Chess.Com getting sued.

[JustinHorton]

I had an argument with Talk Talk, which went to 48 pages of emails (many of them equally as unhelpful as that one) - you might be able to beat that.

I did on one occasion get past fifty emails in a complaint saga (Transport for London, for the record) and I've had exchanges with the Nat West which have literally gone on for more than a year, but I don't think I have quite that amount of energy to waste on these people.

.
I'm sure Ken Regan will be interested to know that they know better than he does.

Obviously that's a matter for him, but yes, it's pretty insulting both to his degree of expertise and to the amount of work he was prepared to put in.

[Roger Lancaster]

Have to say, as a totally disinterested party, that this doesn't appear to reflect too well on Chess.com.

[Simon Brown]

Borrowing an idea from another thread, @chesscom has 107,000 followers who may be interested to hear about this injustice.

[Kevin Thurlow]

I had a look at chess.com and they stated on their site that they use statistical analysis to stop computer use, but in the email to Justin they said they were "100 % certain". That suggests they don't know a lot about statistics. Admittedly, the person sending the email was probably the office cat or someone in marketing. My technique with intransigent and corrupt companies is to start demanding money from them, which seems to trigger an upgrade from the office cat to someone slightly less stupid, although it doesn't always work.

[Roger Lancaster]

While sympathetic towards Justin, I'm also mindful that there are presumably many others where Chess.com have closed accounts because it is "100% certain of unfair play" [their words] but who do not have the time, ability or resources to draw attention to the matter in the way Justin has.

As to the phrase in my parentheses, I can understand Chess.com [or anyone else] concluding from computer analysis that it is highly probable that someone is cheating. But any statistical analysis will always leave open a small but finite possibility that there is no cheating. I'm all in favour of cheats being rooted out but the fact, which Chess.com does not acknowledge, that analysis has its limitations will always be a difficulty.

NB: While I've typed this, Kevin has made much the same point over certainty.

[JustinHorton]

I had a look at chess.com and they stated on their site that they use statistical analysis to stop computer use, but in the email to Justin they said they were "100 % certain". That suggests they don't know a lot about statistics.

Well, just as likely it suggests that they basically operate a practice whereby they won't discuss or reopen decisions because to do so would be to accept the possibility of flaws in their detection model, which in its turn would hurt their business model.

[Kevin Thurlow]

"Well, just as likely it suggests that they basically operate a practice whereby they won't discuss or reopen decisions because to do so would be to accept the possibility of flaws in their detection model, which in its turn would hurt their business model."

Yes, that as well.

[Ian Thompson]

I had a look at chess.com and they stated on their site that they use statistical analysis to stop computer use, but in the email to Justin they said they were "100 % certain". That suggests they don't know a lot about statistics.

Well, just as likely it suggests that they basically operate a practice whereby they won't discuss or reopen decisions because to do so would be to accept the possibility of flaws in their detection model, which in its turn would hurt their business model.

But possibly not as much as their business model would be hurt by making claims that cannot be true and by making defamatory comments about users of their site.

Anyway, I believe Justin has broken one of their fair play rules - "Do not ... do any analysis of your games in progress", on the grounds that his moves do not look entirely random.

[JustinHorton]

I had a look at chess.com and they stated on their site that they use statistical analysis to stop computer use, but in the email to Justin they said they were "100 % certain". That suggests they don't know a lot about statistics.

Well, just as likely it suggests that they basically operate a practice whereby they won't discuss or reopen decisions because to do so would be to accept the possibility of flaws in their detection model, which in its turn would hurt their business model.

But possibly not as much as their business model would be hurt by making claims that cannot be true and by making defamatory comments about users of their site.

Well yes, if more people were able to go down the road of Henry Despres.

[Paul Cooksey]

Anyway, I believe Justin has broken one of their fair play rules - "Do not ... do any analysis of your games in progress", on the grounds that his moves do not look entirely random.

I was tempted to pull on a similar thread in the anti-cheating forum by recommending all first moves except 1.g4, thereby condemning all of them to play the Grob forever to avoid cheating by using my advice. Seems a suitable punishment.

[Jonathan Bryant]

But possibly not as much as their business model would be hurt by making claims that cannot be true and by making defamatory comments about users of their site.

FWIW I rather suspect that, on the contrary, it’s precisely the intention of supporting the business model which is driving chess.com’s behaviour.



I’ve been wanting to contribute for a while but extreme work-related zeitnot has prevented it. I have discuss things a little with Justin in private.

One of the arguments is it is better to ban a few innocent people than have cheaters on the site.

My feeling is that this is exactly what chess.com think. I mean, I can’t imagine they’d could be so stupid to really believe they can "100% certain" on statistical analysis alone with such a small sample.

... Often a very weak player ....

This, I feel, is the fundamental issue.

We - regular club and tournament players - simply aren’t chess.com’s target market. Club players tend to think of weak players as under 100 ECF strength. Perhaps down to 80. Outside of club players who play regularly face to face, though, that level of ability would be considered extremely talented. It’s exactly this group that chess.com target. And it makes sense. After all there are so many many more of that group than us.

All of which means that chess.com is populated for the most part by weak players. Eg. the top thread on the openings forum when I first intended to write this post was
https://www.chess.com/forum/view/chess- ... e4-e5-2-g4

in which the poster earnestly asks for feedback on his idea of playing 2 g4 in the open games.


Which is all very well. there’s nothing wrong with being a relatively inexperienced not particularly strong player. Nothing at all. It’s just this does sometimes create a bit of a culture clash.


E.g. when I started playing there a little while ago and won my first 16 games in a row. To my mind I was winning easily against people who barely knew the moves. To some people I was on obvious cheat (I got a message saying so). Because by the standards what i was doing wasn’t possible.



And Justin, needless to say, is a much stronger player than me. SO I’d expect him to score wins easily against even higher rated players than I did.




So what do chess.com do? cheating is a real problem. During my 30 or so correspondence games in total I had strong suspicions about two opponents that they were cheating. Not by statistical analysis but because their rate of play or the type of moves they made just didn’t feel right.

So chess.com have to crack down otherwise they have no market.

But what if they end up banning genuine stronger players in consequence? I think they have decided it’s a price worth paying, economically.

Because when it comes down to it, there’s relatively few of us. And a lot of casual players. An awful lot.



That, in any event, is what I believe has happened.


Because, when it comes down to it, I can’t believe anybody could be so dense as to be believe that a handful of moves in a handful of games is enough data to draw any conclusions - let alone be 100% certain - from statistical analysis.

[Christopher Kreuzer]

pour encourager les autres

[NickFaulks]

I can’t imagine they’d could be so stupid to really believe they can "100% certain" on statistical analysis alone with such a small sample.

I can easily imagine that. Statistics is a subject where otherwise intelligent people can have the most extraordinary blind spots.

[Jonathan Bryant]

Statistics is a subject where otherwise intelligent people can have the most extraordinary blind spots.

Agreed Nick. But chess.com - chessers and computer cheating in general - don’t really have this excuse.

E.g. the problem of match-up analysis being corrupted by knowing when a player’s opening book moves end has been discussed for around 10 years now.

The issue chess.com have is not about statistics per se and it isn’t a blind spot.



All that said, add me to the list of people who find the policy of "We are against cheats unless they admit it in which case we’re happy they continue playing" to be totally baffling.



Anyhoo, that’s probably going to me done on the subject for a little while. Other than to say i won’t be playing on their website again.

[NickFaulks]

But chess.com - chessers and computer cheating in general - don’t really have this excuse.

You have never been exposed to a meeting of Anti Cheating Commission ( as was ). The rationale does not go beyond "computer says 6.5, which means he is 100% cheating" - probably followed by some guff about the number of atoms in the universe. You ask what is this 6.5 number, where does it actually come from and how is it interpreted? Well, it's 6.5 innit, everyone know what that means.