Grading Cheat or Patzer?

[Nick Grey]

ECF need to sort out conduct. Seems strange to resign if all knew the results were unusual.

More concerned that hounding of volunteers, and voting for none of the above appears to be usual business.

Quite honestly getting volunteers for any chess position In clubs, leagues, associations, congresses, unions, etc is difficult, improbable & unusual.

[Roger de Coverly]

Mr Crockett's results were unusual. The point being?

You've been told the point. It enabled him to enter and win tournaments he would not otherwise have been eligible for.

Some players, knowing they are on bad form will take a break. That includes both not entering Congresses, declining invitations to play in leagues and withdrawing halfway through tournaments that have started badly. If you don't do this and continue with bad results, to what extent are you deliberately allowing your grade to fall?

It was his peers in Minor tournaments, not Horton, who started the comments and they will have observed at close hand what they considered dubious behaviour. If the ECF were aware of a reputation, why did they appoint him?

[Gareth T Ellis]

Roger de Coverly » Fri Feb 05, 2016 12:21 am

You've been told the point. It enabled him to enter and win tournaments he would not otherwise have been eligible for.

Some players, knowing they are on bad form will take a break. That includes both not entering Congresses, declining invitations to play in leagues and withdrawing halfway through tournaments that have started badly. If you don't do this and continue with bad results, to what extent are you deliberately allowing your grade to fall?

So every player that has an alcoholic drink during a game or drinks heavily the night before, doesn't withdraw from events/league matches when not 100% well or is undergoing pressure at work or generally in their life is now fair game for a witch hunt.

As that includes some of the more active GMs/IMs good luck telling them. As for "withdrawing halfway through tournaments that have started badly" speaking as a sponsorer, I wouldn't expect a player to do that, illness/personal matters that's different.

[Roger de Coverly]

So every player that has an alcoholic drink during a game or drinks heavily the night before, doesn't withdraw from events/league matches when not 100% well or is undergoing pressure at work or generally in their life is now fair game for a witch hunt.

From the very nature of the competitions, there's going to be suspicion of players who win grade restricted competitions year in and year out. The usual expectation would be that someone winning lots of games would have grade increases ruling them out of the restricted competitions. The ECF grading history permits results to be examined. Usually the winners of Grand Prix competitions do it by shear weight of numbers. If you play twenty tournaments a year, perhaps you win prizes at five of them. The mediocre results in the other fifteen balance the winning five.

If you have a pattern where for a couple of weeks you get a near 100% score and then in the next few weeks get 0% or near 0% scores, then questions will eventually be asked.

[Gareth T Ellis]

It depends who does the asking and how!
The grader, officials at tournaments, a committee ? Not just anyone with an online presence !
Every "suspect" deserves a fair chance to answer questions, as in this country we are innocent until proven otherwise.

[Roger de Coverly]

The grader, officials at tournaments, a committee ? Not just anyone with an online presence !

As I'm sure you are aware, there are no official means of accusing someone of grade manipulation and it will be other frequent competitors who are the eye-witnesses. Here and there, the ECF and Congress organisers have closed off obvious loopholes. But you are left with "online presence" as the only means of stating a case. As far as this forum is concerned it came to light when someone checked the results of a player who won a couple of tournaments at the British Championship Congress and then won a section at Scarborough.

[MartinCarpenter]

The data is all there, and publically available via the website, so it is not unreasonable for someone to question it. Can go too far perhaps.

That blog does annoy me slightly - treating the individual games in a run of good/bad results as independent statistical events is just plain wrong. Any chess player knows that

Also do have to remember that there are some people who don't have very good levels of 'internal' motivation so can really struggle once there's nothing much external to play for in an event. I spent several years in the Yorkshire league playing quite well while the league destination was 'live' and then awfully once it was decided. 30+ grading pts difference as I remember for when I last checked.

Not at all deliberate but it hardly matters in performance terms once your subconcious has clocked off.

[NickFaulks]

That blog does annoy me slightly - treating the individual games in a run of good/bad results as independent statistical events is just plain wrong. Any chess player knows that

That's true, but the scale of the lack of independence varies greatly between players. It would seem that grade restricted events, and most especially the Grand Prix based on them, will generally be won by a player in whose case, for whatever reason, this factor is particularly high. Does that cast doubt on the whole concept?

[MartinCarpenter]

Come to that, the grand prix very strongly encourages it.

If you're trying to win a grade limited section of the grand prix, then its your best 7 events counting. I guess you want/need most of those to be 4.5 or 5. To do that, presuming you're 'merely' top 1/4 of the field, then that means playing something like 15-20(+?) weekenders and really gambling in the first three rounds of each of them to try and reach 3/3. Huge commitment of course but people do do odd things.

If one of those early round gambles backfires? Well 4/5 probably isn't going to be a counting score anyway, so your motivation for that event suddenly plummets. Results probably will too of course, even if trying.

[David Sedgwick]

Did anyone keep a copy of the original, rather than edited, posting of the announcement (or am I hallucinating?)

The ECF website now links to the full statement from Mr Crockett for those interested

Can I check that I've understood?

The ECF published Mr Crockett's statement, then deleted it, then published a link to the statement.

Is that right?

[Mick Norris]

Did anyone keep a copy of the original, rather than edited, posting of the announcement (or am I hallucinating?)

The ECF website now links to the full statement from Mr Crockett for those interested

Can I check that I've understood?

The ECF published Mr Crockett's statement, then deleted it, then published a link to the statement.

Is that right?

Yes

[Roger de Coverly]

If one of those early round gambles backfires? Well 4/5 probably isn't going to be a counting score anyway, so your motivation for that event suddenly plummets. Results probably will too of course, even if trying.

Looking at the regulations, what they say is

(ii) In a Standard Play event the Member’s Grand Prix points are calculated as: – percentage score in the event, with any fraction of a percent ignored; plus the number of games Played in that event (up to a maximum of 11).

This means, I think, for a 5 round event, you score
105 for 5/5
95 for 4.5/5
85 for 4/5 etc.

Winning scores in the U120 and U140 Grand Prix seem to be between 650 and 700, which as you suggest would need seven scores of 4.5/5 or 5/5.

[Christopher Kreuzer]

Has anyone suggested modifying the Grand Prix rules to take some form of weighted average of all results, rather than just the best ones? (If it was just all results, then you might as well make it grading performance - you need some form of weighting, just maybe not the current form.)

[Christopher Kreuzer]

Did anyone keep a copy of the original, rather than edited, posting of the announcement (or am I hallucinating?)

The ECF website now links to the full statement from Mr Crockett for those interested

Can I check that I've understood?

The ECF published Mr Crockett's statement, then deleted it, then published a link to the statement.

Is that right?

Yes

The practical difference being that if you link to a statement, the person who published that statement can later edit and change it themselves if they wish (add more or withdraw it etc.). That can't be done once it is published on another site.

[Brian Towers]

That blog does annoy me slightly - treating the individual games in a run of good/bad results as independent statistical events is just plain wrong. Any chess player knows that

That's true, but the scale of the lack of independence varies greatly between players. It would seem that grade restricted events, and most especially the Grand Prix based on them, will generally be won by a player in whose case, for whatever reason, this factor is particularly high. Does that cast doubt on the whole concept?

There is a lot of nonsense / misunderstanding being spouted about "statistical independence" that is exactly counter to the facts.

A statistician looking at these results would form the hypothesis that they are statistically dependent. That is, that the 4.5/5 and 5/5 results were obtained by the same person that obtained the other results. If the (calculated) odds of this being true were less than 1 in 20 then this would be flagged as significant at the 95% level and call for some explanation. Two different people, health problems, sandbagging, etc.

The interpretation of these results would be more problematic. If all the good results were in one time period and all the bad in another then this would suggest some health or mental related problems leading to a loss of form. If there were a high level of interleaving of good and bad then this would lend support to a sandbagging theory.