finance Council Meeting on Saturday 27th April 2019.

I am still feeling drained after the battle to save the Counties Championships, successful though it was. I don't have the energy to fight to save Clarke gradings.

The 100th Counties Championships Finals on Saturday 13th July 2019 may be the last event which is Clarke graded.

Does anyone want to organise a wake?

Angus French
Posts: 2151
Joined: Thu May 15, 2008 1:37 am

Re: ECF to scrap its grading system

Post by Angus French » Fri Dec 07, 2018 9:30 am

Mick Norris wrote:
Fri Dec 07, 2018 8:03 am
I think therefore Council gets to discuss this before it comes in?
David Sedgwick wrote:
Fri Dec 07, 2018 8:40 am
Only if requisitionists force such a discussion, I suspect.

If the new system is going be introduced for 2019 - 2020, that would have to happen at the Finance Council Meeting on Saturday 27th April 2019.
The last Home Director did at least commit - at the 2016 Finance Council Meeting - to consult with leagues and congresses about the introduction of monthly grading lists.

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

Re: ECF to scrap its grading system

Post by Roger de Coverly » Fri Dec 07, 2018 10:43 am

ECF anniuncement wrote: This method of calculation is statistically unsound once the frequency of calculation is such that the number of results in each batch for the majority of players is small.
That's not what happens though. Every player is assessed on the basis of at least a year's results, with a top up to 30 games or three year's results if 30 games are not unavailable. I wouldn't be sure how to interpret the results, but the same could be done monthly. Are the monthly fluctuations in an Elo based system of that much significance?

Rather than take a leap into the dark and risk a total wreck of the grading system, wouldn't a parallel run be prudent?

User avatar
Michael Farthing
Posts: 2069
Joined: Fri Apr 04, 2014 1:28 pm
Location: Morecambe, Europe

Re: ECF to scrap its grading system

Post by Michael Farthing » Fri Dec 07, 2018 11:14 am

Roger de Coverly wrote:
Fri Dec 07, 2018 10:43 am
Rather than take a leap into the dark and risk a total wreck of the grading system, wouldn't a parallel run be prudent?
Now that is a positive suggestion, Roger.

J T Melsom
Posts: 1294
Joined: Wed Aug 12, 2009 11:12 pm

Re: ECF to scrap its grading system

Post by J T Melsom » Fri Dec 07, 2018 11:17 am

I hold no emotional attachment to the current grading system. It produces a number to which we then apply meaning. Any new system would do the same thing and we would come to understand it just as we did the old. The test on which I have no specific expertise is not tradition, beware a mis-use of tradition to apply to 20th century ideass and practice. It is whether we can achieve the same level of statistical reliability. And the extent of the administrative burden to get there - I suspect automation may successfully address this.


However, I do wonder quite how many chess players are thought to be potential beneficiaries of this change. In Bucks about 100 people have played league games in the period up to Christmass and half of those have managed in excess of four games. Some of the players may also have played the odd county match, the county individual, or a single round of their club event. perhaps even the odd congress. It's not a lot of chess from which to garner an iniial grade. Is this demand for grades as soon as possible really just a metropolitan thing?

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

Re: ECF to scrap its grading system

Post by Roger de Coverly » Fri Dec 07, 2018 11:33 am

J T Melsom wrote:
Fri Dec 07, 2018 11:17 am
However, I do wonder quite how many chess players are thought to be potential beneficiaries of this change.
It's probably players new to competitive chess who will get their first grade or rating sooner. Players with A grades may have enough games to mean something, but experience with the monthly Elo ratings shows that they just fluctuate and in the case of juniors with a (K=40) modifier, the ratings are unstable.

You could solve the new player problem by just calculating and publishing their grades on a monthly basis. Some technical problems about how to make the new player estimation routine work would remain.

J T Melsom
Posts: 1294
Joined: Wed Aug 12, 2009 11:12 pm

Re: ECF to scrap its grading system

Post by J T Melsom » Fri Dec 07, 2018 11:40 am

I should have added that the new players in Bucks at least just aren't playing enough games, for statistical robustness. There are new players but not only are there only so many local games to go around, but its not easy for them to get selection simply because they are new and weaker. And the more interaction needed with event organisers outside your immediate club the harder to get selected.

John McKenna

Re: ECF to scrap its grading system

Post by John McKenna » Fri Dec 07, 2018 12:13 pm

Is this demand for grades as soon as possible really just a metropolitan thing?
Already answered (see below & above) -
The now deceased Yorkshire system managed to produce continuous grades...

J T Melsom
Posts: 1294
Joined: Wed Aug 12, 2009 11:12 pm

Re: ECF to scrap its grading system

Post by J T Melsom » Fri Dec 07, 2018 12:37 pm

As I've said I'm not wedded to the old system -just wanting to be convinced that there is sufficient benefit in these changes to outweigh the potential administrative burden associated with change. Not sure that quoting the existence of a system in Yorkshire where they had to do something because they had opted out of the national system is evidence of significant demand for more regular grading lists.

John McKenna

Re: ECF to scrap its grading system

Post by John McKenna » Fri Dec 07, 2018 1:00 pm

While I might agree that Yorks are unique they are not exactly 'metropolitan' and had no compunction to make their grading system 'continuous' (that was probably done to show it was superior to the ECF's, I believe that the Swiss Chess Fed. long had - not sure if it persists - a 'continuous' national.Elo system that updated in realtime as results were input.)

J T Melsom
Posts: 1294
Joined: Wed Aug 12, 2009 11:12 pm

Re: ECF to scrap its grading system

Post by J T Melsom » Fri Dec 07, 2018 1:20 pm

Fine, I'm just making the point that there are shire areas with no large urban centres within them, in which the organisation of chess and the demands made on organisers are quite different and that I get a sense that these differences are not always appreciated in the move to change. Plenty of areas of life where the policy makers have to rightly consider impact on various sectional groups, chess administrators shouldn't dismiss diversity either. That is not to say that those voices should be afforded priority, but they should be heard and given due weight. I've just had to answer another e-mail from a parent for whom there is no viable junior coaching locally, though if they lived a bit further away it might be a different story.