English Chess Forum

I've just taken a quick look at the new 4NCL ratings. Three significant factors contribute to causing a problem:

1. The ECF grade ----> FIDE rating conversion formula used is incredibly generous because x8 + 650 really doesn't work well with the 'new' grades.

2. Junior players with FIDE ratings may use ECF conversions instead of FIDE ratings.

3. The 'X' category of ECF grading uses only games played in the last 6 months, but junior players still get an increment because grades are supposedly 'out of date' due to improvement.

We have a junior (Jean-Luc Weller) in the White Rose 2 side.

My FIDE: 2191
Jean-Luc's FIDE: 2074
My ECF: 206
Jean-Luc's ECF: 204

So now because of the 80-point rule I have to play on a lower board than a player who I out-grade and out-rate (the latter by over a hundred points).

As it happens I'd been planning to play on a lower board than him anyway on Saturday but we have been forced to change our board order on other boards.

This rule really needs to be reconsidered (perhaps updating the ECF-FIDE conversion formula might suffice).

The rule was put in because of the result of a captain's meeting. It seemed to have become "generally accepted" that juniors are underrated - moreso than adults, anyway - and that their ECF grade was a better estimate of strength than their FIDE-rating. The concept seemed to win much support from the captains in attendance at the meeting.

_________________
February 16-17: 2013 BUCA Team Championship
April 20: 2013 British Blitz Championship
July 27-August 10: 100th British Chess Championships

Jonathan

Maybe worth your while pulling together a proposal for a future captains' meeting if you think it's an issue? We will probably be holding one at the last weekend (and 'probably' will become 'definitely' if we have some agenda items!). You could also consider canvassing (some of) the other captains beforehand to find out what they think?

As Alex says though, from recollection it was a captains' meeting that agreed the change in the first place.

The ECF conversion formula should revert to x8 +600. Shouldn't have changed in the first place.

The grading team used to be quite keen on doing various fits of the observed relationship between the ECF grade and the international rating. For example they came up with the rather silly (FIDE-1250)/5 = ECF for ratings below 2327 (source 2006 ECF diary). Provided they always keep an 8 in the divisor, a periodic best fit check in the x of (FIDE-x)/8 = ECF would be worth doing. I might suspect that for juniors, the x is going to be a different value from that for adults.

The Scots though prefer to use a 10 as divisor when converting between players who only have national ratings. Supporters of the grading revaluation exercise would suggest that the Scots need to do the same. namely to take 80% of the previous rating, but add something so that the top of the scale hasn't moved relative to the international Elos.


It needs more reverse engineering to establish the position for all activity categories, but has everyone noticed that the grade used for juniors in calculations is neither the brought forward grade at the start of the period, nor the carried forward grade at the end? So if you want a board order based on "current form", you could look at how the junior counted for their opposition.

The White Rose player mentioned earlier was 189 in the published August list, rising to 204 in the January list. His opponents were credited with a performance against a 193 player.

I agree. I won't name names, but there are people high up in the ECF who refuse to recognise and adopt the +650 formula. This is a source of confusion, not surprisingly.

Many junior have negative biased ECF grades ( players are underrated ) so no fixed FIDE conversion formula will work for all juniors. Juniors who play in the 4NCL are likely to be at the top end for underrating. If their games played in the junior 4NCL form a large part of their FIDE games the next rating for some will be even more negatively biased even if their ratings increase.
This smacks of one quarter up. I do hope no probate lawyers have become involved in grading.

The rule works very well. It is indeed optional - captains "opt in" to have their juniors given a fictitious Elo based on grading for board order purposes. Certainly the juniors are almost certain to be under-rated, and in fact the rule should be extended to players under 21 rather than just those under 18.

Beneficiaries so far include Marcus Harvey who beat Simon Williams on board one for Oxford and Sam Franklin who beat Chris Beaumont on board four for Barbican 2. Neither junior could even have played the game if they were stuck with their under-rating.

Suggestion to Jonathan Arnott: discuss this with Rupert. He was there at the captains' meeting last May after all - one assumes he was representing the team?

Whenever you have a large group of players who play relatively few FIDE rated games compared to ECF rated games, there are always going to be some anomalous results every now and then.



This is a very good point Jonathan. I think that introducing this new rule has probably helped to eradicate the problem of under-graded juniors perhaps with the effect of causing some of them to be over-graded.

It doesn't matter what rules we have for this, whilst there are few FIDE-rated tournaments and two different grading systems in use, there will always be some problems, it is just a question of trying to minimize these problems as much as possible.

_________________
http://www.yorkshirechess.org

Well done!
There it is in a nutshell - the 'curious quirk of the system' is that there are two systems.
That begs the question - why not have just one?
No, don't try to answer or another can of worms will open.

_________________
To find a for(u)m that accommodates the mess, that is the task of the artist now. (Samuel Beckett)

Yes, teams had a choice at the start of the season. But choices were made using the August grading list and ECF grades can dart about a lot more than FIDE ratings. So converted grades might be more appropriate for rounds 1-4 and ratings for 5-11. All the captains know before round 1 is what the current grades are.

I suppose my point is that the combination of:

1. Using the updated January list grades for conversions
2. Using the 'generous' conversion formula x8 + 650
3. Including the junior enhancement

is liable to over-correct a problem (under-rated juniors) which we all agree exists. I suspect that any TWO of these three points wouldn't be too bad, but the combination of all three is just too much and leads to over-rating.

I'll expand on the view that x8 + 650 is generous, and have a go at suggesting an appropriate change to the rules to correct but not over-correct, another time.