Chris Majer returns!

[Richard Bates]

The biannual standard gradings are going to be straightforward, and the decision was taken before CM's appointment.

Was it decided if the list was going to be based on a year's worth, or six month's worth, of games? If the latter you are theoretically going to have to play 60 games a year, a practically a lot more, to maintain an "A" grade.

Richard Haddrell is of the opinion (or was, as of the 2010 ECF AGM) that it'll be calculated in the same way as the rapidplay grades. So you'd get a * grade by playing 30 in a 6-month period, and A if you played 30 over the course of a full season.

A "*" grade? Never heard of it It seems to be illogical with the workings of the ECF system to base a grade for established players over a period other than a full year, and I think it might create a pattern of people consistently have a "winter" grade and a "summer" grade, based on different forms of chess (leagues/tournaments/quick time limits/slow timelimits) with noticeable difference between the two but there we are. It's a bit different with Rapidplay since the vast majority of rapidplay games are played at a consistent format.

[Adam Raoof]

It seems to be illogical with the workings of the Clarke system to base a grade for established players over a period other than a full year...

I've always thought using the calendar year was rather arbitrary. Perhaps the grading season, like the Grand Prix, should end with a major event such as the British Championships, and grades published very soon after that and the Hastings International, for example.

[Christopher Kreuzer]

A "*" grade? Never heard of it It seems to be illogical with the workings of the Clarke system to base a grade for established players over a period other than a full year, and I think it might create a pattern of people consistently have a "winter" grade and a "summer" grade, based on different forms of chess (leagues/tournaments/quick time limits/slow timelimits) with noticeable difference between the two but there we are.

I think Alex H may have meant X grades, as described here:

http://grading.bcfservices.org.uk/help.php

"Grade based on 30 or more games in the latest half-season".

Unless "*" is something new that has been proposed?

Possibly some people might take advantage of this by "deflating" their grade over a league season in time to enter certain grading bands in summer tournaments. But most leagues seem to stretch across any dividing line, so shouldn't be a real problem.

League chess is definitely different to county and tournament chess, though, as the standard can vary a lot. In county chess, you invariably face similar strength opposition. In tournament chess, if you do well, you also play players of the same strength or higher. In league chess, the standard of the opposition doesn't change depending on how well you are doing that year.

I agree with Adam that putting the cutting-off point after major events would be logical, but pity the graders trying to rush through results from those events.

[Richard Bates]

It seems to be illogical with the workings of the Clarke system to base a grade for established players over a period other than a full year...

I've always thought using the calendar year was rather arbitrary. Perhaps the grading season, like the Grand Prix, should end with a major event such as the British Championships, and grades published very soon after that and the Hastings International, for example.

The cut-off month may be arbitrary. I don't see how the calendar year can be interpreted as such - it is the period in which every event is covered once and once only (for an A graded player). And using a calendar year renders the former irrelevant. A six monthly basis means that eg. the British Championships would be graded on one list, but not the other.

As for "ending with a major event" - doesn't that by definition mean that the grades used at that event are as out-of-date as is possible to be? The difference with something like the Grand Prix (as was, i don't know about the current incarnation) is that the season ending event (like the World Snooker Champs as an analogy) was given extra weighting and was therefore more relevant to the final outcome.

[Alex Holowczak]

A "*" grade? Never heard of it It seems to be illogical with the workings of the Clarke system to base a grade for established players over a period other than a full year, and I think it might create a pattern of people consistently have a "winter" grade and a "summer" grade, based on different forms of chess (leagues/tournaments/quick time limits/slow timelimits) with noticeable difference between the two but there we are.

I think Alex H may have meant X grades.

Chris is right, I meant X grades. I always thought of it as an A*, given I was at school when I first noticed it...

[Alex Holowczak]

As for "ending with a major event" - doesn't that by definition mean that the grades used at that event are as out-of-date as is possible to be? The difference with something like the Grand Prix (as was, i don't know about the current incarnation) is that the season ending event (like the World Snooker Champs as an analogy) was given extra weighting and was therefore more relevant to the final outcome.

There are 10,000 points for winning the World Snooker Championship, but 8,000 for the next highest profile event (the UK Championship). So there is a weighting. The way the Grand Prix is calculated, you score 'percentage score in event + number of rounds of event'. So 11/11 scores you 111 points. 5/5 in a weekender scores you 105 points. So it is slightly weighted on account of the event being longer.

[Richard Bates]

As for "ending with a major event" - doesn't that by definition mean that the grades used at that event are as out-of-date as is possible to be? The difference with something like the Grand Prix (as was, i don't know about the current incarnation) is that the season ending event (like the World Snooker Champs as an analogy) was given extra weighting and was therefore more relevant to the final outcome.

There are 10,000 points for winning the World Snooker Championship, but 8,000 for the next highest profile event (the UK Championship). So there is a weighting. The way the Grand Prix is calculated, you score 'percentage score in event + number of rounds of event'. So 11/11 scores you 111 points. 5/5 in a weekender scores you 105 points. So it is slightly weighted on account of the event being longer.

I think you've missed my point. The old Grand Prix used to work on multiples of games played, and the final event was (from memory) given a 10x weighting, much higher than any other event. The point being that by maximising the points available in the final event you increase the chances that the competition will still be undecided when that event happens. This obviously isn't relevant to ECF grading where every game in every event is weighted equally.

[Alex Holowczak]

As for "ending with a major event" - doesn't that by definition mean that the grades used at that event are as out-of-date as is possible to be? The difference with something like the Grand Prix (as was, i don't know about the current incarnation) is that the season ending event (like the World Snooker Champs as an analogy) was given extra weighting and was therefore more relevant to the final outcome.

There are 10,000 points for winning the World Snooker Championship, but 8,000 for the next highest profile event (the UK Championship). So there is a weighting. The way the Grand Prix is calculated, you score 'percentage score in event + number of rounds of event'. So 11/11 scores you 111 points. 5/5 in a weekender scores you 105 points. So it is slightly weighted on account of the event being longer.

I think you've missed my point.

I think I was merely confirming your point about the weighting of the system. You're right though, it's not really relevant.

Back to my project write-up I go.

[Roger de Coverly]

The cut-off month may be arbitrary. I don't see how the calendar year can be interpreted as such - it is the period in which every event is covered once and once only (for an A graded player). And using a calendar year renders the former irrelevant. A six monthly basis means that eg. the British Championships would be graded on one list, but not the other.

The grading year runs from 1st June to 31st May. That period or 1st July to 30th June very conveniently puts the whole league season into one grading period and so works well for players who only play a limited amount of chess if at all outside the traditional winter club season.

It will be a problem particularly with a 30th November cut-off, that players taking a break over the summer will have most of their games concentrated in the 1st December to 31st May period.

The grading system as currently constituted makes up the "30 games to count" numbers by borrowing backwards in time. So if you only had 5 games to November 2012, it would borrow from 31st May going backwards until it made 30. This would double count the games included, since they had already been used in the 31st May calculation.

[Jonathan Bryant]

It will be a problem particularly with a 30th November cut-off, that players taking a break over the summer will have most of their games concentrated in the 1st December to 31st May period.

This season I played 8 club games before November 30th and I've played 18 since then (with two still to go). That does seem rather unbalanced as a division.

The grading system as currently constituted makes up the "30 games to count" numbers by borrowing backwards in time. So if you only had 5 games to November 2012, it would borrow from 31st May going backwards until it made 30. This would double count the games included, since they had already been used in the 31st May calculation.

To go backwards in time to make up 30 games for me for the November 30th cut off you have to go back to 20th November 2009.

For the end of May calculation, I'll have 30 games this season (going back to mid-October 2010). In other words, for me the first grading period would in effect be a year long period and the second one would be seven and a half months - a third shorter.

In my case, then, in the circumstances as given by Roger, the end year end grade would be much more reliable than the mid-season one.

[Alex McFarlane]

If you assume a league season runs from October to March then perhaps the grading 'season' should reflect this. Lists could appear on 1st October and mid April. This would mean that the vast majority of league only players would have all their games on the one list (and no change on the other eventually).

[Roger de Coverly]

In other words, for me the first grading period would in effect be a year long period and the second one would be seven and a half months - a third shorter.

I suspect you may have to use imbalanced time periods to get a balanced game count. Just guessing but would time periods of 1st June to 31st January (8 months) and 1st February to 31st May (4 months) give equal numbers of games?

Though if it's going to take 6 weeks or more to process the grades, then no-one will be interested in a grade which didn't come out until the middle of March. It would also get in the way of the league season, since you wouldn't know the grade to publish for the games after 1st February.

Actually to make publishing a grade halfway through a league season anything other than a complete road block for reporting results, you would have to coincide it with the Christmas/ New Year shutdown in club chess and also get the turnaround time down. So you might use a 15th December data cutoff with a publication date of the first week of January.

The other cutoff date doesn't matter so much. It could be 30th April, 31st May, 30th June or even 31st July without making a lot of difference other than to Congresses. The big constraint being to get it published before the league season so that clubs can decide on teams and board orders.

October to March

Some English league competitions (or clubs) start in September and continue to May. Usually it's where you have teams in more than one league but with the same players in both. Encroachments into the league season from events like the London Classic don't help either.

[Alex Holowczak]

If you assume a league season runs from October to March then perhaps the grading 'season' should reflect this. Lists could appear on 1st October and mid April. This would mean that the vast majority of league only players would have all their games on the one list (and no change on the other eventually).

That's not necessarily a sound assumption. League seasons in the Midlands run from September until April by design, but sometimes overshoot into May. So your lists would have to be on 1st September and 1st June at best.

I think it would make sense to move the grading lists back 1 month, I think. If the lists ended on December 31st and June 30th, rather than November 30th and May 31st. Leagues will be graded at the end of May as normal, so it wouldn't massively delay things. With no membership pre-check, and the results being printed online, that would even speed up the system. Not only does that fit better with the Gregorian calendar, but you get around the problem of equalising the split.

[Roger de Coverly]

Lists could appear on 1st October and mid April. This would mean that the vast majority of league only players would have all their games on the one list (and no change on the other eventually).

The problem is that the ECF system as currently constituted needs 6 to 8 weeks of processing time. So with the present 31st May cutoff, the results appear just in time for the British. So a list published on 1st October is a cutoff date of 1st July and mid April is cutoff mid February.

In practice you need the list 1st September at the latest rather than 1st October because if you don't know player's grades you don't know what teams you can enter if you have grading restricted competitions, or how to rank your players if you don't.

[Sean Hewitt]

The problem is that the ECF system as currently constituted needs 6 to 8 weeks of processing time.

I don't know if the ECF needs this amount of time or not. They certainly take that long currently but FIDE produce their rating list just 7 days after the cut off.