Questions About Cheaters

The other day I played a casual game with a player that had a 93% win rate at Rapid and a record of 80 wins in a row. I'm not saying they cheated in that particular game with me, however it seems suspicious at best.

Because it was a casual game should it matter if players cheat?

Should I report suspiciously high win rate percentages ?

What is a high win rate percentage?

How do I use an engine to investigate a suspicious game?

I wish I could just filter out players in casual games if their win rate was much higher than mine in that time control because it would eliminate a lot of stress and wasted time of looking up stats.

Call me paranoid but I'm suspicious of win rates over 65% because it seems like something is being manipulated.

@Sacmaniac #1
I don't know that I have answers to your questions, but I do have some observations:

1) I played against a person who was new to Lichess and had 59 wins, 1 draw, 0 (zero) losses -- a win rate of 99%. They had a provisional rating of 1500 because they never played rated games. They weren't cheating, but they were effectively sandbagging by intentionally (as they admitted) concealing their strength.

2) Yes. Cheating always matters. It's always wrong.

3) I suspect Lichess admin protocols check for high win rates already, but I also suspect that there simply are some really strong players who enjoy playing against much weaker opponents.

4) My win rate is over 68%, and I don't really feel like I'm that strong a player. So a suspiciously high win rate would be quite a bit higher than that, in my opinion.

5) One time I tried looking at a player's stats and at some of their games after what seemed like suspiciously strong play in a series of games with them, but I didn't see any clearcut pattern. I think maybe I just didn't play well. One thing that does suggest illegal use of an engine during play is if a player continually plays the very best move -- especially a really subtle move when the second-best move was obvious and nearly as good. But maybe someone using an engine has access to second-best move, third-best move, etc., and can select from them -- I don't know.

6) Filtering based on win rate is something you could propose in the Lichess feedback forum.

7) My win rate is over 65% and I'm not manipulating anything. Although I don't play many rated games (because I feel pressure to win but I am frequently interrupted and need to be able to step away or resign, even in an equal or winning position), I do play enough rated games to have a rating that's not completely bogus, so my opponents have at least some idea what to expect from me. But I agree that a 93% win rate and 80 wins in a row is, um, remarkable.

@Foolacious said in #2:
>
Woohoo, I'm one who's beaten you!: lichess.org/4y9LvgwS#108
Just ignore 3/4 of the takebacks

@Sacmaniac said in #1:

> Because it was a casual game should it matter if players cheat?
>
Yes, cheating in casual is forbidden as well as in rated. Feel free to report suspicious games.

> Should I report suspiciously high win rate percentages ?
>
We prefer if you report a game or multiple games and explain your reasoning. You can use any public information to investigate and add it to your report. Although reporting a player based only on an apparent suspicious win streak, without mentioning a game and the reasoning behind, is not something I would advice you to do.

> How do I use an engine to investigate a suspicious game?
>
You can use our computer analysis to have a general idea of the accuracy of the player in the game. You can also use the local analysis to investigate multiple lines.

> Call me paranoid but I'm suspicious of win rates over 65% because it seems like something is being manipulated.
>
We encourage players to report any suspicious activity, misbehavior or gameplay, but we also recommend not to report every single player you have lost to (the probability that you have played x cheaters in a row in y time is close to zero) and also not to report players based on a personal feeling, like "I think that player is cheating" or " something seems off", because that is not how the report feature is supposed to be used.
For more informations about reports please see lichess.org/page/report-faq.

@The_Chess_Star_100 #3
"Just ignore 3/4 of the takebacks"
But 6 isn't evenly divisible by 4! ;-)

@Foolacious said in #5:
> @The_Chess_Star_100 #3
> "Just ignore 3/4 of the takebacks"
> But 6 isn't evenly divisible by 4! ;-)
Oh yeah xD

@Sacmaniac said in #1:
> Call me paranoid but I'm suspicious of win rates over 65% because it seems like something is being manipulated.

It can be a stronger player that enjoys playing with newbies ^^

My victory rate is 38%, and defeat rate 55%. I'm the opposite, I like playing against stronger players, very often :)

@Sacmaniac said in #1:
> The other day I played a casual game with a player that had a 93% win rate at Rapid and a record of 80 wins in a row. I'm not saying they cheated in that particular game with me, however it seems suspicious at best.
>
> Because it was a casual game should it matter if players cheat?
>
> Should I report suspiciously high win rate percentages ?
>
> What is a high win rate percentage?
>
> How do I use an engine to investigate a suspicious game?
>
> I wish I could just filter out players in casual games if their win rate was much higher than mine in that time control because it would eliminate a lot of stress and wasted time of looking up stats.
>
> Call me paranoid but I'm suspicious of win rates over 65% because it seems like something is being manipulated.

I think the sample size should be considered as well. If they have 1000 games and a 90% win rate, well they are either sandbagging or something fishy is going on.

It could also be the case that they play in tournaments here (maybe not always finishing the tournament) and are being disproportionally paired against opponents who are a few hundred points beneath them. Or they could be just very choosy about their opponents, rather than using the quick pairing function. Someone could look at my classical win rate and see no losses and be suspicious, but in my defense i have only played 8 classical games (7 wins, 1 draw) and a few of those wins were the first few rounds of a swiss tournament where I got paired against 1800 level opponents. I have also been using the lobby to find longer games, as I dont always get paired too quickly at 30/20 or whatever the classical quick pairing is. So usually I just grab whatever reasonably rated opponent is available if I am happy with the time control.

@Le_Patzer83 Their sample size was over 3000 casual games with a 93% win rate and a 1500 provisional rating.

Now, I realize that they could be a master only playing much weaker players because that can be fun too.

However, speaking about the player admitting to cheating, I can't even fathom how this can go on for years and so far I've noticed they are still active and have not been banned. Either they are a troll or a very good cheater.

A guy at chess.com said the engine won't really help you at catching cheaters, sure low acp is a thing but yeah, we have the tools for catching cheaters and you don't really have a clue(my interpretation of his words).

And thats why it should be enough to ask lichess :"is this guy honest or not?" when sending in cheat reports. Imo.