Fair Play stats for September: 26,546 Fair Play closures (including 15 titled players)

These stats are from chesscom. After the Niemann incident where he is plastered up and down the news with his online escapades, one would think that people would cool it with cheating in online chess but no. Fifteen titled players caught, GTFO.

Fair Play stats for September:

26,546 Fair Play closures (including 15 titled players)

www.chess.com/article/view/chesscom-update-september-2022#FairPlay

@kaotic99 said in #1:
> These stats are from chesscom. After the Niemann incident where he is plastered up and down the news with his online escapades, one would think that people would cool it with cheating in online chess but no. Fifteen titled players caught, GTFO.
>
> Fair Play stats for September:
>
> 26,546 Fair Play closures (including 15 titled players)
>
> www.chess.com/article/view/chesscom-update-september-2022#FairPlay
Cheating is a big issue in online chess . Usually people cheat because they want to acgieve high rating in chess like 2800 or 3000 like super GM and IM . Chess.com is the no.1 site and i know people do cheat there in chess tournaments and simple gameplay . Thanks for posting .

Does Lichess publish their statistics for fair play closures, Or can they be found somewhere?

@userfriendly2 said in #3:
> Does Lichess publish there statistics for this? Or can they be found somewhere?

not so far, would be nice if they did though.

@kaotic99 said in #4:
> not so far, would be nice if they did though.

Yeah I admire chesscom publishing these stats - it highlights the issue and shows the hard work behind the scenes to address it.

It's also good in the interest of transparency. There is no reason to assume there are less cheaters here but would be interesting to see if the account closures match up.

> Does Lichess publish their statistics for fair play closures, Or can they be found somewhere?

The amount of account closures is a completely meaningless metric.

For example: A while ago, I, a single moderator, closed approximately 1000 accounts within one day. Did I work incredibly hard? Was it a great day for cheat detection? Not at all. I found a user who played a single digit number of games per account, then cheated in the last game and created a new account right after that. It was in fact quite embarrassing that they managed to get away with that strategy for so long.
Did cheating on the site decrease measurably after I banned this particular user? Not at all, a single user can only cheat so much. Even if they do create thousands of accounts, they cannot cheat on all of them at the same time.
In other words, a thousand closed accounts still had the same effect as closing the account of one user who sticks to their account.

If one closed account and a thousand closed accounts can have the same meaning, how meaningful could the total amount of closed accounts be?

@userfriendly2 said in #5:
> Yeah I admire chesscom publishing these stats - it highlights the issue and shows the hard work behind the scenes to address it.
>
> It's also good in the interest of transparency. There is no reason to assume there are less cheaters here but would be interesting to see if the account closures match up.

Actually, it doesn't really show anything : these stats are extremely easy to inflate. Chess.com can decide how "sensitive" their cheat detection will be.
They can decide to ban only the accounts for which they are 99.99% sure the player cheated, but then many cheaters will remain undetected. Or they can decide to ban every account for wich they are 60% sure they will ban more cheater but countless innocents too.

So a high number of Fair Play closures is easily obtainable : they just have to consider everyone even remotely suspicious as a cheater.
Most innocent people unfairly banned won't try to sue Chess.com. It's much simpler to create a new account.

Keep in mind I don't accuse Chess.com of closing more accounts than they should, I just say it's possible, and even in their interest to do so since it allow them to tell everyone they have the best cheat detection in the world.

@anonmod said in #6:
>
> If one closed account and a thousand closed accounts can have the same meaning, how meaningful could the total amount of closed accounts be?

Just one example doesn't mean such statistics are totally meaningless. If the average number of accounts/cheater is fairly stable over time (which doesn't seem to be far fetched) then the number of closed accounts over time give an indication on whether there are more or less cheaters over time (even if one doesn't know the average number of accounts/cheater).

Now, if in one month, all the cheaters created a 1000 accounts, and in the next month, no cheater created more than one account, then yeah, the statistic would be meaningless.

I'd argue, the number of closed accounts isn't totally meaningless, but on the same time I acknowledge people will give too much value to that. Why not publish them with a disclaimer?

> Why not publish them with a disclaimer?

I would argue that publishing misleading statistics with a disclaimer is unethical when you have the option of not publishing misleading statistics in the first place.

Anyway, it is not my decision.

@anonmod said in #9:
> I would argue that publishing misleading statistics with a disclaimer is unethical when you have the option of not publishing misleading statistics in the first place.
>
> Anyway, it is not my decision.
and chess.con certainly have precedent in that.