Reluctance to Align with FIDE ELO?

John Upham · Post by **John Upham** » Wed Apr 04, 2007 10:52 am

I'd like to be made aware of the rationale behind the ECF of not aligning its national rating system with FIDE ELO.

I know its the English way of doing things to stick two fingers up at "Johnnie Foreigner" as we always know best, but, apart from this, are there sensible arguments to consider?

John

John Upham · Post by **John Upham** » Wed Apr 04, 2007 6:21 pm

Martin,
I guess the ECF could always canvass its membership and convert any "feelings" into real opinion.

One thing we could debate (to slow things down) would be what *is* the membership and the only real answer could be the Direct Members.

Maybe we could poll the 4NCL players first and leave the people who only play local league chess (and probably never consider such lofty matters) until last. If the 4NCL players don't want it then nobody will. However, I think I can guess their view...

John

Ian Jamieson · Post by **Ian Jamieson** » Thu Apr 05, 2007 10:27 am

John,

When you say align do you mean changing the scale and location of ECF grades i.e. 250 to 2600, 3 digits to 4 digits, and / or the method of calculating ECF grades?

Ian

Greg Breed · Post by **Greg Breed** » Thu Apr 05, 2007 11:01 am

The thing with the ECF grading system is that is fairly simple to calculate.
A Win is opponents grade +50.
A Loss is opponents grade -50.
A Draw is your opponents grade.
The only stipulation is the +/- 40 rule so that no-one can benefit from losing or deficit from winning.

The ELO system is complex and involves formulas and coefficients and stuff. Urgh.

At least with ours anyone can work out their performance over a number of games.

Chris Majer · Post by **Chris Majer** » Sat Apr 07, 2007 12:04 am

I'd like to be made aware of the rationale behind the ECF of not aligning its national rating system with FIDE ELO.

If by this you mean moving to an ELO methodology for calculating grades, then it only became technically feasible (at least in principle) about 18 months ago. (When we got rid of an obsolete method of submitting grading information). Now the primary challenge in changing would be to ensure that all the graders and all the leagues/congresses were bought in to any change and that Council approved it.

Changing to an Elo methodology is a possible outcome from the ongoing investigation into grading anomalies. Otherwise, I would question whether the benefit of a change is worth the very considerable disruption.

Regards

John Upham · Post by **John Upham** » Sat Apr 07, 2007 12:21 pm

Chris,
I mean not only the methodology but also using ELO type ratings day-to-day and phasing out the old style gradings.

We could also allow authenticated result submission for more frequent rating updates.

Once a year really demonstrates we've got our finger on the pulse showing the rest of the world how to stay in the past.

Perhaps we should outsource some foreign IT consultants to show us how it could be done?

John

Nigel Wright · Post by **Nigel Wright** » Tue Apr 10, 2007 9:27 pm

How about giving people both types of grade and let them choose which they prefer? I like the ECF one personally, and would always use it in tournaments. However, I have many online friends, and it is rather annoying not being able to compare grades accurately with them. If it is too much effort to convert everyone's grades, why not provide a step-by-step instruction guide on how to convert it ourselves?

Chris Majer · Post by **Chris Majer** » Tue Apr 10, 2007 11:32 pm

How about giving people both types of grade and let them choose which they prefer? I like the ECF one personally, and would always use it in tournaments. However, I have many online friends, and it is rather annoying not being able to compare grades accurately with them. If it is too much effort to convert everyone's grades, why not provide a step-by-step instruction guide on how to convert it ourselves?

ECF to ELO conversations (and much else) can be found at http://grading.bcfservices.org.uk/help.php

However, I don't believe that accurate conversations between two different systems exist. USCF Elo ratings are traditionally claimed to be about 100 points higher than FIDE Elo

Greg Breed · Post by **Greg Breed** » Thu Apr 12, 2007 4:32 pm

For those wanting to compare their rating to others online:
ECF * 8 + 600 = ELO
ECF * 8 + 700 = USCF
This is fairly accurate, however I think most online grades are artificially high.

I've read the entire 7 pages of discussion at the Atticus Forum and I'm impressed with the participation and ideas there. The one thing that didn't get discussed enough was the effect of using the ELO system. One of the members clarified it:
We can use the ELO numbers system but not the ELO calculation system as that requires each game to be graded after it is played so that it has an effect upon the next game. Sure that works in Tournaments once a month but not in a League.

I for one would instantly resign as Grader if I had to submit everyone's grade as soon as I got the result. There isn't enough time in the day for me and I'll bet I'm not the only one. Running a League is time consuming enough! - And I was the only person willing to stand-in as controller and grader (two separate posts)!

I might be persuaded to submit the results twice a year but no more.
I wouldn't be fussed if someone could come up with a way of using ELO numbers and the ECF system.

Mike Gunn · Post by **Mike Gunn** » Fri Apr 13, 2007 4:13 pm

The last one is easy: just add 400 for a win instead of 50. (And you'd need some starting values, of course.) But the argument as to whether the ECF or Elo systens are "working properly" still remains, of course.

LeamingtonGrader · Post by **LeamingtonGrader** » Thu May 03, 2007 2:20 pm

As a grader, I fully understand the reluctance to change both the system and the way in which grading is carried out.

For anyone who is a league organiser / fixture secretary and a grader, a change to an ELO type system would be difficult to put it mildly.

Would web based systems where league results are input and validated be a step in the right direction. If this data could be made available to a grading system it may be possible to contemplate the direct and immediate grading of league games.

I know that there are a whole host of logistical problems to overcome and probably an extensive software development and data processing task to undertake.

How do the Elo based chess federations tackle these problems?

Sean Hewitt · Post by **Sean Hewitt** » Thu May 03, 2007 4:25 pm

LeamingtonGrader wrote:For anyone who is a league organiser / fixture secretary and a grader, a change to an ELO type system would be difficult to put it mildly.

Why do you think that?

I am a grader also and as far as I can see our work wouldnt change. We collect results and send them to the ECF in the prescribed format. Then the ECF do the calculations. Our bit would remain unchanged. It is their calcualtion that would be different.

In all honesty, all we do is data input!

Carl Hibbard · Post by **Carl Hibbard** » Thu May 03, 2007 8:31 pm

LeamingtonGrader wrote:Would web based systems where league results are input and validated be a step in the right direction. If this data could be made available to a grading system it may be possible to contemplate the direct and immediate grading of league games.

Funnily enough I expect a discussion about this at the Leamington AGM (a league in which we both play...) with the fixtures secretary stepping down, perhaps that may come to something but it depends if I have the time to be honest!

Cheers
Carl

Greg Breed · Post by **Greg Breed** » Fri May 04, 2007 10:09 am

Sean Hewitt wrote:
LeamingtonGrader wrote:For anyone who is a league organiser / fixture secretary and a grader, a change to an ELO type system would be difficult to put it mildly.
Why do you think that?

I am a grader also and as far as I can see our work wouldnt change. We collect results and send them to the ECF in the prescribed format. Then the ECF do the calculations. Our bit would remain unchanged. It is their calcualtion that would be different.

In all honesty, all we do is data input!

If it were that simple then our job wouldn't change, but unfortunately it isn't. If we were to use the ELO numbers but still calculate on an annual or bi-annual basis then, as you say, our job would remain unchanged.

If the ELO system is used though, the difference would be that results would have to be sent in immediately to have a bearing on the following matches. In a local league such as mine (and many others I suspect) there are many adjournments and postponements. It's better and easier to just send all the results in at the end of the season.

Sean Hewitt · Post by **Sean Hewitt** » Fri May 04, 2007 4:44 pm

Greg Breed wrote: If the ELO system is used though, the difference would be that results would have to be sent in immediately to have a bearing on the following matches. In a local league such as mine (and many others I suspect) there are many adjournments and postponements. It's better and easier to just send all the results in at the end of the season.

Greg, you are confusing the ELO system with frequency of publication. Changing to ELO would not change the work we do whatsover. Changing to more frequent publication of grading lists (whether the calculations are done using ECF or ELO methodology) would mean graders having to be more punctual in the submission of results.

And whilst I agree it might be easier for the grader if he sends all the results in at the end of the season, it certainly isnt better!!

English Chess Forum

Reluctance to Align with FIDE ELO?

Reluctance to Align with FIDE ELO?

Rationale for sticking with ECF System

ECF to elo conversations