The controversy involved the unusual pattern of his results, specifically the way in which he seemed to score either very well or very badly, but much more rarely achieve a mediocre, mid-table score.
Using the ECF's grading database (for assistance, see yesterday's posting) I looked up Mr Crockett's standard play tournament performances for the season 2014-15 (I've omitted club games) which was the last full season in which he competed. It was also one in which he won the Grand Prix for the grading category 120-139. If I have rendered them correctly, they run as follows:
4/5
4.5/5
4.5/5
3.5/5
2/4
0.5/4
4/4
0/4
0.5/4
1/5
1/4
1/4
3.5/5
4/5
3.5/4
1.5/5
0/4
4.5/5
4.5/5
5.5/6
3.5/4
So out of twenty-one events, he was within a point of scoring 100% ten times and within a point of scoring 0% seven times. There was one score of 1.5/5, two scores of 3.5/5 (on both of which occasions, as it happens, he lost in the last round) and only one genuinely middling score, the 2/4 he achieved in the Bristol Summer Congress in August 2014.
That's an eccentric pattern, to put it mildly, with its near absence of middling results. It's a pattern which continued into the 2015-16 season, with the following scores:
3.5/4
1/4
1/5
5/5
0.5/4
0/5
0/4
0.5/4
1/5
4.5/5
1.5/4
That's another eleven tournaments, and in all but the last - played two days before he "quit being a circuit regular" - he was within a point of either a maximum score or a minimum. I wonder how common that kind of record is?
Seriously, if there are any comparable records among Grand Prix contestants, could anybody point me to them?
I mean some players do achieve consistently very good records, and there are some who suffer very bad ones. What's particularly curious about Mr Crockett's results is that they're almost evenly distributed between the two. Middling results, though, are harder to find. Particularly in his more recent record.
One other noteworthy feature of Mr Crockett's record is a tendency to score poorly across a tournament even when the opponents are all, or nearly all, lower-graded players, something that might not raise one's eyebrows quite so much if the same player wasn't winning four consecutive Grand Prixs (and for that matter, four consecutive Regional Grand Prixs too).
For some reason the grading database doesn't give opponents' grades for the 2015/6 season, but here's a couple of choice examples from 2014-15.
Here's another couple that caught my eye from the 2012-13 season.
Of course those tables don't prove anything in themselves, because anybody can have an off-day. Indeed anybody can have an off-weekend. But what is harder to explain is when a player who wins grading-limited tournaments time and again, in both standardplay and rapidplay, indeed a "record-breaking" number of titles
also scores zero, or half a point, an awful lot.
On top of that he seems rather more likely to score zero, or close to it, against weaker opposition, than to achieve normal, middle-ranking scores.
Perhaps these things happen, now and then. But so regularly? So coincidentally? So advantageously?
I?m not a competitor in the Grand Prix. But if I were, I think I would want a record like this to be looked at, properly, if I were to have any faith in the integrity of the competition in which I was taking part. What I would not want is for everybody involved to turn a blind eye. Nothing could do more to convince other players that they have not been competing on a level playing field, and for that matter that they may not be doing so in the future.
Because what's going to happen next time we have a bizarre set of results that affects prizes and prize money?
- - -
As it goes, the most eyewatering sequence I have found in Mr Crockett's record came in the 2010-11 season, just before his long run of Grand Prix success, when he suffered a long run of failures in rapidplay games.
A very long run.
A very long run indeed.
We begin in November 2010 at the British Rapidplay Championships Open, where after a promising start of 3/5, our man suddenly collapsed and lost his last six games.
As I say, these things happen. Matter of fact they happened again at the next rapidplay in which he competed, where he lost all six.
Next time out it wasn't so bad. One of his opponents was careless enough to let slip a half-point, albeit only one...
...but he was not so fortunate in the tournament that followed
nor in the one that followed that.
Happily in his very next tournament he won the very first game, thus putting an end to a sequence of twenty-seven rapidplay games of which he had lost twenty-six and drawn just one.
You can calculate the odds on that whatever way you like: there's no way they´ll come out as anything less than an astronomical figure.
A run like that would be extraordinary - beyond extraordinary - if we were talking about the worst rapidplay player on the circuit. But we're not.
In the rest of the season, in rapidplay, he scored 23.5/33.
His standardplay scores for the 2010-11 season, incidentally, were perfectly normal.
Look at those numbers again. Twenty-seven games. Twenty-six losses and a draw. A reasonable explanation eludes me.
- - -
In Mr Crockett's valedictory Facebook post he states:
the amount of success I had was way off the scale of what I or anyone else ever expectedand "way off the scale" is about right, if you ask me, where that run of rapidplay defeats is concerned. But he goes on:
I hope to share some of the secrets that helped me do it on this page.That would be most welcome, I think. We'll look at some of Mr Crockett's explanations tomorrow.
[Entirely anonymous comments will not be accepted on this series of articles.]
75 comments:
Certainly a very long losing sequence- but it was four or five years ago wasnt it? Seems strange to be bringing it up now when the informations been freely available for you, colleagues, the ecf, congress organisers and anyone else that cared to look for the past 4 years..if they accepted his entries 3 or 4 years ago why should that change now..and re your blog, Why not say something at the time if you were that bothered that someone was legitimately entering tournaments, paying their entry fees, travel costs etc losing their games and no doubt spending quite a bit of money travelling to all these events. If you were in some way trying to rig your grade why would you spend so much time effort and money to do it over a period of time rather than just chucking a few league games?
Andy -it's irrelevant to the question of whether or not these results are legitimate, but the obvious and correctanswer is - because I wasn't aware of them at the the time.
As for league games, for the moment I'll leave that as an exercise for the reader.
What's 2010-11 results got to do with whether he should agree to volunteer to be a gofer in 2016? I don't get it..
When compiling this sort of thing it's also important that you see how others have done according to the criterium you laid out. You might also have mentioned that he has often entered Major events when he could have gone into lower rated sections.
Without mentioning these things, linking to a forum thread that made accusations of grade rigging and failing to mention the 'alternative hypothesis' (a downward spiral that is possible for some people when things go wrong) this doesn't exactly seem very fair. Hopefully you will be making amends to give a more complete picture, mentioning too that Crockett is seen to be a very pleasant man from those who have actually met him (unlike some of his accusers I might add).
"He's a very nice man" is not a defence really, is it?
As for how others have done, the post invites readers to submit comparative examples, an offer I'd be glad for you to take up.
Personally, I don't find any problem with losing intentionally. Just like in women's badminton, it can be to your advantage. And in chess you won't have the pompous "Lord Coe" clucking that it's "unacceptable" that you didn't perform like a circus animal for the spectators. Maybe every other "grading" ECF GP prize is rigged the same way, just not as clearly. That's the only fault, that it was "too obvious" just like women's badminton. If you were a rower in 1948, you could admit it out loud then every one would applaud when you won the gold at the end, and call it "gamesmanship" or "sportsmanlike dumping". In bridge, it happens too, you lose to not to face a better opponent. So why is chess different?
What effort have you yourself made to find players scoring comparable results?
The databases exist: so while it would certainly take non-zero effort, it must be possible to define some criteria and search for players with result like this.
For what it's worth, and offering no opinion on possible explanations, my intuition / experience suggests that the sequences you list are indeed unusual. But: if that's wrong then it would only be fair to know that _before_ going ahead and blogging about it, and if that's right then it would considerably strengthen your case to provide actual evidence showing as much.
David
Is it possible to take the grades/ratings and give an expected outcome? If so, you can potter along and total up the probability of his results.
With ELO there is, but for ECF I am not sure...
-theblueweasel
I dare say it's possible to do an analysis by expected outcome, but it's hardly necessary here; the most generous of calculations, say assigning a 50% chance of losing to a player of equal or lower strength, would very quickly show the nature of the sequences involved here.
I find it astonishing that people can come up with such flimsy defences as "he's a nice guy" and "why don't you look at other people" or "it was a long time ago".
Truly ridiculous.
What about the other grand prix section winners for the last few years..what makes you so convinced they are so much different to this guy..where's the comparative evidence? Stats about one single person can be used to prove pretty much whatever you want depending on what you choose to highlight..
Meh this sort of thing goes on in all sports and games at all levels.except usually it involves significant amounts of money like in horse racing or in super gm chess tourneys with pre agreed results etc. Hard to see why this case warrants such special attention to elevate the guy to the same level of imfamy as the top grand masters you usually like to mention for various reasons.
Why not get rid of these not especially relevant low level sections for hobby players and just play opens like in other countries so the best players get the best rewards?
@theblueweasel - under the ECF grading system, the expected % score for player 1 graded A vs player 2 graded B is simply A-B+50 (where A-B is limited to +/- 40 for grading purposes)
What about the other grand prix section winners for the last few years..what makes you so convinced they are so much different to this guy..where's the comparative evidence?
I'm happy to have anybody else provide whatever comparative evidence they choose.
Readers are reminded that entirely anonymous comments are not going to appear! Apologies.
ejh: you keep inviting other people to do your work for you. Can we take it that the answer to "What effort have you yourself made to find players scoring comparable results?" is "Absolutely none"?
When you go public with insinuations of dishonest behaviour, don't you think that the burden is on you, rather than you readers, to make it fully convincing?
There are a lot of people out there playing a lot of chess. Finding someone with unusual results may or may not be interesting. Do you intend to make any quantification of how unusual these particular results are?
(It's perfectly plausible that these results are in fact very unusual and that a sensible analysis would show this beyond reasonable doubt - I'm not supporting either defence or prosecution here. I'd just like to see a more concrete discussion).
David
David - no, I'm kind of inviting other people to do their own work for themselves. (Further reiteration of this point will not, I think, be profitable.)
Quick question Justin. Are you claiming the player in question has actually broken any rules and or cheated here?if so which rule have they broken please as your readers would like some clarification. Thanks, Mike
I'm not really sure what the aim of this campaign is or what it's meant to achieve exactly to be honest. Surely by attacking the credentials of people for volunteering to help the ecf it will just make it even harder than it is already to persuade others to volunteer to help out in future and put themselves through this sort of thing?
You do not have to be a maths graduate (though I confess that I am) to understand how incredibly unlikely it is that the results shown are not being deliberately manipulated. For the 26 losses alone the probability of this happening is greater that a trillion to 1 against. Factor in the Grand Prix results and it gets even more absurd. So the apologists either haven't thought through the maths, or are a bit thick, or perhaps have some other agenda (or aren't who they say they are?).
Anyway, this all misses the point of the original post. Whether or not there is an explanation that does not involve Crockett deliberately sand-bagging his way to unjustified prizes, there is a completely compelling statistical case for having it investigated. If my ECF membership money is to be put to some use, I would think that looking into this sort of thing is as good a case as anything.
And, guess what, this should be done before appointing him to a position within the ECF.
Quite. In my view it's not possible (if I may also answer Mike above) to be satisfied, on the basis of the evidence I have, that these results have been properly obtained.
(Once again, I have to reiterate that entirely anonymous comments will not be published today, regardless of their content. Sorry to anybody who has put in the effort to comment. Comments simply repeating points that have been made before are also liable to be rejected, regardless of signature.)
Chess is naturally a game of fact and justice.if you win a game your grade goes up and if you lose it goes down. The grades are a guide to form effectively.you have to pay a game fee to play a graded game.he paid his game fees, he win, he lost, his grade was set accordingly beyond his power to control. What rule has been broken?
If he was losing g games then he wouldn't have been winning prizes while losing those games so he wouldn't have benefited in any great way regardless of how and why he lost surely? He would have to pay money to play and lose those games
Interesting set of reactions. Most people seem unable to accept the statistical case. Suppose the ECF had investigated C's results (they may well have - I simply don't know). They are deeply suspicious - what objective numerate observer could fail to be? - but what do they do about it? If they hand down some punishment (what exactly?) there would be a storm of protest from C's supporters, including (to my knowledge) at least one influential member of the ECF Council. Could they actually make a statistical case stick? The line of least resistance would be to do and say nothing. Perhaps that's what has happened?
Brenog - that's quite plausible but of course the outcome of such a policy would be that suspicious results were just brushed under the carpet and other competitors could have little confidence that similar occurrences wouldn't happen in the future. These aren't at all easy situations to handle, but as ever, if you neglect a problem it doesn't mean that it goes away.
I don't propose to say anything at all specific about rules: that's for the tournament organisers and the, ah, officials responsible for the Grand Prix. I do not however think that "what rules has been broken?" adequately addresses the isues that this case raises.
(Commentor Tim - I'm not going to publish what you posted, but if you want to email the blog at the address on the front page, I'll get back to you.)
My view is that the results are curious but I also don't see any major problem with them per se if there's no sognificant financial benefit been gained. The grand prix is a nice little trinket designed for people who play a lot at events.C played a lot at events. I understand he doesn't play often at them anymore.mayve to cut out any arguments they should just simply say that whoever controls the grand prix isn't eligible to be included in it while they hold that role?Surely that would address any conflict of interest or grade manipulation argument that he's benefiting from the ecf role
On the subject of rules, isn't there one allowing the punishment of those bringing the game into disrepute?
Even on the chance of rules having been brokwn,Surely talking about taking some form of unspecified action against someone who won the grand prix would have been better taken at the time before they were awarded it than several years in retrospect.is it really credible to question someone's results five years after the event when there been openly available for all the last 5 years to anyone who cared to look?ny implication these results were already accepted as valid a long time ago when they resulted in a grade, grand prix results for that year etc.whats changed in last few months to mean they should be looked at again now?
I think there is been of, but as this would be a charge made by the ecf we see probably safe to assume they already considered stuff like that before appointing C..otherwise are they really likely to suddenly do a u turn and charge him now even if appropriate after giving him a role?
Chris - it's an interesting point but I'm not sure it'll appeal to everybody who plays in grading-limited competitions, since for them the whole point is surely to play people of roughly their own strength, and not people who shouldn't be there. The money ain't the point.
Paul - I am afraid that there is plenty of form for results that were once considered legitimate being overturned at a later date - indeed, that's the norm much more than it is the exception. The name "Lance Armstrong" kind of makes that point for me.
Regarding the question as to whether any rules have been broken - we are not the ones to decide if they have been in this instance. Clearly though, sandbagging should not be tolerated by the ECF, or for that matter FIDE, since it brings the game into disrepute. I don't believe that anyone could credibly argue against that.
If I had the time, i would pop the probability of each congress result into a spreadsheet. I don't.
-theblueweasel
I appreciate that you can work our statistical probabilities of various players results but the problem is it can tell you they're unusual but it can't tell you why they're unusual..when you look at cess results online you can check them to pattern match with chess engines. For over the board results how do you take into account all the v stable reasons for results and that some people see affected more by these than others. Also where I the line for someone losing on purpose..If someone plays faster than usual because they have somewhere to be is that sandbagging..how about if someone doesn't like the opponent and can't be bothered to carry on the gm and resigns..how about agreeing a draw after 2 moves to get a prize??
In future if someone pays to enter a chess event and does a lot worse than statistically expected or doesn't seem to be trying hard should that mean we ban them from playing or something?
Well, that might depend on how serious we consider their infringement to be, mightn't it? As it would it any other field.
GW, I mean Paul, no I mean Chris, raises a interesting point. Statistical evidence is an excellent starting point to see whether or not something is unusual. Once you have established that something is statistically unusual, you move to the next level.
Sorry about the confusion with the names there.
-theblueweasel
To be fair to C as others have mentioned,his chess record, inconsistent as it is has been out there for anyone to see for years. He h's been subject to the same conditions as others when entering events and organisers could have told him to play in another section if hey had any concerns. Also it seems he quite regularly plays in higher sections than he needs to.for instance at Blackpool he played in the open last time he played according to his record. Blackpool is the sandbagger paradise because of the amount of prestige on offer and money.i find it hard to believe he wouldn't play in the lowest possible section there is he was truly up to something dodgy.
To be fair to C as others have mentioned,his chess record, inconsistent as it is has been out there for anyone to see for years
Well it's not wholly unknown for dubious pracrice in chess, in various guises, to be "out tere for anyone to see for years" and yet for nothing to be done about it, for all sorts of good reasons, most of which usually amount to "other people didn't deal with it so it must be all right". This being so, inaction by organisers, or the ECF, or anybody else, isn't something from which I'm prepared to draw any particular inferences.
Chris, the whole point is that this is a consistent pattern over a number of years and a large number of games; it's not about the odd poor weekend, or the occasional loss in the last round when your heart isn't in it.
You really can't compare it to the quick 'GM draw' to ensure cash in the last round; that might not be to everybody's taste, but it is at least transparent.
Whether it is cheating is another thing; maybe it isn't, but it does have the effect of ruining the weekends of a lot of people who might have won prizes, but for this pattern of behaviour. How would you feel if it happened to you?
Make him pay back all his wages from the ecf for this job.not fair he is getting a job with chess fee payers money
I don't think there's any reason to think that this is a paid post, is there?
Wouldn't he have had a legitimate expectation of being able to be included in grand prix etc if nothing was said to him at the time regardless of whether any rules have been broken or not?Surely C should have been given a first warning initially at the time if there were any concerns about his results?
Re: sandbagging in general, whether any individuals think it matters, or whather it's cheating, or whatever, isn't so important: what's important is whether the organisations running grading systems and grading-related competitions think it matters.
If they don't, cool - it's not my view but it's up to them. However, if they don't think it matters, they need to say it's OK.
And if it's not OK, they need to show that it's not OK by properly looking into potential cases of sandbagging when they happen.
Paul - you're maybe putting that question to the wrong people.
I have played a few events in the past Peter and from my experience there are always sharks there who have hsd higher ratings or unusual drops or who regularly seem to win.if something could be done about them in general then great but I'm not sure whether singling out one individual like this, before putting out any rules against it even, is the right way to go here.
Well Chris it's hard to see how we deal with many potential examples if we don't even deal with one potential example.
Should it be compulsory for players to have to try equally hard and think equally long in every game regardless of how important the game is to them? If someone plays loads like that what incentive is there for then to be motivated to play at their toughest every event if out of contention?
What do you think should be done about C if he was found guilty ejh? Is this something that should go to the courts?
Maybe in the first instance they should put it some rules or uidance to say if it's ok or not then ejh.clean slate for everyone but cross the line hey lay out and you will be investigated.
What do you think should be done about C if he was found guilty ejh? Is this something that should go to the courts?
I very much doubt it. Besides, the question seems a bit premature.
Maybe in the first instance they should put it some rules or uidance to say if it's ok or not then
Maybe, although I think we do in fact assume that unless a competitition says otherwise, this sort of thing is not OK, and it is not as if anybody is saying "I thought it was OK if I engaged in sandbagging". So talk of laying down rules is essentially a distraction at best.
A quick consultation with the Director of Home Chess reveals that, as suspected, the Manager of the Grand Prix is not a paid post.
Honestly if he isn't even being paid for this role and is just doing the chess community a free service and doesn't even have any control over the gp rules is his really such a big deal?it's nice that people are actually prepared to volunteer to do this sort of thing so he should get some credit for that whatever you honk of his results.
I think a reasonable conclusion to this debate would be that no actual statistical evidence has been offered, just some raw tournament scores of one player plus a link to a thread on a forum in which accusations were made. Of course if an accusation were to be made it could well be actionable, which may well be why this has been avoided.
With regard to cheating in lower sections I've certainly witnessed a fair bit of it in tournaments, but these have in general been perpetrated by scumbags with shifty eyes. Someone's character is seen as a measure of possible innocence or guilt in courts but not apparently on this thread. Has 'niceness' been disregarded because it's not a very common trait here and so therefore can't possibly count?
Meanwhile there might be a case for adjusting rating limits for sections and prizes based on the volatility of someone's results, regardless of whether the results came about via unscrupulous means. There are players of every level who have highly variable results (Veselin Topalov for example) and they tend to be rewarded with a greater share of prize money than if they were more consistent. The same effect could be achieved by not 'top loading' prize funds so that smaller prizes were spread over a larger number of players.
I think a reasonable conclusion to this debate would be that no actual statistical evidence has been offered
Tell you what. In your next comment I'd like you to offer your own statistical analysis of the chances of scoring no more than 0.5 out of 27 consecutive rapidplay games.
How many millions to one, do you think?
There are players of every level who have highly variable results (Veselin Topalov for example)
On the other hand, I think anybody who comes out with a comparison like this probably shouldn't be anywhere near the term "statistical analysis", should they?
With regard to cheating in lower sections I've certainly witnessed a fair bit of it in tournaments, but these have in general been perpetrated by scumbags with shifty eyes.
Ha ha ha ha.
But seriously, please do better if you want to comment again.
@ND
So niceness = not guilty.
FFS grow a brain.
In an amusing coincidence, here's a story about a football team breaking a 26-game losing streak this week.
What are the odds? Millions to one against? Trillions, according to Jon H! No, they must surely be trying to get relegated, right?
Or, perhaps:
- whatever model you're using to come up with these numbers isn't a good one
- and, there are a lot of football teams out there and some of them go on long losing runs.
David (also a maths graduate).
This is a very surprising series of results, but because a lot of people are talking about probability as if it i close to proof, I'd like to say something about my own results.
I dropped 20 ECF points, all in team games played in six months, with a very high instance of blundering. I subsequently won prizes in a couple of tournaments. I get that could look suspicious.
I hope I'm over my minor health problem now; but either way, I am sure a person with a condition that can impact concentration could have a mix of good and bad results with not much between.
Paul C
I subsequently won prizes in a couple of tournaments.
But not, I suspect, 48 first prizes over a four-year period. It's quite a noticeable difference.
there are a lot of football teams out there and some of them go on long losing runs
However, few of them also win the league and several cups in the same season as losing 26 games out of 27, do they?
But you knew that.
Games to download or play through
British Championship Congress 2015
Under 140
http://chess-results.com/PartieSuche.aspx?lan=1&id=50023&tnr=181825&art=3
Yates Weekender (under 125)
http://chess-results.com/PartieSuche.aspx?lan=1&id=50023&tnr=182158&art=3
British Championship Congress 2014
Under 120
http://britishchesschampionships.co.uk/wp-content/uploads/2014/07/U120.pgn
Yates Weekender (under 125)
http://britishchesschampionships.co.uk/wp-content/uploads/2014/07/Yates.pgn
An amusing start to the 2014 U120 event
1. e4 c5 2. f4 e6 3. Nf3 Nc6 4. Bb5 d5 5. Ne5 Bd7 6. exd5 . If someone blunders a piece on move 6 in round 1, I suppose you have to spot it and take it. So 6. .. Nxe5 winning eventually. The winner of the piece would go on to win the tournament conceding a single draw.
RdC
As yesterday, I'm locking commnts overnight so I don't wake up in the morning with dozens of comments to moderate.
Re-opening (and part three) tomorrow morning
..and we're back in the room.
Just to make it clear anybody called Lee or Fbi Lee is not me (Lee Bullock) if I post which I will soon I will be posting as my full name. Just for the record I have been having the debate over Crockett for over 4 years now in online groups. Gaining support of many people who agree this is totally unacceptable and billions of trillions to 1 if we count all the strange patterns and runs.
Anyway I am currently in Gibraltar having fun and will find a moment over the next few days or when I get back to answer some people's questions. I knew Steve very well and was his friend before I became certain this was all on purpose. Then we did stay friends although my constant accusations of him made the friendship impossible as I have morals and could not accept what he was up to. Beating poor Alan Fraser to the GP title one year was the final straw. He was the deserving and real winner and i just couldn't accept it.
I even took the liberty of getting Alan Fraser a trophy and engraved which he really appreciated.
Noting Lee Bullock's comment, what is actually the supposed method? Is it left to chance as to how the first round goes whether it's a "winner" tournament or a "loser" one? I was looking at the games from last year's British Championship Congress, in which to recap, he won the under 125 rapidplay over the first weekend, scored one from five in the Under 140 and then 100% in the second weekend U125. Do his openings change with the objective? In round 4 of the Under 140, the game started 1. f4 e5 2. e4 exf4 3. Nf3 d6 4. d4 g5. That's now transposed to the Fischer Defence to the Kings Gambit and I would have thought the normal 5. h4 was widely known. Instead the wimpy 5. h3 was played and White never did break up the king side pawns.
In the first round of the weekend tournament, 1. b4 was played. It was hardly possible to avoid playing for a win when his opponent captured on b4 with the Bishop from f8, without having ensured that B b2xg7 wasn't possible.
RdC
> But you knew that.
My point is simply that apparently extraordinary things happen all the time. But surely you knew that.
Another example for you: I see that Maria Sharapova has just lost her 18th consecutive match against Serena Williams. What odds would you have given against that? Hundreds of thousands to one, presumably?
Of course these examples are not exact analogs for Crockett's performances; that would be too much. But so far as the 26-game losing streak goes: the fact that two roughly comparable events can be found just in this week's news should give you pause for thought when you claim that such a thing is a millions (or trillions) to one against.
For all I know Crockett is indeed deliberately losing games. But simply observing that he has some results that you find surprising is not sufficient to make the case.
On a different note: I took a look over some of the games that RdC points at. I can only say that at U125 level, it looks to me as though pretty much all games are thrown! Presumably players who are 40-50 points stronger than I am would view my own games as similarly blunder-filled.
David
the fact that two roughly comparable events can be found just in this week's news
Of course they're not remotely comparable events, but you knew that, too, unless we're pretending that "losing a lot of games to a manifestly superior rival" is similar to "losing a lot of consecutive games to players many of whom are manifestly weaker than you".
At 9/2 on a throw it's still quite a streak, but I make the odds against it less than 100/1. As the judge said to Kevin Costner, if that's your case, you have no case.
The Sharapova-Williams odds were of course much closer to evens, sometimes even favoring Sharapova, towards the beginning of the streak.
(And of course recent odds reflect the streak itself; if tennis players had Elo ratings then I strongly suspect that the difference between Sharapova and Williams would not be sufficient to support odds of 2/9).
You are claiming that Sharapova's results (say) are indicative of her strength, whereas Crockett's results are indicative of something else. But your entire argument seems to be "well, it looks that way to me". Maybe you're right; but it's no way to make case.
David
Regarding the point about the reliability of the play, it can be next to impossible to tell the difference between incompetence and deliberate poor play.
What can be spotted is the inconsistency between being able to consistently beat players under 130, 125 , 120 to score near 100% when playing well and being unable to stop the rot with a few draws when playing badly.
Cynical theory - you would like to win a tournament or several. It's not for the money, rather the prestige. You observe that the weaker the opposition, the easier this will be. Even at sub 125 level players aren't complete mugs, so if you don't try hard enough, maybe they will get the better of you and your grade will collapse.
RdC
The Sharapova-Williams odds were of course much closer to evens, sometimes even favoring Sharapova, towards the beginning of the streak.
I'm not seeing anything in your link to say so - am I missing something? (It may be right, but presently I don't have anything to back it up.)
You are claiming that Sharapova's results (say) are indicative of her strength, whereas Crockett's results are indicative of something else.
No, I'm saying we're looking at two entirely different situations which don't bear scrutiny as comparisons.
For what it's worth, you may be right that the odds would be different if tennis players had Elo ratings, and I'm sure the odds do reflect the streak. (We can think of similar streaks in chess (at least where failure to win is concerned, since in chess the draw complicates things where, say, Shirov-Kasparov or Nakamura-Carlsen are concerned, though curiously less so regarding Mr Crockett.)
The trouble is that the odds against Crockett's streak are of such extraordinary magnitude that it is very, very difficult to entertain a theory other than the "something else" to which you refer. We can't address a problem of many-millions-to-one by just saying "oh, maybe his head goes down when he starts off with a loss": not when you have a player who apparently has the fighting spirit to win forty-eight tournaments in four years. I think William of Occam would be able to see a straightforward explanation for Crockett's losing streak which without difficulty would fit all the known facts.
I think there will be another posting on very much this theme.
(Moderating may be slow today I'm travelling, and Blogger doesn't like smartphones.)
There have been detailed analyses of the probabilities involved in tennis. The calculations become very complicated quite quickly. The scoring system distorts the probabilities of outcomes.
Suffice to say, a better analogy would be to say what are the odds of Sharapova losing 26 consecutive games to Williams, rather than matches.
But also, results against the same player are probably not independent enough for the standard binomial probability calculations to be valid.
Anyway, it is very sad when someone suffers from mental health issues. Often, those issues cause a person to act in an unconventional way. In this case, that unconventional way might have been to use a sandbagging approach to lower his grade so that tournament victories in lower sections were easier to come by. And thus gaining self-esteem in such a way.
So I do not feel anger just sympathy for the man. Of course, the real point of this whole discussion is not whether he is guilty of sandbagging or not. It is to question the approach of the governing body.
Quite so.
> am I missing something?
The column towards the right hand side titled "Odds".
> such extraordinary magnitude ... many millions-to-one
Please show your working! If such estimates are correct, of course you'll have a point. Help us to evaluate your claim.
> William of Occam would be able to see a straightforward explanation ...
Sure. But if he's a decent guy he'll also prefer to believe the best of his fellow man, and start with a presumption of innocence. So he'll want some realistic quantification of how strange a thing it is that he has found, and he'll want to know whether anything similar has happened to other players.
David
Apologies for the delay in moderating, I said I'd be travelling!
The column towards the right hand side titled "Odds".
OK, got you. It's certainly more than 100/1, but am I right that, working from Williams' odds (I also used this tool) we're still looking at 1200/1 or so? This still isn't anywhere near the kind of figures we're looking at as far as the Crockett rapidplay sequence is concerned. (As far as showing my working is concerned - future post, hopefully either tomorrow or Wednesday.)
he'll want to know whether anything similar has happened to other players
So he might and so do I, but I really don't believe that it's necessary for me to research everybody else's results ever in order to observe the great unlikelihood of a given sequence by a given player.
(it's also worth noting that in tennis, the absence of a draw makes an all-losing sequence much more likely than it would be in chess. Also that we're missing one set of odds, which would increase the overall odds a bit.)
Just for the record where someone asked has JH checked others results in the GP. I have. I checked all the top 20/25 active players over 3 years Crockett was winning GPs not 1 player was even remotely close to Crocketts high and low scores. Every player I checked had a Huge % of 30-70% scores. But Crocketts were all below 30% and above 70% with a very rare one in between.
Post a Comment