new grades wow brandon clarke shot up 206

[Carl Hibbard]

Nonetheless it does lead me to another line of enquiry, which I am just beginning. If anyone is interested drop me a private mail and I'll send copies.

Send it to me and I will turn it into a .PDF document and attach it into here for you

[Roger de Coverly]

but I was hoping it could show the various influences in the system. Frankly, they are not obvious to me!

To my mind, the logical approach to reviewing the various influences is to parallel run. In other words if you want to see the practical effect of the junior increment, just remove it and rerun. Particularly if you ran over several years, you could see how the effects evolved over time. Similarly if you made the 40 point rule a 45 point rule or you changed the estimation process for new players. At the end of the process you should have a reasonable idea of which parameters cause or counter inflation, deflation and spread.

[Brian Valentine]

Carl,
I can do the pdf, but its rather useless without the spreadsheet. Can you do something for both orshould I investigate Alex's idea?

Roger,
I think the spreadsheet can be adapted!

[Carl Hibbard]

Carl,
I can do the pdf, but its rather useless without the spreadsheet. Can you do something for both orshould I investigate Alex's idea?

Roger,
I think the spreadsheet can be adapted!

Up to you....

Or send to "carl AT ecforum.org.uk" and I will see what is possible

[Paul McKeown]

Brian,

I would be interested in seeing your work - I value the clear and well thought out statement you make regarding grading issues, on a forum that otherwise sometimes seems to drown in nonsense. You might find scribd generally useful for exchanging or online publishing documents, although I can't remember whether or not it supports Excel.

Regards,
Paul McKeown.

[E Michael White]

I've been playing around with a matrix version of the ecf grading method. The final equations support many of the points made in this and other threads.

Clearly a matrix approach to a 10,000ish player population is not sensible, but I was hoping it could show the various influences in the system. Frankly, they are not obvious to me!

Is your approach similar to mine ? in - (E Michael White » Tue Aug 18, 2009 8:33 pm) - when the matrix goes singular it indicates one or more pockets of players who only play against one other in a closed group. Subject to certain assumptions the iterative approach always produces the same result as solving the equations.

[Paul McKeown]

Is your approach similar to mine ? in - (E Michael White » Tue Aug 18, 2009 8:33 pm) - when the matrix goes singular it indicates one or more pockets of players who only play against one other in a closed group. Subject to certain assumptions the iterative approach always produces the same result as solving the equations.

That seems a very interesting and intuitive approach - matrix singularity indicating a closed group! I have been away from the forum for a while and have obviously missed some interesting postings, which I shall have to read. Michael, a perhaps stupid question whilst I get my head around this, do you anticipate that the whole grading pool should form a closed group? Obviously not, as the whole pool going singular is not useful (!), but why isn't it? They have to form an APA or APA (more than once), is I assume the answer. Right?

Regards,
Paul McKeown.

[Roger de Coverly]

Roger,
I think the spreadsheet can be adapted!

I was rather more thinking of using the actual data. The "red" grades were computed using a process which took results from 2002 onwards. Given the number of different versions of the "red" grades that have been produced, there must have been numerous computer runs. My approach would have been to use the "2008" system grades as the control and test the effect of various modifications. Spread, deflation and inflation are quantities that you need several years of observations to monitor.

[Brian Valentine]

Michael,
It is closely related, but eliminates the need to iterate. Singularity can always be a problem, since there is a matrix inversion.

[Brian Valentine]

Roger,
Nice idea, but two things:
1) shame the data required is not available (in a suitable form),
2) one would think the grading team had done something similar already. Otherwise this would be a fruitful project to understand the dynamics of the influences.

I am still interested in something that measures:
1) Inflation/deflation in the system
2) Accuracy of ranking

and such measures must be useful for any future investigation. I suppose some would argue that stretch should also be measured, but to me this is a secondary issue.

[Alex Holowczak]

He just won the U210 section in Leamington Spa (rapidplay), ahead of Russell James (who he narrowly lost to in one of the games) and two others whose names escape me. His 206 is 206 on merit, and no inflation result. Interestingly, there were still a glut of juniors doing well.

[Richard James]

Yang-Fan Zhou just won the Open section of the Richmond Rapidplay ahead of IMs Cech, Buckley, Wall, Basman etc..

However...

A new 12-year-old chess player has just joined the school where I teach. His rapidplay grading history is as follows:

2006 0
Jan 2007 0
Jul 2007 43
Jan 2008 46
Jul 2008 49 (also standardplay 47)
Jan 2009 55
Jul 2009 134 based on 13 games last season and 20 games in total.

These 13 games would, I guess, have come from Kent junior tournaments.

I encouraged him to play in today's Richmond Rapidplay. His new rating made him top seed in the Intermediate (U135) section, having improved a massive 79 points (well, let's say about 50 points given the new system) in the last six months. Needless to say, he was outclassed in most of his games, losing the first three but recoving to score 2/6, but the Minor (U100) would really have been the right level for him. Had his grade been just one point higher he would have been forced to play in the Major section.

I really don't mind whether Yang-Fan, Callum or Brandon are or are not a few points too high. They know how good they are without having to ask a grader, and I'm sure they're more interested in their FIDE ratings anyway.

What I object to is that the, in my opinion, absurd decisions taken by the ECF graders have the potential to put off average strength juniors for life by giving them grades which are 30 or 40 points too high.

[Matt Harrison]

It can of course work the other way and give them confidence. For example, my son's rapidplay grade shot up from 145 to 184 in the recent regrading, based on some good results at Golders Green. Yesterday he played in the Open there, had a 190+ grading performance and a draw with black against IM Sherwin.

[Richard James]

It can of course work the other way and give them confidence. For example, my son's rapidplay grade shot up from 145 to 184 in the recent regrading, based on some good results at Golders Green. Yesterday he played in the Open there, had a 190+ grading performance and a draw with black against IM Sherwin.

Yes, but it's not the same thing at all. Your son is a strong player who plays regularly and deserves his high grade. I'm talking about average players who only play two or three tournaments a year and have been assigned grades 30 or 40 points above their true strength.

[Alex Holowczak]

Well, I scored 2/6 today, and I was just under the 150 boundary, but that was more to do with messing up openings as black... I wasn't losing because I was outplayed, just because I made mistakes rushing given it was a rapidplay. Was he outplayed, or did he just make mistakes?