Page 1 of 18

Yorkshire vs ECF disparity

Posted: Tue Oct 30, 2012 5:06 pm
by MartinCarpenter
Something intriguing from the Yorkshire grading list main page:

"A Yorkshire grade should indicate the same playing strength as an equal ECF grade. The 2010 Yorkshire grades were adjusted to match the national scale. A study comparing 2011 grades found little divergeance, but in 2012 the standardplay scales had diverged by about 1.7 points. To restore parity, Yorkshire standardplay grades will be adjusted down by 0.1 points per month during the year 2012/13."

A summary of the major differences:
1) Yorkshire grades are calculated 'live' with a weight that drops over time then increments after a game is played,
2) There is thus no junior increment of any form (its not needed),
3) The +-40 pt rule is iirc currently +-45 instead.

The basic calculation is ECF. Not obvious how these might cause this though. The current best guess seems to be that its random fluctuations, although the gap does seem to be quite large for that.

Any ideas?

Re: Yorkshire vs ECF disparity

Posted: Tue Oct 30, 2012 5:34 pm
by Sean Hewitt
MartinCarpenter wrote:A summary of the major differences:
1) Yorkshire grades are calculated 'live' with a weight that drops over time then increments after a game is played,
2) There is thus no junior increment of any form (its not needed),
3) The +-40 pt rule is iirc currently +-45 instead.
(2) makes no difference. The ECF grading system has no junior increment in it's calculations.
Need to know more about (1) to understand how it might affect things. (3) will affect things though you would need to look at the ratio of games played between players graded more than 40/45 points apart to understand by how much.

It's also worth asking how they measured the divergence. i.e. what are they comparing?

Re: Yorkshire vs ECF disparity

Posted: Tue Oct 30, 2012 5:40 pm
by Rob Thompson
Could it be simply that people may tend to play slightly better when relatively close to home (Yorkshire games) than travelling a longer way (games outside of Yorkshire)?

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 7:52 am
by Kevin Thurlow
"Could it be simply that people may tend to play slightly better when relatively close to home (Yorkshire games) than travelling a longer way (games outside of Yorkshire)?"

A very sensible point. Although judging by the way I played a mile from home on Monday, I had better not go any further away.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 10:19 am
by MartinCarpenter
Are you sure about the increment? There's definetly some non trivial systematic difference regarding juniors because their ECF grades are routinely ~8 pts higher than their Yorkshire ones. (Which actually seem slightly more accurate to me in this case, not that I have any strong feelings about the two systems overall.). This effect is pointing the wrong way of course.

All that I can see 1 doing is producing (short term) randomness at the start of seasons and allowing for live updates. More details here - http://www.chessnuts.org.uk/ny5/Rulesof ... ading.html .

Quite which group of players Jon checks for the discrepancies I'm not totally sure. Maybe it could be travel, but some of the people graded do travel into Yorkshire to play and quite often a fair distance within it too. Makes as much sense as anything else though.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 10:36 am
by Roger de Coverly
MartinCarpenter wrote:Are you sure about the increment? There's definetly some non trivial systematic difference regarding juniors because their ECF grades are routinely ~8 pts higher than their Yorkshire ones.
The ECF's published grades for juniors, whilst based on results for previous seasons, are not used as the estimate of strength for the new season. Those published do however have an age related supplement. The only effect of this is on eligibility for graded restricted competitions, board orders and seeding orders for tournaments.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 11:29 am
by MartinCarpenter
I knew they were doing some odd things with junior grades but that particular option hadn't occured to me :) Thanks.

It was actually a very mild worry cf the two grading systems because of course it would have been acting as a non trivial deficit for Yorkshire grades vs ECF ones, while the difference was in the other direction. If its only the published junior ones affected it'd be a rather small effect. (or even none at all depending on the populations used to compare the two grading systems.).

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 2:06 pm
by Michele Clack
Does the ECF get a fee for its grading system being used in this way? If not why not?

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 2:22 pm
by IM Jack Rudd
michele clack wrote:Does the ECF get a fee for its grading system being used in this way? If not why not?
Why would we be paid a fee? The ECF grading system isn't patented; it may well not even be patentable.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 2:33 pm
by Sean Hewitt
michele clack wrote:Does the ECF get a fee for its grading system being used in this way? If not why not?
The ECF grading system is a mathematical formula (and as such cannot be patented or copyrighted). The data (i.e. the grades themselves) is copyrightable.

Which means you can set up your own grading system if you wish, but you can't use ECF grades to get you started.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 3:27 pm
by Ian Kingston
Sean Hewitt wrote:
michele clack wrote:Does the ECF get a fee for its grading system being used in this way? If not why not?
The ECF grading system is a mathematical formula (and as such cannot be patented or copyrighted). The data (i.e. the grades themselves) is copyrightable.

Which means you can set up your own grading system if you wish, but you can't use ECF grades to get you started.
I'm not sure you're right about the copyright issue. Copyright protects only the physical expression of an idea, and does not apply to a collection of facts (which is what the grades are). It does cover the program(s) used to generate the grades and the published form of the grading list. The UK Copyright Service describes the basics of UK copyright law. It's not clear to me where the grading list would fit into this framework, but I'm not a lawyer. Mere data does not, on the face of it, seem to be covered.

Cases in the USA and Australia have found that telephone directories, for example, are not protected by copyright. There is currently a case in the UK regarding football fixtures: the European Court of Justice says they're not copyrightable; the Premier League and others say that they are.

The final outcome of that case might well affect the status of the grading list, but that's one for the lawyers to argue over!

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 3:38 pm
by Roger de Coverly
Ian Kingston wrote:I'm not sure you're right about the copyright issue. Copyright protects only the physical expression of an idea, and does not apply to a collection of facts (which is what the grades are).
From getting on for ten years ago.

http://www.sccu.ndo.co.uk/0203/sccu.htm
(9) Grading: copyright charge. The BCF, as you will know, has announced a copyright charge for local grading lists using BCF data. Local reactions have not been favourable. The SCCU will put a motion on the agenda of the October Council meeting, to the effect that this charge be dropped. The charge will undoubtedly be resented, but the Union's principal reason is that it sounds almost impossible to police.
This dates back to when the grading list was only published in paper form.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 3:43 pm
by Sean Hewitt
Ian Kingston wrote:I'm not sure you're right about the copyright issue. Copyright protects only the physical expression of an idea, and does not apply to a collection of facts (which is what the grades are). It does cover the program(s) used to generate the grades and the published form of the grading list. The UK Copyright Service describes the basics of UK copyright law. It's not clear to me where the grading list would fit into this framework, but I'm not a lawyer. Mere data does not, on the face of it, seem to be covered.
Data was a poor choice of words on my behalf. I should have simply said the grades are copyrightable because they are not factual (in the way that a telephonne number is) but are the product of the ECF's work. At least, that's what my legal adviser said when we looked into it.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 4:01 pm
by Ian Kingston
Sean Hewitt wrote:
Ian Kingston wrote:I'm not sure you're right about the copyright issue. Copyright protects only the physical expression of an idea, and does not apply to a collection of facts (which is what the grades are). It does cover the program(s) used to generate the grades and the published form of the grading list. The UK Copyright Service describes the basics of UK copyright law. It's not clear to me where the grading list would fit into this framework, but I'm not a lawyer. Mere data does not, on the face of it, seem to be covered.
Data was a poor choice of words on my behalf. I should have simply said the grades are copyrightable because they are not factual (in the way that a telephonne number is) but are the product of the ECF's work. At least, that's what my legal adviser said when we looked into it.
That kind of issue is what the football fixtures case is about. The ECJ says 'A football fixture list cannot be protected by copyright when its compilation is dictated by rules or constraints which leave no room for creative freedom'. To me that would imply that ECF grades can't be copyrighted.

I think it's a slightly grey area. I doubt that the ECF would have the funds to pursue a copyright infringement case anyway.

Re: Yorkshire vs ECF disparity

Posted: Wed Oct 31, 2012 7:12 pm
by Michele Clack
Perhaps not, but it doesn't seem to be quite cricket for Yorkshire to piggyback their system onto the ECF system. Given that this is Yorkshire I'm shocked!