GRADING ANOMALIES

[Sean Hewitt]

What irritates me about all this "anomalies" talk is that no one, as far as I have seen, has shown average grades are going down. As far as I can discover, the average ECF/BCF grade remains roughly in the 114-120 range, where it has been for eons.
I know there is a lot of anecdotal evidence about people unhappy with their own declining grades, but if the average is the same ...?
Meanwhile, we know ECF grades are out of sync with Fide ratings, but bearing in mind the obvious inflation in the Fide system, this doesn't seem like evidence of anything being wrong with grades.

Try reading my paper, written and published in 2006. http://www.lrca.org.uk/lrca/Grading/Gra ... ction1.doc

I think you'll find that provides the proof, the solution, and the fix. It IS broke, and it does need fixing!!

It's taken two years to do something about it but, to be fair, the solution to be implemented is far more accurate than the one I proposed. I confess I hadnt envisaged the will to implement something as radical and as much work as they are doing, but hats off to them for doing so.

To answer the question about the new system - the system (+50 for a win ; -50 for a loss) isnt being changed. You will get a new grade (almost certainly increased, and in general it will increase more the lower graded you are) but then the calculation going forward will still use the +50 ; -50 formula.

[Ian Kingston]

Sean: do you know what's being done to prevent deflation in the future?

[IM Jack Rudd]

Sean: do you know what's being done to prevent deflation in the future?

One of the things David Welch mentioned was that juniors would have their new grades worked out first, and adults' games against juniors would be graded against the juniors' new grades, not their old ones.

[Howard Grist]

Could anyone quickly summarize for me what the new grading rules will be?

The new grading rules are the same as the old ones. There are likely to be some changes to the junior enhancements though.

Also on a personal level, I have been tracking my grade this year on the old formula. It should increase this year by 21 or 22 points. Under the new formula, is this likely to be the same, or higher, or lower, can anyone tell me?

A rough conversion guide for the corrected grade is old grade x 0.8 + 50, so a 21 point increase under the current system will translate to a 17 point increase under the new one.

Btw, I have beaten a few players with grades more than 50 points below me. Relative to my current grade, this wins bolster my incoming grade; seen only as a part of my incoming grade, these wins actually lower it! Is that what this is all about?

Not really, no. The 40 point rule is rather generous towards the higher graded player, but I'm not convinced that this is actually worth changing, although I am investigating it.

What irritates me about all this "anomalies" talk is that no one, as far as I have seen, has shown average grades are going down. As far as I can discover, the average ECF/BCF grade remains roughly in the 114-120 range, where it has been for eons.

This is true if you look at standard play grades, but not if you look at rapid play grades. The July 2007 figures are 111.7 for standard and 73.13 for rapid. The 1996 figures are 113.5 and 109.7 respectively. It is difficult making year on year comparisons like this as there are more juniors included now than there used to be.

do you know what's being done to prevent deflation in the future?

Hopefully it will go away, but we'll certanly be keeping an eye open for it.

One of the things David Welch mentioned was that juniors would have their new grades worked out first, and adults' games against juniors would be graded against the juniors' new grades, not their old ones.

David Welch may have mentioned this but it would mean that you wouldn't be able to calculate your new grade if you'd played a junior. There are no plans to incorporate this in to the grading calculations.

Howard
ECF Grading System Programmer

[Dr Andrew Cula]

A rough conversion guide for the corrected grade is old grade x 0.8 + 50

Do we use the existing BCF to FIDE formulas, or will this need tweaking?

[TomChivers]

Thanks Howard. I'm afraid I'm still confused about one thing...

A rough conversion guide for the corrected grade is old grade x 0.8 + 50, so a 21 point increase under the current system will translate to a 17 point increase under the new one.

My current grade is 163. Next year it under the old system ought be 184 or 185 (I played one ungraded player.) So instead it will be 180 next year ie in the 2008 list, just a multiplication of my old 2007 grade by the formula, ignoring all the games I've played this year?

[Dr Andrew Cula]

Thanks Howard. I'm afraid I'm still confused about one thing...

A rough conversion guide for the corrected grade is old grade x 0.8 + 50, so a 21 point increase under the current system will translate to a 17 point increase under the new one.

My current grade is 163. Next year it under the old system ought be 184 or 185 (I played one ungraded player.) So instead it will be 180 next year ie in the 2008 list, just a multiplication of my old 2007 grade by the formula, ignoring all the games I've played this year?

you will be 197/198

[Howard Grist]

Do we use the existing BCF to FIDE formulas, or will this need tweaking?

This will change to something like FIDE = 7 x ECF + 800. Theoretically we ought to be able to go back to the 'traditional' formula of FIDE = 8 x ECF + 600, but we won't be able to as the FIDE system has a similar problem to the ECF system, although it's not quite so severe.

My current grade is 163. Next year it under the old system ought be 184 or 185 (I played one ungraded player.) So instead it will be 180 next year ie in the 2008 list, just a multiplication of my old 2007 grade by the formula, ignoring all the games I've played this year?

Tom,

Next month the new (i.e. 2008/9) grades will be published. This will be calculated using the old system and so for you will be 184/5. This will be your official grade for the coming season. Alongside this, there will be another standard play grade published for you. This will be calculated from scratch along the lines described by Roger de Coverly further up this page. In your case this would be in the region of 197. In July 2009, the current grades will be dropped and only the newly calculated grades will be published. Both sets of grades are being published this year so that organisers can see how the new grading ranges look and amend their events accordingly and also so that you can calculate what your 2009/10 grade will be.

[TomChivers]

Thanks Andrew & Howard - I think I've got it now!

[Tim Spanton]

What irritates me about all this "anomalies" talk is that no one, as far as I have seen, has shown average grades are going down. As far as I can discover, the average ECF/BCF grade remains roughly in the 114-120 range, where it has been for eons.

This is true if you look at standard play grades, but not if you look at rapid play grades. The July 2007 figures are 111.7 for standard and 73.13 for rapid. The 1996 figures are 113.5 and 109.7 respectively. It is difficult making year on year comparisons like this as there are more juniors included now than there used to be.

Thanks Howard. This confirms what I thought. So there may be a case for playing around with rapid grades, but if standard grades aren't suffering from deflation (and if the average is roughly the same, they can't be) I still reckon there is absolutely no case for inflating them.

[Sean Hewitt]

...if standard grades aren't suffering from deflation (and if the average is roughly the same, they can't be) I still reckon there is absolutely no case for inflating them.

Tim - Standard play grades ARE suffering from deflation. Just because the mean average is approximately constant proves absolutely nothing - either way. Even if you dont understand the maths, the graphs in my paper are fairly intuitive.

[Tim Spanton]

...if standard grades aren't suffering from deflation (and if the average is roughly the same, they can't be) I still reckon there is absolutely no case for inflating them.

Just because the mean average is approximately constant proves absolutely nothing

!

[Sean Hewitt]

...if standard grades aren't suffering from deflation (and if the average is roughly the same, they can't be) I still reckon there is absolutely no case for inflating them.

Just because the mean average is approximately constant proves absolutely nothing

!

Tim, I dont know where you got your maths degree from, but anyone with one will tell you that I am right!

[John Upham]

Whilst preparing league results databases for export to the Rating Officers a thought occurred...

Some of new junior players have Date of Birth information to go to the ECF and some don't

AFAIK, those with will benefit from the junior rating bonus and those without won't

So, hard luck on those who have lost to improving juniors w/o date of birth information as they will score 10 less for that game

Now, THAT IS RANDOM!

John

[John Upham]

Perhaps a stupid question but does anyone have an estimate for how many players are in the ECF rating database who are 18 years of age or less but their Date of Birth is not known?

I believe I'm right that unless the DoB is known they will always be calculated as adults with no junior bonus?

What impact will this have on deflation?

John