cheating detection

@Munich said in #50:

for what?
For high accuracy? Some one already shown some of your games.

it is not ok what you write, and certainly not funny.

Same metric you are using. Not that funny right? Exactly the point.

it is one thing to talk about how to detect players (high accuracy is not the metric on its own), and completely a different thing if you start accusing me of unfair play.

(Ignoring the pseduo-flame-war) Nothing really substantive to add on the original point, except to point out that if you noticed subtle timing stuff or whatever that really does point to a difference in consistency / the way cheat investigations work, this is possibly a correction of an error as well as possibly an introduction of an error -- in other words, if timing / action is different maybe false positive rate was too high and something was rightly corrected to have better confidence before action.

But it does seem unlikely that anything can be concluded about how the process is working / not working based on a single subjective experience.

@Munich said in #1:

I dont expect every report of mine will lead to a ban. But so many players have recently scored against me high accuracies with very low average centi-pawn loss - and none of those reported games resulted in a ban.

@Munich said in #54:

it is one thing to talk about how to detect players (high accuracy is not the metric on its own), and completely a different thing if you start accusing me of unfair play.

So, it is the metric, but not when it comes to you right?

Im not accusing you of anything. All I am saying is that there are a lot of great chess players in the site. High level play means high accuracy and very consistent level around a vast number of games. If you are going to report every single player that is good enough to have a good game against you, you will have to do a lot of reporting, and it is going to get worse if you keep improving, as the level of the opposite player will increase as well.

If you are going to come up with cheat accusations, bring proof. Reporting a player is not proof they are cheating, it is only proof that you cant take a defeat with grace (there is a better wording combination, but Ill leave it as is).

You have to learn to deal with defeat, the world does not spin around you. There are plenty of players that are way better than you and you will not be able to defeat them all.

@Munich said in #1:

But so many players have recently scored against me high accuracies with very low average centi-pawn loss

Please check the game analysis before jumping to any kramnik-type conclusions.
If you make a blunder, like getting a fork and losing a piece, then I might win against you with 90% accuracy and 15 CPL, even if you are 1000 higher rated. People tend to forget this factor.

If you got outplayed from start to finish by a lower-rated opponent + you didn't blunder obvious tactics .... then ok, you probably right about them cheating

I am not accusing anyone, so I can decide to report. And I did. I do not expect all to be positives, but none?

And indeed, after I opened this thread, shortly afterwards I got a positive, and the user got banned. This isnt jumping to conclusions. it what you all should do: if you think it is possible your opponent cheated, dont tell them during or after the game. Simply use the report option and let the admins decide. If we all do that, lichess will get rid of cheaters. Yes, they often respawn with fresh accounts (that is why I opt out and abort the game if my opponents have an unstable rating)., but if they get caught all the time, they will go somewhere else and stop trying to cheat here on lickess.

chess.com was quite slow in banning people, and when I changed to lichess, chess.com was quite infested with cheaters. I have no idea how the situation at chess.com is now, This was years ago, and I like that lichess reacts to reports - if the reported player was a cheater.

They have metrics that we dont have, such as they can see if a bot engine recommends moves (by highlighting the squares and point an arrow, so cheating in bullet is easily possible with a [redacted] many players still are not aware of that). It is also suspicious if the user does task-switching (focus of the open tab is on and off at each move), etc.

All I do is: if my opponent have very high accuracy, and very low centipawn loss per move, and the thinking time looks uniform, then I report the player, well aware that he could not be a cheater. I let the admins decide.

@Alientcp said in #56:

So, it is the metric, but not when it comes to you right?

Im not accusing you of anything. All I am saying is that there are a lot of great chess players in the site. High level play means high accuracy and very consistent level around a vast number of games. If you are going to report every single player that is good enough to have a good game against you, you will have to do a lot of reporting, and it is going to get worse if you keep improving, as the level of the opposite player will increase as well.

If you are going to come up with cheat accusations, bring proof. Reporting a player is not proof they are cheating, it is only proof that you cant take a defeat with grace (there is a better wording combination, but Ill leave it as is).

You have to learn to deal with defeat, the world does not spin around you. There are plenty of players that are way better than you and you will not be able to defeat them all.

a lot of accusations towards me. proof for what exactly, what are you talking about? where did I write what you accuse me of?
This is strawman argumentation and actually kind of trolling. None of it was what I was talking about. you dont even know whom or when I report.

Fully wrapped up, you live in your own world, nothing of what you say is related to what I said. you act as if I had reported players without proof. this thread is about cheating detection, but... you ignore that.

I received yesterday to more notices that 2 players were banned. I looked up the history of my games, and can thus see that these are most likely players I reported more than 1 month ago. The queue of games that are going to be checked might be quite long, so if you report someone, he might not get banned that quickly.

Or, the game that was reported was maybe not sufficient on its own, but other users reported then this TOS-violator, too, and together with those additional reported games, maybe that was finally sufficient evidence?

As always, if we report someone, we will not get told who of those reported is banned.
Since I have not reported that many users (but Alientcp says so, without knowledge, just for trolling), I can still find out in the history.

Approximately every third game I report leads to a ban.
Definitely most of those reported were rather fresh account with few games.