New ECF grades are now online.
-
- Posts: 8839
- Joined: Fri Aug 06, 2010 2:34 am
- Location: London
Re: New ECF grades are now online.
I was a bit disappointed to find that there weren't three pages of people talking about the new grades (this thread was started in July 2011). Still, having two threads on this can't be a bad thing...
-
- Posts: 3
- Joined: Sat Jan 26, 2013 2:12 pm
Re: New ECF grades are now online.
Another point is to allow people to measure performances.Richard Bates wrote:
What is the point of any grade if not for seeding purposes?
Understood, although I'd then expect the difference between their new grade and that used in the calculation to be the same as the bonus (which isn't the case)Richard Bates wrote:
Junior grades reflect past performance plus anticipated future improvement. For the purposes of grading their opponents in the next period their ACTUAL performance then replaces this grade (so no bonus is necessary).
-
- Posts: 3340
- Joined: Fri Nov 14, 2008 8:27 pm
Re: New ECF grades are now online.
It looks as if his performance (over 26 games) was 178. How his grade is worked out i don't know. Maybe it is only based on his last 20 games?Tim Seymour wrote:Understood, although I'd then expect the difference between their new grade and that used in the calculation to be the same as the bonus (which isn't the case)Richard Bates wrote:
Junior grades reflect past performance plus anticipated future improvement. For the purposes of grading their opponents in the next period their ACTUAL performance then replaces this grade (so no bonus is necessary).
-
- Posts: 3562
- Joined: Wed Jul 02, 2008 4:31 pm
- Location: Awbridge, Hampshire
Re: New ECF grades are now online.
In reverse order of what then?Angus French wrote:It also appears that, to make up a quota, results are taken from previous grading periods in reverse order (this doesn't necessarily mean that the most recent results are used!).
-
- Posts: 2153
- Joined: Thu May 15, 2008 1:37 am
Re: New ECF grades are now online.
I meant that if there are insufficient results in a grading period to achieve a grade category then results are taken from the immediately preceding period and then, if there are insufficient in that second period to achieve a grade category, results are taken from a third grading period – the one immediately preceding the second - and so on until either a grade category is achieved or it’s established that it’s not possible to achieve an ‘E’ grade category (where the ‘buck stops’).Ian Thompson wrote:In reverse order of what then?Angus French wrote:It also appears that, to make up a quota, results are taken from previous grading periods in reverse order (this doesn't necessarily mean that the most recent results are used!).
That’s my understanding of how the thing works and it might seem fine but I believe it could still throw up something unexpected.
For illustration:
10 results were submitted for January 2013. This isn’t enough for an ‘X’ grade so we try for an ‘A’ grade. However, only 15 results were submitted for July 2012. We don’t have enough for an ‘A’ so we try for a ‘B’ grade. Since another 25 results were submitted for January 2012 we now have 50 results and that’s more than sufficient for a ‘B’ grade. We make the ‘B’ grade by taking the taking the 25 results from the two most recent submissions and then we add the 5 most recent results from the third submission.
This may all sound fine. But, what if the January 2012 submission included, say, 9 results from the 2011 London Chess Classic, played in December 2011, and the July 2012 submission included, say, 5 results from the London League which were played in October and November 2011? (Last season, the London League submitted all of its results in one go, for July 2012 grading, and none for Jan 2012 grading.) We’ve ended up with a situation where 5 later results have been discarded in favour of five earlier results. Probably the situation is a one-off (since nearly all leagues, and the London League in particular, now submit half year results in January) and probably not many grades are affected.
-
- Posts: 21322
- Joined: Tue Apr 15, 2008 2:51 pm
Re: New ECF grades are now online.
I think the grading team look for false accuracy in counting back results by date. In my opinion it would be better to just average the results from the previous period. So if you consistently play 20 games each half year, your grade is always 2/3rd of the most recent six months plus 1/3rd of the previous six months. This avoid the problem of having results at the end of a grading period perpetually double counted. That's for adult players. For Juniors and anyone "rapidly improving" (as opposed to "rapidly fluctuating"), the count back method has more merit.Angus French wrote: We’ve ended up with a situation where 5 later results have been discarded in favour of five earlier results. Probably the situation is a one-off (since nearly all leagues and the London League in particular now submit half year results in January) and probably not many grades are affected.
-
- Posts: 8839
- Joined: Fri Aug 06, 2010 2:34 am
- Location: London
Re: New ECF grades are now online.
I fully agree. I had a moderately bad result at the London Chess Classic FIDE Open, which falls at the end of the grading period used for the January 2013 list, so I want to avoid that being counted again in the next calculation!Roger de Coverly wrote:I think the grading team look for false accuracy in counting back results by date. In my opinion it would be better to just average the results from the previous period. So if you consistently play 20 games each half year, your grade is always 2/3rd of the most recent six months plus 1/3rd of the previous six months. This avoid the problem of having results at the end of a grading period perpetually double counted. That's for adult players. For Juniors and anyone "rapidly improving" (as opposed to "rapidly fluctuating"), the count back method has more merit.Angus French wrote: We’ve ended up with a situation where 5 later results have been discarded in favour of five earlier results. Probably the situation is a one-off (since nearly all leagues and the London League in particular now submit half year results in January) and probably not many grades are affected.
Is there a way to find out which of my 44 games from that period would be used in any countback if I for some reason played less than 30 games for this period? Admittedly, that is unlikely, having played 6 in this period already, and planning to play another 23 over the next few months, without even taking into consideration the odd congress here and there, but I'd like to have some idea of where things stand.
The obvious answer, I suppose, is to go by the dates on the list (44 games):
http://www.ecfgrading.org.uk/?ref=15528 ... =543946487
I would need to arrange those results chronologically by date and treat games played at a congress as a block, but spread the league games around in chronological order? So there, if I play zero games in the following period, my new grade would be calculated by taking the last 30 games and then continuing to count back until the start of the event in which that 30th game fell (it currently falls in the Hinckley Congress)? So as I've played 6 games, the countback is only 24, and so on? But as I play more, and approach 19 games, it will be the LCC FIDE games counting in the countback, plus the three games played after that and before the end of the grading period.
CORRECTION: 30th game countback falls after the Hinckley Congress, I mistakenly counted the Kazakhstan block as a single game when that is eight games. Still, the simple calculation is that playing 29 games might be a bad idea, as countback would then bring in my loss to a 143, the last game I played in the preceding period...
It's almost enough to make one give up trying to calculate such things. Hmm, that might be the point!
-
- Posts: 89
- Joined: Wed Jul 07, 2010 4:13 pm
- Location: Surrey
Re: New ECF grades are now online.
A revision of the January 2013 list has gone up today (4th Feb) "following the discovery of numerous incorrect grades." And indeed, a glance at the amended grades section reveals no less than 2,099 (!) grades have changed. A very large proportion of these are junior players, which leads me to wonder if there was a slip-up with the junior calculation process somewhere?
-
- Posts: 21322
- Joined: Tue Apr 15, 2008 2:51 pm
Re: New ECF grades are now online.
Published junior grades are works of fiction anyway, since they aren't used in the calculation for the next period.Daniel Young wrote:A very large proportion of these are junior players, which leads me to wonder if there was a slip-up with the junior calculation process somewhere?
-
- Posts: 2193
- Joined: Sat Mar 10, 2012 8:18 pm
Re: New ECF grades are now online.
Two things have happened. Firstly, Cumbria have submitted their league results for grading and these have now been included resulting in changed grades for most Cumbrians. But far and away the biggest impact was the discovery that one junior player had a rogue grade allocated to her in the grading system. This affected her grade, that of every player she played, and every player who played someone she had played, and so on and so on.Daniel Young wrote:A revision of the January 2013 list has gone up today (4th Feb) "following the discovery of numerous incorrect grades." And indeed, a glance at the amended grades section reveals no less than 2,099 (!) grades have changed. A very large proportion of these are junior players, which leads me to wonder if there was a slip-up with the junior calculation process somewhere?
Ultimately, 2000 grades were affected; most, I'm told, by 1 point.
-
- Posts: 89
- Joined: Wed Jul 07, 2010 4:13 pm
- Location: Surrey
Re: New ECF grades are now online.
Thanks Sean. Yes, on the basis of my old school's grading entry (where 18 of 38 grades moved), almost all the changes are upwards by one point. Amazing how one player can cause a chain reaction like that! But good to see that it's fixed so quickly anyway.Sean Hewitt wrote:Two things have happened. Firstly, Cumbria have submitted their league results for grading and these have now been included resulting in changed grades for most Cumbrians. But far and away the biggest impact was the discovery that one junior player had a rogue grade allocated to her in the grading system. This affected her grade, that of every player she played, and every player who played someone she had played, and so on and so on.Daniel Young wrote:A revision of the January 2013 list has gone up today (4th Feb) "following the discovery of numerous incorrect grades." And indeed, a glance at the amended grades section reveals no less than 2,099 (!) grades have changed. A very large proportion of these are junior players, which leads me to wonder if there was a slip-up with the junior calculation process somewhere?
Ultimately, 2000 grades were affected; most, I'm told, by 1 point.
-
- Posts: 427
- Joined: Sun Jul 06, 2008 4:56 pm
Re: New ECF grades are now online.
The odd junior here is up by as much as 5 rapidplay.
I'd've preferred ECF to wait until the July list: these changes make it difficult for captains and organisers working in grading-limited Leagues and other competitions.
PB
I'd've preferred ECF to wait until the July list: these changes make it difficult for captains and organisers working in grading-limited Leagues and other competitions.
PB
-
- Posts: 1420
- Joined: Fri Jun 01, 2007 6:31 pm
Re: New ECF grades are now online.
When 1 junior grade is altered it is quite likely that every grade of every player, who played any junior, will require checking. In addition grades of many adult new starters would also need checking and thus grades of adults and juniors who played those adults etc. In many cases the revisions would be lost in rounding, being too small to change the published grade. All this of course assuming the system still works the way I think it does.
Paul Buswell mentions a 5 point increase; it is possible for a small change in a junior or adult grade to cause substantial changes, even 10 - 15 points in several thousand grades.
Paul Buswell mentions a 5 point increase; it is possible for a small change in a junior or adult grade to cause substantial changes, even 10 - 15 points in several thousand grades.
-
- Posts: 21322
- Joined: Tue Apr 15, 2008 2:51 pm
Re: New ECF grades are now online.
There's a statement on the ECF site on what happened. With a recalculation system as used by the ECF, you can always get relative grades, but with a completely closed set, you will not get absolute ones. They appear to have assigned one very active player an arbitrary grade as this unlocked grades for many other players. Obviously this arbitrary grade should have been a one off, but they continued to use it as it became increasingly incorrect as an estimate of the strength of this player. The widespread nature of the effects is an example, I would think, of perturbation theory. That is an approach which attempts to gain insights into complex models by changing one parameter and reviewing the effects. It was suspected when the ECF announced that it was treating all juniors as new players, that this could make the absolute level of their grades conditional on the performance of a handful of players, this rather confirms it empirically.E Michael White wrote:When 1 junior grade is altered it is quite likely that every grade of every player, who played any junior, will require revision.
-
- Posts: 577
- Joined: Fri Apr 03, 2009 1:30 pm
Re: New ECF grades are now online.
It is a welcome improvement that the graders have explained what they are doing.
It remains striking that they have an arbitrary arrangement for the treatment of strong foreign players rather than a fixed rule.
One would hope that the situation that has emerged for juniors will encourage them to check the stability of the system when they tinker with junior grades albeit for the greater good of grading more players.
It remains striking that they have an arbitrary arrangement for the treatment of strong foreign players rather than a fixed rule.
One would hope that the situation that has emerged for juniors will encourage them to check the stability of the system when they tinker with junior grades albeit for the greater good of grading more players.