Farewell

Once again, this forum shows that most people on here have very basic, inaccurate views on the Lichess cheat detection system. The system here assesses much more than ACPL, overall accuracy and even move times.

Remember, cheaters don't always have to cheat in every game, every move. They don't even have to always play the best move. In fact, the stronger the player, the less he has to consult with the engine. This is why cheating is so hard to detect at top-level GM play. Hence the cheat detection is much more complex and considers much more variables and statistical data than most think...

@cashcow8 said it best in #101:
"I mean, how many of us have played a good game of chess, totally fairly, but then been worried we maybe played too well that nobody would believe us? So then, maybe, we run it through the engine and are "relieved" that actually we didn't play as well as we thought we had. But this is the fear we face - if we're not believed, we can be booted out with no way to return."

What if we played what we thought were good games of chess but not great but our worse fears come to fruition and we get 2 inaccuracies 0 mistakes and 0 blunders with an average centipawn loss of 7. This was a game that I played in Classical format 20 10 against an opponent 600 rating points less than mine in an arena that I late regged for so I wasn't playing for 1st place and I wasn't playing to gain rating points as most of the players I played ended up having a far less rating than myself. In this game, I won a pawn and the exchange without any great tactics or an amazing plan. My overmatched opponent made blunders and I took advantage of them and exchanged pieces until I got a winning endgame. I ended up making the most basic mate. with 2 rooks. I now wish I didn't play in this arena as I went undefeated but ended up with 2 draws one where I berserked against an opponent 400 rating points below me and I regretted decision that as not having an increment puts one in peril against an opponent who is playing with double the time and has a large increment. I offered a draw which was refused and somehow got a perp draw. I played one player my own rating and played an opening that I have played for many, many years. Although, I was white and quite familiar with opening I was quite happy to get a draw with repetition of moves. Long story short I have now been marked as a cheater on Lichess for trying my best in an arena with nothing to gain as I finished in 13th place and didn't gain many rating points. There is no appeal process. I was told by someone that if you don't cheat you will not be branded a cheater. Unfortunately, on Lichess that is not true. I have an account in good standiing on chess.com with a higher rapid rating. with many, many more games played there and many times I get rating points refunded for playing someone who cheated and I get destroyed by them like I am nothing while playing the same strength as I did in this arena. It sucks to lose to cheaters and it sucks to be falsely accused of being one. Only I know for sure that I didn't cheat in this arena. Only Pepe and his wife know for sure if he cheated. I feel bad if he is innocent but I don't know for sure for him. I can only speak for myself. I do think that these sites have to be totally sure if they are going to ban someone. I have total confidence in chess.com. I have no confidence in Lichess. I could write a lot more to prove my innocence but Lichess thinks they have a special program with AI and we have seen with ChatGPT that AI NEVER makes any mistakes. I will never play here again but I do hope that Lichess decides to regain the confidence of the players and not say that smart people are too stupid to understand and we can't tell you what evidence we have as that would make cheating too easy. Hopefully, something positive will come from this and we will not see so much doubt about someone like Pepe in the future.

@Robola99 said in #112:
> @cashcow8 said it best in #101:
> "I mean, how many of us have played a good game of chess, totally fairly, but then been worried we maybe played too well that nobody would believe us? So then, maybe, we run it through the engine and are "relieved" that actually we didn't play as well as we thought we had. But this is the fear we face - if we're not believed, we can be booted out with no way to return."
>
> What if we played what we thought were good games of chess but not great but our worse fears come to fruition and we get 2 inaccuracies 0 mistakes and 0 blunders with an average centipawn loss of 7. This was a game that I played in Classical format 20 10 against an opponent 600 rating points less than mine in an arena that I late regged for so I wasn't playing for 1st place and I wasn't playing to gain rating points as most of the players I played ended up having a far less rating than myself. In this game, I won a pawn and the exchange without any great tactics or an amazing plan. My overmatched opponent made blunders and I took advantage of them and exchanged pieces until I got a winning endgame. I ended up making the most basic mate. with 2 rooks. I now wish I didn't play in this arena as I went undefeated but ended up with 2 draws one where I berserked against an opponent 400 rating points below me and I regretted decision that as not having an increment puts one in peril against an opponent who is playing with double the time and has a large increment. I offered a draw which was refused and somehow got a perp draw. I played one player my own rating and played an opening that I have played for many, many years. Although, I was white and quite familiar with opening I was quite happy to get a draw with repetition of moves. Long story short I have now been marked as a cheater on Lichess for trying my best in an arena with nothing to gain as I finished in 13th place and didn't gain many rating points. There is no appeal process. I was told by someone that if you don't cheat you will not be branded a cheater. Unfortunately, on Lichess that is not true. I have an account in good standiing on chess.com with a higher rapid rating. with many, many more games played there and many times I get rating points refunded for playing someone who cheated and I get destroyed by them like I am nothing while playing the same strength as I did in this arena. It sucks to lose to cheaters and it sucks to be falsely accused of being one. Only I know for sure that I didn't cheat in this arena. Only Pepe and his wife know for sure if he cheated. I feel bad if he is innocent but I don't know for sure for him. I can only speak for myself. I do think that these sites have to be totally sure if they are going to ban someone. I have total confidence in chess.com. I have no confidence in Lichess. I could write a lot more to prove my innocence but Lichess thinks they have a special program with AI and we have seen with ChatGPT that AI NEVER makes any mistakes. I will never play here again but I do hope that Lichess decides to regain the confidence of the players and not say that smart people are too stupid to understand and we can't tell you what evidence we have as that would make cheating too easy. Hopefully, something positive will come from this and we will not see so much doubt about someone like Pepe in the future.

I want to clarify that I'm not saying you cheated, but I do admit that your perfect games raise some doubts. If pepe had as many perfect games as you, I would be a lot less skeptical.

Some bans make sense, and if tomorrow I legitimately play ten games in a row with perfect accuracy and mods bans me for them, well so be it. I could understand them I'd be the one false positive in a million (I'll prob. start playing OTB too!) but it makes sense.

It's a really difficult problem to solve, and tied to the easiness of playing/cheating wich is why I do think some people should get a special treatment: if you're streaming the game, filming yourself, showing your real identity and if any accusations could have some real world repercussions or reputation damage then yeah the threshold should be a little higher.

Public false positives hurt the community as much, if not more than actual cheating. It feeds into the existing paranoia, while causing people to lose trust into Lichess's ability to detect cheating which is a very dangerous combination.

@Hitsugaya said in #113:
> I want to clarify that I'm not saying you cheated, but I do admit that your perfect games raise some doubts. If pepe had as many perfect games as you, I would be a lot less skeptical.
>
> Some bans make sense, and if tomorrow I legitimately play ten games in a row with perfect accuracy and mods bans me for them, well so be it. I could understand them I'd be the one false positive in a million (I'll prob. start playing OTB too!) but it makes sense.
>
> It's a really difficult problem to solve, and tied to the easiness of playing/cheating wich is why I do think some people should get a special treatment: if you're streaming the game, filming yourself, showing your real identity and if any accusations could have some real world repercussions or reputation damage then yeah the threshold should be a little higher.
>
> Public false positives hurt the community as much, if not more than actual cheating. It feeds into the existing paranoia, while causing people to lose trust into Lichess's ability to detect cheating which is a very dangerous combination.

I just want to point out that Pepe berserked and played fast blitz games. When I play 3-0, I almost always make several blunders. I can't imagine rebounding from a mouseslip against a GM and beating one especially at that time level. So we are comparing apples and oranges. I didn't play a GM. I played players much weaker than me and when that happens and you take advantage of blunders and play basic chess, any decently strong player will play "perfect" games. Pepe is much better than me at faster time controls. My blitz rating on chess.com is 1750ish which is much lower than my blitz rating on Lichess. I made my insights public and my accuracy increases with longer time controls. I am pretty sure that is very common and everyone has that.

I watched his stream today and he seems very confident in his abilities. I don't know if he streamed these games. I wish that I was on a webcam while playing. I do agree that the harm of cheating for a streamer is very significant and causes a lot of harm to them and even the perception of a false positive would cause people to lose confidence in Lichess's ability to effectively cheat detect.

@Robola99 I can't tell just by looking at your profile. I had a look at your insights and my own (and I am not marked). Our centipawn loss rates at rapid and classic were about the same, at blitz mine were far better but that's probably because my idea of "blitz" is 5+5 or 5+3 rather than 3+0 which I don't play at all. And I've hardly played any blitz.

Surprised that both of us are more likely to gain rating points against slightly stronger opponents, and we both lose rating points playing weaker opponents.

Most of my insights were in variants with huge centipawn loss but that's no surprise as in anti-chess or crazyhouse you get massive swings on a single move you don't get in standard chess. In chess if you're -5.0 you're pretty much lost unless your opponent makes a major blunder, in Crazyhouse these are often turned around.

If they would at least tell the person whose account was marked that "we strongly believe you cheated in these games, and on these moves.." that would at least be something, although it's possibly an accumulation over many games.

Your account was marked, pepellou was directly closed so it's difficult to see his games over a long period for us to get our own opinions.

By the way, I would not want to be on the cheat detection committee either on here or on any other site. Would you?

I would be happy, however, to be an arbiter at a Hybrid Chess location. I'd also be happy to potentially help "develop" a hybrid chess "feature", or whatever is necessary to make it a possibility.

And I think it would be a possible idea to "restrict" some accounts to playing only hybrid chess, rather than ban them completely.

Can't read all of these wall of texts obviously. I'd just gently ask everyone to read my blog. Because you probably have nothing else to go with anyway. Can you defend someone as clearly innocent just because of their name and reputation? No you can't. There's a very high chance you're not serving the truth.

Can you always defend the mod decisions and claim they're never wrong? Well that's a double-edged sword too (I know successful appeals...)

Were you able to investigate the case on your own and prove it? Well, good luck with convincing others, chances are the people close to the banned person are too stubborn to listen. I know this from experience. Besides, within the space of Lichess communication, we have a policy of discretion anyway.

This is sad to hear :(
I hope Lichess would fix this soon and everything will bee alright again :)

I trust Pepellou so this makes me lose faith in the cheat detection done by lichess. explanations are needed my conclusion is that Lichess just sux at cheat detection.

well you are just naive, like the majority of people.

Surprised by the fact that many titled players were caught cheating?

i don't get how you can look at a person who had (1) chess.com close their account for violating their fairplay policy, and (2) lichess mark their account for tos violations – and still go "yes, i am 100% certain this person would never cheat, there is no doubt in my mind". how well do you have to know a person to be this certain? personally, i'd have to be this person myself.