When is the new grading list coming out?

[Roger de Coverly]

The ECF grading system is (or perhaps I should say was until they started tweaking it last season) a relative zero-sum system measuring performance only, not an absolute one which the Fide system is with its different K factors, rating floors, results against unrated players not being counted, and cumulative results structure.

At the risk of anticipating some of EMW's comment, the ECF system is not as zero-sum as you might think. A player playing 30 games a season beats a player who plays 60 games. The "less active" player gains twice as much (one and two thirds) as the "more active" loses ( five sixths). This contrasts with Fide which is zero sum for the subset of established rated players on the same K factor.


Under some conditions, the ECF estimation routine now seems to converge to extreme values (219 performance against a 150 field for example). The speculation is that the lack of zero sum may be a reason.

[Geoff Chandler]

Well I started reading this thread, don't know why, these grading threads
are always the same.

However this differs;
Instead of a lonely voice howling in the wilderness about a lost 5 pts it does appear
that someone has cocked up the grades for the entire country.

The bad news you lot will playing Scottish and other European players.
You are going to spread these inflated grades around the world like a pandemic infection.

Every Chess player on Earth will be affected.

I have calculated that in a few years we will all have the same grade of 2525.
(Zager and Evans predicted this would happen).

Well Done.

To err is human, but to really screw things up you need a computer.

[Ola Winfridsson]

The ECF grading system is (or perhaps I should say was until they started tweaking it last season) a relative zero-sum system measuring performance only, not an absolute one which the Fide system is with its different K factors, rating floors, results against unrated players not being counted, and cumulative results structure.

At the risk of anticipating some of EMW's comment, the ECF system is not as zero-sum as you might think. A player playing 30 games a season beats a player who plays 60 games. The "less active" player gains twice as much (one and two thirds) as the "more active" loses ( five sixths). This contrasts with Fide which is zero sum for the subset of established rated players on the same K factor.


Under some conditions, the ECF estimation routine now seems to converge to extreme values (219 performance against a 150 field for example). The speculation is that the lack of zero sum may be a reason.

Yes, of course to a certain degree it's not zero-sum (also underrated juniors would tend have a deflationary effect), and I understand your point about the more inactive player's individual results will have a greater effect on his/her performance, but doesn't that go both ways? (losses are also compounded) In my experience, once you regularly play a certain number of games in a year (whether or not 30 is the right number is difficult to say, but in general I would have thought that another 10 or 20 games/year would be even "more" optimal), I'd say that the difference in number of games played by the individual players in question no longer becomes relevant - and the predictability of the outcome will remain relatively stable (and more or games will just serve to confirm the average performance).

Therefore (before the tweaking started) the ECF grade tended to be a more accurate reflection of a player's strength than the Fide system, since the ECF grade was more current and based on a sufficiently large number of games. Since there are so few Fide-rated events in this country (although Sean Hewitt does a sterling job of remedying it!) it can often take several years before a Fide-rated players reaches a stable rating.

[Brian Valentine]

It isa complete fluke if the system comes out as zero sum. It depends on the relative grades of leavers and joiners. Here is an analysis of last year's movements.

The 2008 list I refer to is the 2008 grades posted in 2009 or, if missing, the red grade from the 2008 list

2008 list: 10257 entries at an average of 133.8
leavers: 1589 at an average of 126.8
in both lists: 8668 at 135.0 in 2008 increasing to 135.4
joiners: 1489 at 122.5.

Giving an overall 2009 average of 133.5

The joiners are only so high because a significant proportion of them are overseas with a high grade. My guess is these go in at an estimate of their elo conversion (something I don't think has been discussed here). You only need the numbers or the averages of these leavers and joiners to change slightly to start observing material inflation or deflation. The real "currency change" is how and why the survivors increase by 0.4. Note that this in in the opposite direction to the overall average.

[John Upham]

Is GC Taylor (now 216) entitled to purchase an FM title and continue to play for Morley College in Division Three of the London League? That would be most amusing!

What are the criteria for obtaining an FM title?

[James Coleman]

You just have to reach 2300. It used to have to be for 24 games but now you can claim it as soon as your rating gets to 23.

[E Michael White]

I would like to see the BM title on offer again. It would give those lower than FM something to aim for. Of course it would need to be dumbed down a bit from the past; well why not they've done it with A'levels.

In the 50s it was thought that the top BMs were stronger than some IMs. What would your view on this be Mr Leonard Barden as a BM if you are reading this ? Would you mind it being reintroduced ?

[Roger de Coverly]

[quote

The joiners are only so high because a significant proportion of them are overseas with a high grade. My guess is these go in at an estimate of their elo conversion (something I don't think has been discussed here).

In fact as far as I'm aware they are trundled through the estimation routine in the same way as a new junior. So the "joiners" average is conditional on whether events like the "Liverpools" or the Isle of Man run and the success or otherwise of Hastings, Gibraltar and other international events in attracting new players to these shores. Without proof I would make the assertion that the recursion is sound when used, in effect, at an event level but something destabilises it at a system level. No-one takes much notice of ECF grades for Nakaruma etc., but they never seemed to stand out as obviously wrong.

You only need the numbers or the averages of these leavers and joiners to change slightly to start observing material inflation or deflation.

.

Last year there was a long thread about amongst other things the meaning and measurement of inflation. Most of the usual suspects participated. As part of that thread, the historic averages were unearthed and show remarkable stability (at around 115) from about 1994 onwards. At best you can only see a very slow downwards drift. The averages we really need are those before 1994. In particular there is debate in the late eighties about inflation and the average was quoted then to be about 125 to 130. So somewhere in the years where we have no recorded data, the average dropped dramatically. Was this the influx of new players in 1993 after KvS match? Was it the result of counties leaving the BCF when the BCF had a whip-round to pay for the 86 KvK match?

in both lists: 8668 at 135.0 in 2008 increasing to 135.4

Last year, I looked at 2008 actual v 2007 actual - same idea survivors only. I think the averages were almost identical. You would have expected to see systemic deflation if the deflation theorists were to be believed.

[E Michael White]

in both lists: 8668 at 135.0 in 2008 increasing to 135.4

Last year, I looked at 2008 actual v 2007 actual - same idea survivors only. I think the averages were almost identical. You would have expected to see systemic deflation if the deflation theorists were to be believed.

A few points occur to me:-

1. if you want to look at the average grade of survivors as some indicator of in/deflation, you should remove the games, from the calculations for survivors grades, played against leavers/joiners .
2. the 2008 figures will also have been calculated including games against those who have now left or treated as junior.
3. Its very difficult to remove starters and leavers without going back to raw games as some of the <30 games per year players will have some past games included from up to 3 seasons ago against those who have now left.
4. the figure which should stay the same, starters/leavers excluded ,is the weighted mean grade of survivors where the weights are the number of games played.

[Roger de Coverly]

Last year there was a long thread about amongst other things the meaning and measurement of inflation. Most of the usual suspects participated. As part of that thread, the historic averages were unearthed

Here's part of the debate with the historic analysis and the movements summary from 2007 to 2008.

http://www.ecforum.org.uk/viewtopic.php ... &start=180

[Roger de Coverly]

A few points occur to me:-

1. if you want to look at the average grade of survivors as some indicator of in/deflation, you should remove the games, from the calculations for survivors grades, played against leavers/joiners .
2. the 2008 figures will also have been calculated including games against those who have now left or treated as junior.
3. Its very difficult to remove starters and leavers without going back to raw games as some of the <30 games per year players will have some past games included from up to 3 seasons ago against those who have now left.
4. the figure which should stay the same, starters/leavers excluded ,is the weighted mean grade of survivors where the weights are the number of games played.

Both Brian and I are using an outcomes based approach. It's the published grades that people are interested in and we are looking to see whether there is untoward movement in the published grades. To this we define "in force throughout" as being someone on both the yyyy and yyyy-1 published lists. We define joiners as someone only on the yyyy published list and leavers as someone only on the yyyy-1 list. No other assumptions are needed. It may come down to what you mean by inflation or deflation. By my definition inflation (at one point a year) is where 20 years ago you got a grade of about 175 by scoring 50% in opens but today it's 195. Deflation (at one point a year) is where today you only get 155.

At the simplest level of observation the 2008 published grades are 20 points or more on average below the 2009 published grades. From that simple fact you can deduce that something major happened between 2008 and 2009.

[Brian Valentine]

I was trying to highlight the effect of leaversand joiners being an ongoing issue to control and probably more important than the points transfers from active versus relatively inactive (which may be a problem. I accept these bald statistics can be mis interpreted but I can't see a way of digging into this with the information to hand.

I'm sure Micky Adams is pretty unperturbed by his ECF grading. However he played 24 games last year. 11in the Staunton Memorial (mainly against Dutch), 10 at the EU championships in Liverpool and 3 in the final 4NCL weekend (thisreconciles but might be wrong -I am putting 2and 2 together). 2 of these were against Short who isi n asimilar position. If the elo conversion is not used (I have no reason to disbelieve Roger's assertion) then Mr Adams is hardly getting a reasonable grading. Although he has 24 games "counting" he is effectively being graded over perhaps 6 games. Other grandmasters may have a similar effect. The new overseas players grades may not appear misleading, but the situation is plausibly as unstable as the junior issue, albeit affecting fewer players.

[Richard Bates]

I was trying to highlight the effect of leaversand joiners being an ongoing issue to control and probably more important than the points transfers from active versus relatively inactive (which may be a problem. I accept these bald statistics can be mis interpreted but I can't see a way of digging into this with the information to hand.

I'm sure Micky Adams is pretty unperturbed by his ECF grading. However he played 24 games last year. 11in the Staunton Memorial (mainly against Dutch), 10 at the EU championships in Liverpool and 3 in the final 4NCL weekend (thisreconciles but might be wrong -I am putting 2and 2 together). 2 of these were against Short who isi n asimilar position. If the elo conversion is not used (I have no reason to disbelieve Roger's assertion) then Mr Adams is hardly getting a reasonable grading. Although he has 24 games "counting" he is effectively being graded over perhaps 6 games. Other grandmasters may have a similar effect. The new overseas players grades may not appear misleading, but the situation is plausibly as unstable as the junior issue, albeit affecting fewer players.

There is a rather major difference - gradings at the top end of the list are usually no more than an issue of personal pride for those who care about them. They don't affect ability to enter tournaments, and chances of winning prizes. There is no issue with "under-rated" juniors hoovering up prizes, or over-rated juniors becoming disillusioned with the game by being forced to play at a level too high for their ability.

[Leonard Barden]

I would like to see the BM title on offer again. It would give those lower than FM something to aim for. Of course it would need to be dumbed down a bit from the past; well why not they've done it with A'levels.

In the 50s it was thought that the top BMs were stronger than some IMs. What would your view on this be Mr Leonard Barden as a BM if you are reading this ? Would you mind it being reintroduced ?

The British Master title was worth around 2350, so it was harder to achieve than the current FM title.. You acquired it automatically after some appearances in the grading list at 213+ or 218+ (I forget which) the number of required appearances depending on how high up you were on the list. I guess Oliver and Jonathan Penrose, Clarke, Franklin, John Littlewood, Cafferty , myself and perhaps two or three others might be the survivors.

The BCF at various times also awarded CM (candidate master) and regional and county master titles. I think there were too many, and nowadays such awards would not be competitive with Fide titles.

[Roger de Coverly]

It would give those lower than FM something to aim for.

There's also the FIDE CM title - awarded at 2200. A problem with the prestige of both this title and FM is that FIDE have awarded the title on occasions in effect as prizes in Zonals and Olympiads. So there are some fairly low ranked players out there with these titles, one or two even in the lower reaches of the ECF grading list.