ECF to scrap its grading system

[Michael Farthing]

the ECF League Management System provides a far better interface.

Does it accommodate the input of any ad hoc game between two players?

Not by the players themselves - rightly so.
However, an authorised user could in theory generate a competition entitled "M Farthing v R de Coverley Challenge Match" or, more realistically, "Morecambe Club Friendlies". It can, and already does, support internal club competitions.

Also, of course, it is still an evolving system.

[NickFaulks]

I'm fairly sure that London, like Manchester and many others, uses John Upham's system that predates the ECF one

Yes, I'm also sure that is the case. It is a good system to use and I hope it does not get thrown out, through either the front or the back door.

[Adam Raoof]

I am currently using it to record results of graded games played between Hendon Chess Club members, and it's very useful in this regard.

[Nick Grey]

There was a suggestion that the league management software was going to be free. But not clear what it was & left leagues with little choice.
Plenty of LMS available - FAs had a great one.

Nothing wrong with John Upham or other systems. But I see little need for a monthly update or get the support to grade. May be only 6-8 matches in a league.

May be available for supporting grading in July. Not for the other 11 months for example.

[Alex Holowczak]

There was a suggestion that the league management software was going to be free. But not clear what it was & left leagues with little choice.

The website is free for graded leagues to use. Lots of leagues are using it on that basis already.

[Nick Grey]

Alex that may be the case but did the ECF give those LMSs already being used a chance to tender for the national system?

My main gripe is that I cannot see the benefits of a monthly production of grading results against the admission involved in collecting them.

Perhaps those juniors have time on their hands to help.

[Alex Holowczak]

Alex that may be the case but did the ECF give those LMSs already being used a chance to tender for the national system?

Yes. We had a large number of responses, from memory.

My main gripe is that I cannot see the benefits of a monthly production of grading results against the admission involved in collecting them.

Which is why we included in the tender a means for that to happen seamlessly if people used the LMS. That was one feature that is yet to be added, although you can do it at the press of a few buttons already if you want to.

[Roger de Coverly]

My main gripe is that I cannot see the benefits of a monthly production of grading results against the admission involved in collecting them.

Collecting and publishing game results as they happen shouldn't be a major issue even though it appears a cultural shock for some club officials to think in those terms. Once you have collected the game results, the grading problem is down to how you extract and process them. That's non-trivial though, depending on what you are expecting the grading or rating system to measure. If you want a stable estimate of strength, updating this after every individual league game introduces fluctuations almost for the sake of it. If you require a system of perceived rewards and punishments, then perhaps you do want an update after every game.

Tournaments using the chess-results server now routinely update the pairings with results as they become available. As does the 4NCL for that matter.

[Nick Grey]

Quite OK on chess results server it is just you need to convince Council & league officials to input onto the system.

If monthly updates are quick, easy & cheap then do it. Quite irrelevant for many players. Annual was fine. Not sure that bi-annual adds anything.

I suspect there may need to be more graders not less. That's all.

[IM Jack Rudd]

Internal club competitions are the sticking-point here because of the prevalence of the "post a crosstable up on the wall and people fill in results when they play them" method of reporting results. It's hard to get monthly updates when the tournament organizer doesn't know what month a game was played in.

[Angus French]

If Month Played is unknown (or not easily discoverable) cannot Month Reported be used?

But is not a greater issue with the idea of monthly grading the fact that it currently takes something like 23 days (I think two weeks for the local results officers to submit and then nine days for the central grading team to calculate, check and publish) to tournaround a grading list? If this is maintained and there's a switch to monthly grading then January's list wouldn't be published until a just over a week before the cut-off date for February games and the start of the March period...

I'm wondering if it mightn't be an idea to have *definitive* six-month grades (as currently) and *indicative* faster-turnaround monthly grades which reflect only those results reported quickly.

Anyway, I think it would be an extremely wise idea to consult the local results officers about any proposed switch to monthly grading since it is they who, collectively, do most of the work to produce grading lists and because a promise was made not so long ago to consult by the director then responsible for grading.

[Roger de Coverly]

But is not a greater issue with the idea of monthly grading the fact that it currently takes something like 23 days (I think two weeks for the local results officers to submit and then nine days for the central grading team to calculate, check and publish) to tournaround a grading list?

To make it credible, monthly grading would have to be done to a similar timetable to FIDE, namely that you have a reporting cutoff a few days before the month end and you aim for publication on the last day of the month, or the first day of the following month.

With all the estimation processes for new players and juniors, I could imagine the ECF method is more calculation intensive than FIDE's or the USCF's Elo methods. So there are perhaps two questions here. The first being how long it takes to collate and check for consistency all the result data. The second being of how long it takes after the results are believed to be complete to generate and validate a grading list.

[Angus French]

To make it credible, monthly grading would have to be done to a similar timetable to FIDE, namely that you have a reporting cutoff a few days before the month end and you aim for publication on the last day of the month, or the first day of the following month.

Hmm. For evening chess leagues, teams need time to report match results. Really you want both teams to report the result so that it can be verified. The results officers then needs time to verify results / resolve descrepancies and produce a submission file for the ECF.

[Alex Holowczak]

But is not a greater issue with the idea of monthly grading the fact that it currently takes something like 23 days (I think two weeks for the local results officers to submit and then nine days for the central grading team to calculate, check and publish) to tournaround a grading list?

The deadline for July (e.g.) is normally about 15th July. Leagues are normally finished in the middle of May, so a league grader has maybe 6 weeks to get the results in. I suspect there's a heavy element of graders working to a deadline, and if the deadline was tighter (as it is in January), they'd get the results in sooner. But if the deadline is 15th July, why do the grading on 29th May?

I see that there might be an argument that a weekend congress finishing on 30th June, as may be the case this year, may struggle to get its results in that day for a list that's meant to be published on 1st July. But equally, if the next list after 1st July is 1st August, rather than 1st January, the issue of getting the results in that night isn't so big a deal. In fact, for FIDE-rated events three days before the end of the month is the deadline. So, for example, the World Rapidplay Championship will be FIDE-rated on 1st January, but the World Blitz Championship will be FIDE-rated on 1st February.

There's probably a genuine human resources issue at the ECF end of this process; will the right people be in the right place when the time to push the button goes live? It's the sort of thing where I imagine if it were a monthly issue, a more robust process will be put in place to make sure that happens than the prevailing process now. Whereas with a once-every-six month list, "it can wait a few days until I get back from Gibraltar", or "it can wait a few days until I'm back from holiday in India" is less likely to cause delays than in recent years, because that sort of thing will happen so frequently that the process will be designed to make sure that can't delay the system; whereas now, no one will notice if there's a 2-day delay because it takes so excruciatingly long anyway.

[Roger de Coverly]

For evening chess leagues, teams need time to report match results.

If you are allowed to switch on your phone, you could report from the venue. Failing that you report once you get home. Same day if you are home before midnight. The opposing captain only needs to confirm.

If you had the headaches of adjournments, adjudication, unable to win claims etc. then yes, you need time. As monthly grades would be random fluctuations for most club players, a cut off a few days before the month end (as FIDE do it) seems sensible enough.