Carlsen resigns on move 2

[Ian Thompson]

People have often been banned on the strength of one game, eg when caught with an engine running in the background, and I seem to remember a case where a female - cannot recall whether a child or an adult - was disqualified for looking away from her screen too often.

I assume you mean Iulija Osmak.

Edit: A more informative link.

[Mick Norris]

I see she played in Chennai in the Olympiad, so presumably FIDE are happy with things now

[JustinHorton]

People have often been banned on the strength of one game, eg when caught with an engine running in the background

Although this is not what is meant by being banned on the strength of one game

[Geoff Chandler]

Hi Tim,

I was of course referring to Roger's; ...the game-score fallen into the hands of chess.com or lichess.'
and them scanning one game and banning you. It has to be more games.
You can however get banned from any chess site for one wrong post in any forum.
It would have to be particularly offensive. I know of a case at RHP were a player was banned for
such behaviour and everyone thought they had been cheating. I was practically alone
in saying the games showed no signs of silicon use and I suspected a stupid post as the person
sometimes sailed close to the wind with insults etc. (they must have been warned in the past.)

The player concerned joined under another name and sent me the banning PM. (language and insults)
I then passed on the PM (with permission from the player.).

Regarding 'Your Cheating Eyes.' A vid on the net regarding cheating is Nakamura and Petrosian's eyes.

https://www.youtube.com/watch?v=Gq4Rs9-rZNI

If you have not seen it give it a go. There are some light-hearted moments in a serious matter.

Edit: we crossed posted Justin.

[Roger Lancaster]

Hi Roger,

They cannot, or should not ban anyone on the strength of one game.

You don't have to but it would sure make a pleasant change to see some chess in this thread.
Show us the T.N. perpetual. if you do and use it online at a later date and someone cries cheat
then you can send them here. Your TN will have been date stamped previous to the online game.
(and it might appear in books as 'The Chandler Variation.' Yes it's your TN but sticking here is my idea.)

Geoff. I deliberately didn't state whether the game was played online or OTB because there was another facet that interested me, namely, given that cheating is accepted (both empirically and, I think, by Ken Regan's analyses) as occurring more frequently online, would Ken's methods have attached different probabilities to the hypothesis that cheating occurred in such a game depending on whether it was played online or OTB? Bayes' Theorem, perhaps, for those statistically inclined? (As to your request, the game ended in a 28-move perpetual which wasn't a TN. It was only the prepared variation, which didn't occur, which was - as far as I know - a TN and I'm keeping that as a future surprise for someone).

[Matt Bridgeman]

Danny Gormally has an interesting link to analysis by FM Yosha Iglesias, which seemingly unearths a number of perfect OTB games by Hans according to the Let’s Check function of Chessbase. This being a number of games separate from the Carlsen game and seemingly at a level beyond Fischer and Carlsen at their peak. There’s a link to the YouTube video and the spreadsheet of results. According to the analysis the highest ever recorded result is supposed to be 98% by the discredited Sébastien Feller, yet Hans has a number of games at 100%.

https://m.youtube.com/watch?v=jfPzUgzrOcQ

https://docs.google.com/spreadsheets/d/ ... k/htmlview

[John Upham]

Danny Gormally has an interesting link to analysis by FM Yosha Iglesias, which seemingly unearths a number of perfect games by Hans according to the Let’s Check function of Chessbase. This being a number of games separate from the Carlsen game and seemingly at a level beyond Fischer and Carlsen at their peak. There’s a link to the YouTube video and the spreadsheet of results. According to the analysis the highest ever recorded result is supposed to be 98% by the discredited Sébastien Feller, yet Hans has a number of games at 100%.

https://m.youtube.com/watch?v=jfPzUgzrOcQ

https://docs.google.com/spreadsheets/d/ ... k/htmlview

I am somewhat of a novice on this subject but perhaps someone can explain....

When you say 100% does that suppose that the player is allegedly using the same engine as the one used for the analysis for the cheat detection process?

I guess in many middle game positions different strong engines will come up with different top move suggestions.

Thanks for any explanations

[Matt Bridgeman]

Danny Gormally has an interesting link to analysis by FM Yosha Iglesias, which seemingly unearths a number of perfect games by Hans according to the Let’s Check function of Chessbase. This being a number of games separate from the Carlsen game and seemingly at a level beyond Fischer and Carlsen at their peak. There’s a link to the YouTube video and the spreadsheet of results. According to the analysis the highest ever recorded result is supposed to be 98% by the discredited Sébastien Feller, yet Hans has a number of games at 100%.

https://m.youtube.com/watch?v=jfPzUgzrOcQ

https://docs.google.com/spreadsheets/d/ ... k/htmlview

I am somewhat of a novice on this subject but perhaps someone can explain....

When you say 100% does that suppose that the player is allegedly using the same engine as the one used for the analysis for the cheat detection process?

I guess in many middle game positions different strong engines will come up with different top move suggestions.

Thanks for any explanations

Let’s Check I think automatically stores the deepest engine analysis across multiple engines that pass through the Chessbase servers I think, and will update as deeper analysis appears. Presumably someone will fire all Carlsen’s games through the same process soon and there will be a spreadsheet to compare to Hans’ OTB one.

The interesting guide chart for making sense of the percentages has;

98% Sébastien Feller in Paris 2010
72-75% Correspondence World Champion (pre-engines)
72% Bobby Fischer at peak 20 game winning streak
70% Carlsen at best
69% Kasparov at best
62-67% Super GM’s
57-62% Normal GM’s

[Geoff Chandler]

Thanks Roger,

Again the one game online or OTB would not be enough.
If you were 'hot' they would look for further evidence and may stumble upon a game
where you have had that position before and NOT played your TN.
Then they would have to assume that you had run a computer over that game when it was finished,
saw the better move and saved it for a rainy day. Noting illegal in good preparation.
Let us know when you use the TN.

I wonder how many potential TN's Ken Regan's computer has discovered when going though the thousand of games.
I'm now waiting for the day when Ken starts playing on chess.com, uses these TN's and gets banned.

[Ian Thompson]

Danny Gormally has an interesting link to analysis by FM Yosha Iglesias, which seemingly unearths a number of perfect games by Hans according to the Let’s Check function of Chessbase. This being a number of games separate from the Carlsen game and seemingly at a level beyond Fischer and Carlsen at their peak. There’s a link to the YouTube video and the spreadsheet of results. According to the analysis the highest ever recorded result is supposed to be 98% by the discredited Sébastien Feller, yet Hans has a number of games at 100%.

https://m.youtube.com/watch?v=jfPzUgzrOcQ

https://docs.google.com/spreadsheets/d/ ... k/htmlview

Let’s Check I think automatically stores the deepest engine analysis across multiple engines that pass through the Chessbase servers I think, and will update as deeper analysis appears.

So what's the score going to be if you have a game that follows known theory for many moves, your opponent then blunders and the game's over after a few accurate moves to finish him off?

[Matt Bridgeman]

Danny Gormally has an interesting link to analysis by FM Yosha Iglesias, which seemingly unearths a number of perfect games by Hans according to the Let’s Check function of Chessbase. This being a number of games separate from the Carlsen game and seemingly at a level beyond Fischer and Carlsen at their peak. There’s a link to the YouTube video and the spreadsheet of results. According to the analysis the highest ever recorded result is supposed to be 98% by the discredited Sébastien Feller, yet Hans has a number of games at 100%.

https://m.youtube.com/watch?v=jfPzUgzrOcQ

https://docs.google.com/spreadsheets/d/ ... k/htmlview

Let’s Check I think automatically stores the deepest engine analysis across multiple engines that pass through the Chessbase servers I think, and will update as deeper analysis appears.

So what's the score going to be if you have a game that follows known theory for many moves, your opponent then blunders and the game's over after a few accurate moves to finish him off?

It’s really hard to say isn’t it. Will the results be commonly replicated by other players or are they really out there on their own? Hans has some scores from OTB games at the 2021 US Junior Champs coming in at 100, 98, 85, 81, 80 and 72 for example. It would be interesting to know if anyone has ever done that before at the US Junior Championships?

[Tim Harding]

Hi MJM,

I think the reference made regarding fake hearing aids was they could easily be disguised as a radio receiver.
It has been noticed I've been turning the telly up quite loud of late so I am probably heading in that direction.
(I'm falling to bits, my body is being held together by over worked pieces.)
As well as searching toilets for mobiles (If they find one they should just remove the battery and put it back.)
Arbiters now have to sneak up on players not wearing their hearing aids and burst balloons behind them
to see if they really are deaf.

Arbiter Handbook, page 49.
Look out for any player wearing a hearing aid, (receiver) a glass eye (transmitter) with a walking stick (ariel).

Sounds like there is a black market for devices used for cheating in chess.

This isn't actually in the Arbiter Manual 2022, certainly not on page 49 which doesn't deal with anti-cheating.
I think Geoff was trying to make some kind of weak joke rather than providing an actual quotation.

[Roger Lancaster]

Thanks Roger,

Again the one game online or OTB would not be enough.
If you were 'hot' they would look for further evidence and may stumble upon a game
where you have had that position before and NOT played your TN.
Then they would have to assume that you had run a computer over that game when it was finished,
saw the better move and saved it for a rainy day. Noting illegal in good preparation.
Let us know when you use the TN.

I wonder how many potential TN's Ken Regan's computer has discovered when going though the thousand of games.
I'm now waiting for the day when Ken starts playing on chess.com, uses these TN's and gets banned.

Hi Geoff. I won't prolong this too much as it's some way removed from the issue of Carlsen's resignation. In fact, the 28-move game was played OTB with zero toilet breaks and no suspicious circumstances. But suppose I tomorrow play an identical game (or, worse still, the TN) online and rapidly. Since an anti-cheating program won't know I've analysed the line in depth, it will inevitably regard the sequence of moves as suspicious. That's no reflection on the statistical methods as such but of the fact that the program, through no-one's particular fault, was unaware of that key piece of information. That's one of the problems of applying mathematical techniques to human behaviour.

[Matt Bridgeman]

Yosha makes an interesting point on Twitter;

‘A clever cheater can outsmart the tools they know about, like avg centipawn loss, SpyPGN, chesscom CAPS, or Regan's ROI, but they can't outsmart the tools they don't know.

It turns out that almost no one knew about the Chessbase Let's Check Analysis.’

[Ian Thompson]

Yosha makes an interesting point on Twitter;

‘A clever cheater can outsmart the tools they know about, like avg centipawn loss, SpyPGN, chesscom CAPS, or Regan's ROI, but they can't outsmart the tools they don't know.

It turns out that almost no one knew about the Chessbase Let's Check Analysis.’

I don't understand. As I understand it, Let's Check is just another engine with a large database of previously analysed positions. How would having access to it during a game be any different from having access to some other engine?