Board order rules

[Sean Hewitt]

For what it's worth I don't think there should be a choice in which rating / grading to use. If a player wants to play higher / lower than they should then that is what the 80 point rule is for. This should be about getting the rating for the purposes of the 80 point rule right.

[Jonathan Rogers]

That's true and it is more accurately put than my effort last night (saying that the rule would be there to "help" the juniors).

But we do hit the borderline problem. We can be very sure that virtually all undersixteen year olds are more accurately graded than elo rated, because they have played so few rated games and are likely to be improving quickly. Between 16 and 18 this is perhaps still true for a clear majority, but not necessarily everyone. Between 18-21 it is obviously a diminishing majority - some will play regularly (taking gap years or just filling their University vacations) and their Elos will become just as accurate as their grades. And are these players generically still improving in leaps and bounds at 21 years old? I think that was what I was after when I wondered about allowing choice, because of the variation in accuracy of Elo due to playing opportunities and rates of improvements.

[Sean Hewitt]

To be fair Jonathan, if players "Elos will become just as accurate as their grades" then it doesn't matter that you use their grade for board order purposes.

The only issue would be if their rating was more accurate than their grade.

[Jonathan Rogers]

Understood.

What is the feeling then - grades for players up to 21 years old? Always trusting that the ECF graders don't simply go mad as in 2009, of course, which possibility may be the last remaining objection.

[Alan Walton]

A simple point would be for the junior players to play more FIDE chess tournaments away from the 4NCL and try and get their ratings up quicker, I would reckon out of all the junior who have FIDE ratings that alot of them are reasonably inactive in terms of FIDE chess

Another piece of analysis is to see what results these juniors are actually having when they play rated games, it could be that when they play the "odd" game against a FIDE player they are not getting the results consistently to improve their rating

Also, due to the inflation of the ECF system in the past year or so, some of the junior's ECF grades have become ridiculously high, some have become 190s when they should be 170s, and 170s should be 150s. I played Craig Whitfield at Leyland last August and his FIDE was 1980ish (compared to my 2147), but the equivalent ECFs were 192 to my 193, the game was reasonably straightforward. So his conversion would be 2186, but with his ECF being too high in my view (I think at best he is 175 in old money), then at best his conversion would be around 2050.

Maybe the current conversion calculation needs revisiting again

[LozCooper]

A simple point would be for the junior players to play more FIDE chess tournaments away from the 4NCL and try and get their ratings up quicker, I would reckon out of all the junior who have FIDE ratings that alot of them are reasonably inactive in terms of FIDE chess

Another piece of analysis is to see what results these juniors are actually having when they play rated games, it could be that when they play the "odd" game against a FIDE player they are not getting the results consistently to improve their rating

I suspect that the number of weekends that have to be devoted to playing in junior qualifying tournaments for the European & World Youth inhibit the time that juniors spend playing rated chess. Thankfully the number of high quality e2e4 events held over weekends and 5 day schedules give our juniors a lot more opportunity to play rated games than used to be available

[Alex McFarlane]

In Scotland to convert an ECF grade into real money we use the formula 10xECF + 180. A number of northern events also use this formula in reverse when Scots enter.

Rob Thomson's 173 ECF would convert to 1910 which is only 4 different from his FIDE.

Prior to the ECF revision of grades we used x8 + 600 but did some maths to find a better conversion factor based on players Scottish and English grades.

[Roger de Coverly]

In Scotland to convert an ECF grade into real money we use the formula 10xECF + 180.

The reforms to the ECF grading system in effect reduced everyone to 80% of their previous value. They then added back 40 to 50 points to make the top grades look the same. That's why 10*ECF now works for the Scots, because it reverses out the 80%.

The general problem seems to be that in games between players of different ratings, the lower rated players is a bit tougher than linear rating formulae suggest. It doesn't mean that the ordering of people by strength is wrong, but the value of grades as a predictive measure is lessened. Linear systems predict that if 200 points represents a 3-1 score over 4 games, then 100 points is 2.5 - 1.5. This may not be the case in practice.

The ECF practice of treating juniors as new players every year must have some side effects. You also need to observe that the ECF junior grade isn't just the performance, but also contains an age related factor not present in the international Elo system.