Grading Lists: Twice a Year?

[J. Moore]

Speaking as player who returned back in January 2009 and therefore still no official grade, I think its a very good idea to have a grading list that comes out more often

[Ben Purton]

The gentlemen john moore? so good they named a university after him. Demands more grading lists!

Does anyone know if my grade is below 190, i defo wont be able to coax nevil to give me board 1 in London league if im sub 190.

Ben

[Carl Hibbard]

Does anyone know if my grade is below 190, i defo wont be able to coax nevil to give me board 1 in London league if im sub 190.

No, drop to 185 and I will be one point higher

I could manufacture that scenario anyway of course that might be fun

[Richard Haddrell]

Congresses mostly report results with little delay after the event.

“Hmm," I thought. “How does Roger know this? And what does he mean, anyway?” So I did some counting. I won’t have got it exactly right, but the figures are something like this.

Congresses 2009-10: time taken to report
same day ............ 5
1 week ...............51
1 month .............74
2 months ............40
3 months ............16
4 months ............15
5 months ............14
6 months ............ 7
7 months ............ 8
8 months ............ 2
9 months ............ 1
10 months .......... 0
11 months .......... 3
12 months .......... 4

I don’t know if that’s the sort of “little delay” Roger had in mind. Congresses are actually requested to report within one month, and about 54% did so. Some were much quicker. Of the remaining 46%, some were numbingly slow.

I’m not suggesting this kills the idea of two or more lists a year. I’d like to see two or more lists a year. But there’s no harm in bringing some facts into the debate. And let’s remember that congresses only account for about 40% of results. Leagues will always be slower.

[Kevin Thurlow]

"No you can definitely look up the last published grade as well as the new 'current' one. "

Thanks - perhaps people got confused because they didn't realise they should look at the "old" one?

"Leagues will always be slower."

I do hope that no league I play in behaves like this!

[Adam Raoof]

Congresses have to report results with a little delay; they can't report them before the event takes place.

Richard's statistics are actually quite reassuring; most events report within two months, and I believe with a little encouragement we could capture most of the data in one month. We need to approach organisers with help and advice about grading, guidance on how to use grading software / pairing programs to automate the process. We are already introducing a penalty for events submitted late to the Grand Prix. Is it reasonable to apply a penalty to events submitting grading data up to 12 months late?! Richard - let me know who they are, and I will chase them!

Leagues are under no pressure to submit grading data any more speedily than before; though it would make it simpler to use the bespoke software available to enter match results. As I said below, in the Netherlands (and Belgium) you just use the grade published at the start of the league throughout the period of the league. You should probably use new grades where available. Leagues end in May, and the grading list is already published in July, and therefore the situation for grading leagues would not change.

[Mike Gunn]

No, because:

1. The Clarke system was designed to be operated on a yearly basis and suits the league structure we have in this country. Leagues tend to run for about 8 months of the year and so there will be an inevitable imbalance between the two halves of the year.

2. I don't think that recent tinkering with the Clarke grading system has been done well. Examples include the algorithm for introducing new players (specifically the iteration invoved) and the way that junior players are graded. Based on this experience I'm not confident that further tinkering with the system will respect the (statistical) principles on which the system is based.

3. Leagues will probably stick to grades in force at the start of the competition (which is the most straight forward thing to do), but the existence of the new grades will lead to arguments and confusion. This is because even if the rules stipulate that initial grades are used a significant minority of players will argue that the new grades are relevant to matters such as determining relative strength board order, even if the league rules specify otherwise.

The desire on the part of players to have a more regularly updated grade is a reasonable one, but if you are going to do it you have to use a system properly designed for that purpose (i.e. an Elo type system), but you could do that using a linear prediction function so that ordinary players could check their grades using simple arithmetic just like they can (sorry, used to be able to) using the current ECF system.

[Roger de Coverly]

“Hmm," I thought. “How does Roger know this? And what does he mean, anyway?”

It was the observation that cross-tables are usually up on a website very soon after the event. On the premise that data on one computer is available to another computer, then the data can potentially be included in the grading database at much the same time. It also comes from noting the congress names appearing (or not) in Richard's event list.

Leagues have problems with match captains who find difficulty in recording the names of who played who, what the results were and communicating the match result to third parties.

[Ian Thompson]

Leagues are under no pressure to submit grading data any more speedily than before; though it would make it simpler to use the bespoke software available to enter match results. As I said below, in the Netherlands (and Belgium) you just use the grade published at the start of the league throughout the period of the league. You should probably use new grades where available. Leagues end in May, and the grading list is already published in July, and therefore the situation for grading leagues would not change.

If you are suggesting that leagues should only report results once, at the end of the season, then I vote for once a year. If leagues are required to report results to date prior to each grading list then I might vote for twice a year, provided an appropriate mathematical method is used to calculate the grades.

It's worth pointing out that the Upham websites that you refer to make entering match results easy, but they do not provide output in the format required by the ECF grading system, so there is still work for the grader to do. Most importantly, they don't guarantee that the player details entered are complete. The grader still has to make enquiries to establish the correct identities of players reported as new on the website who may already have a grading code.

[Roger de Coverly]

It's worth pointing out that the Upham websites that you refer to make entering match results easy, but they do not provide output in the format required by the ECF grading system, so there is still work for the grader to do.

It's a key point on any acceleration of the grading system that the grading data is produced automatically or near automatically from whatever system is used to input and collate results. It would be helpful if the data format needed by grading system was more widely circulated. It might also help if the ECF grading system was able to accept data in the format required for Fide rating.

[Peter Rhodes]

If the league grader feels that they have to submit results during, rather than just at the end of, the season, they might not be happy

Maybe it's time for some people to move on. Submitting results is not rocket science.

[Adam Raoof]

It would be helpful if the data format needed by grading system was more widely circulated. It might also help if the ECF grading system was able to accept data in the format required for Fide rating.

Attached is the guide Richard Haddrell supplies graders, which details the format of the Excel spreadsheet that he requires. I am also attaching the Excel template. There is a Checker - a piece of software supplied to graders which checks the format for validity. Tournament Director provides the output that the ECF needs automatically.

If the data is in a spreadsheet it is relatively simple to make it FIDE compatible.

I have volunteered to grade our internal events at Hendon Chess Club, so that I actually know what's involved in the grading process and can have a more informed opinion.

[John Upham]

It's worth pointing out that the Upham websites that you refer to make entering match results easy, but they do not provide output in the format required by the ECF grading system,

I am working on a specification for an XML standard for the transmission of results information. My goal would be to create a data standard (schema http://www.w3schools.com/schema/schema_intro.asp for chess results data.

However, the motivation for doing so would be enhanced by being welcome by those who would use it!

This, I believe, would help to smooth the way to more frequent rating updates regardless of use of the Elo or Clarke or some other system.

[Ian Thompson]

Attached is the guide Richard Haddrell supplies graders, which details the format of the Excel spreadsheet that he requires.

... and the alternative text file format (which I thought was the preferred submission format)?

[Roger de Coverly]

Attached is the guide Richard Haddrell supplies graders, which details the format of the Excel spreadsheet that he requires. I am also attaching the Excel template.

I have volunteered to grade our internal events at Hendon Chess Club, so that I actually know what's involved in the grading process and can have a more informed opinion.

There's a load of extraneous data being asked for - which must make the submission process more complex. For example for an established player you should only need the event PIN and the ECF grading code - date of birth, club etc. will already be known. Event pin isn't really necessary as long as every player has a grading code.

I doubt if there's appetite for tampering with a (presumably) working system - but some streamlining would be desirable for faster submission.