- Blind mode tutorial
lichess.org
Donate

Cheaters getting better at cheating

The only thing you can do is report the players

The only thing you can do is report the players

@aescht said in #3:

Can you link two or three games when you thought the opponent is cheating?
Please don't encourage others to express suspicions in public. This is not allowed! If you are suspecting an opponent of cheating, then you can report the opponent.

@aescht said in #3: > Can you link two or three games when you thought the opponent is cheating? Please don't encourage others to express suspicions in public. This is not allowed! If you are suspecting an opponent of cheating, then you can report the opponent.

Lichess moderators care infinitely more about people complaining about cheats, then the cheating itself.

99% of people on this site are anonymous. So, what difference does it make is someone accuses someone else? Nobody's reputation is besmirched. It literally does not matter.

What public accusations do achieve, however, is forcing Lichess to actually do something with evidence (which they don't want to do).

Private reports are ignored. Public reports cannot be ignored. So, they have a policy of prohibiting public reports so that they can continue .... to do nothing.

Lichess moderators care infinitely more about people complaining about cheats, then the cheating itself. 99% of people on this site are anonymous. So, what difference does it make is someone accuses someone else? Nobody's reputation is besmirched. It literally does not matter. What public accusations do achieve, however, is forcing Lichess to actually do something with evidence (which they don't want to do). Private reports are ignored. Public reports cannot be ignored. So, they have a policy of prohibiting public reports so that they can continue .... to do nothing.

Guys I respectfully disagree with you. I have played here a lot for several years and I hardly ever encounter cheaters. There are just some very good players here and even lower rated players can play exceptional games occasionally. Learn to accept losing to better players or even weaker players. They are not all cheating.

Guys I respectfully disagree with you. I have played here a lot for several years and I hardly ever encounter cheaters. There are just some very good players here and even lower rated players can play exceptional games occasionally. Learn to accept losing to better players or even weaker players. They are not all cheating.

@ColonelMusgrave said in #7:

@sheckley666

Because this would take a team of computer engineers and analysts that would rival most tech companies in silicon valley. Nobody is doing this -- certainly not with a team of unpaid volunteers.

That is just wrong. While this is a difficult problem indeed, a team a few people can definitely tackle this, especially as there are already published worked on related topics.

The real problem is that any tool to find more cheaters that are more subtles that playing every top engine moves in exactly 10 seconds per move when they are 1200 rated are statistical in nature. With any statistical procedure, you have to set a boundary: if you want to catch every cheater, the simpler is to flag all players as cheaters. The problem is that you now also flagged all legitimate players as cheaters, also known as false positives.
If you want to never flag a legitimate player as a cheater, then you simply never flag any player as a cheater. In that case, obviously all cheaters get through.

You will allways have to make a tradeoff between these two extremes: the more you want to avoid flagging legitimate players as cheaters and the more real cheaters will go through. As it would be disastrous for a website to flag legitimate players as cheaters, they obviously have to be conservative with how suspicious a players needs to be before being flagged as a cheater.

The fact that a website can't afford to ban legitimate players for flagging them as cheaters (or at least not often enough for this to become a known problem), they will generally set the bar quite high for their algorithms to flag a user as potentially cheating for review by their staff. (Because to avoid this flase positive issue, I'm pretty sure that appart from some trivial cases they have all cases reviewed by someone). the problem is the same here: if you set the bar too low, your team will have to be gigantic to go through all the reports.

On a side note, this is one the reasons most websites offer premium plans to titled players. The better a player and the more likely he can be flagged as cheater by mistake. By offering premium plans to titled players, they have to identify themselves, which means they won't be flagged as cheaters simply for playing way better than normal users. This could also serve a purpose by allowing an algorithm to learn to discriminate between good player behaviors and cheaters if they want to use this type of learning, but I don't know if they do. In theory, this gives some labelled game generators so it can also be helpful for this.

@ColonelMusgrave said in #7: > @sheckley666 > > Because this would take a team of computer engineers and analysts that would rival most tech companies in silicon valley. Nobody is doing this -- certainly not with a team of unpaid volunteers. That is just wrong. While this is a difficult problem indeed, a team a few people can definitely tackle this, especially as there are already published worked on related topics. The real problem is that any tool to find more cheaters that are more subtles that playing every top engine moves in exactly 10 seconds per move when they are 1200 rated are statistical in nature. With any statistical procedure, you have to set a boundary: if you want to catch every cheater, the simpler is to flag all players as cheaters. The problem is that you now also flagged all legitimate players as cheaters, also known as false positives. If you want to never flag a legitimate player as a cheater, then you simply never flag any player as a cheater. In that case, obviously all cheaters get through. You will allways have to make a tradeoff between these two extremes: the more you want to avoid flagging legitimate players as cheaters and the more real cheaters will go through. As it would be disastrous for a website to flag legitimate players as cheaters, they obviously have to be conservative with how suspicious a players needs to be before being flagged as a cheater. The fact that a website can't afford to ban legitimate players for flagging them as cheaters (or at least not often enough for this to become a known problem), they will generally set the bar quite high for their algorithms to flag a user as potentially cheating for review by their staff. (Because to avoid this flase positive issue, I'm pretty sure that appart from some trivial cases they have all cases reviewed by someone). the problem is the same here: if you set the bar too low, your team will have to be gigantic to go through all the reports. On a side note, this is one the reasons most websites offer premium plans to titled players. The better a player and the more likely he can be flagged as cheater by mistake. By offering premium plans to titled players, they have to identify themselves, which means they won't be flagged as cheaters simply for playing way better than normal users. This could also serve a purpose by allowing an algorithm to learn to discriminate between good player behaviors and cheaters if they want to use this type of learning, but I don't know if they do. In theory, this gives some labelled game generators so it can also be helpful for this.

A Lichess Moderator said in #12:

Please don't encourage others to express suspicions in public. This is not allowed! If you are suspecting an opponent of cheating, then you can report the opponent.

Ok I see and accept this.

It's just that I almost never encounter any cheaters at ~2000 (or I am not aware of them) and was wondering if there is so much cheating in the 1200 rating range or if people are just paranoid. So I was curious and asked to see some games.

A Lichess Moderator said in #12: > Please don't encourage others to express suspicions in public. This is not allowed! If you are suspecting an opponent of cheating, then you can report the opponent. Ok I see and accept this. It's just that I almost never encounter any cheaters at ~2000 (or I am not aware of them) and was wondering if there is so much cheating in the 1200 rating range or if people are just paranoid. So I was curious and asked to see some games.

@DoctorFuu

"[A] team a few people can definitely tackle this ..."

A "few" people? There are millions of players, and tens of millions of games. A team of a "few" people isn't catching anything. If the algorithm isn't catching them, they're not being caught.

Nobody is sitting around at Lichess headquarters digging through hundreds of thousands of games in their spare time, seeing as how allegedly everyone working for this site is a volunteer.

@DoctorFuu "[A] team a few people can definitely tackle this ..." A "few" people? There are millions of players, and tens of millions of games. A team of a "few" people isn't catching anything. If the algorithm isn't catching them, they're not being caught. Nobody is sitting around at Lichess headquarters digging through hundreds of thousands of games in their spare time, seeing as how allegedly everyone working for this site is a volunteer.

@Lilsnottynose said in #14:

Guys I respectfully disagree with you. I have played here a lot for several years and I hardly ever encounter cheaters. There are just some very good players here and even lower rated players can play exceptional games occasionally. Learn to accept losing to better players or even weaker players. They are not all cheating.

This also has to do with rating. Look at the discrepancy between your rating and OPs. I don't remember being suspicious of an opponent since when I crossed 1800 rating. Unfortunately, there are more cheaters at lower ratings. My guess is simply that cheating in normal random games on internet has close to zero appeal once you invested enough time to actually find the game interesting. So, yeah, more casual players are more likely to have cheaters at their rating than other players. This is very sad because this is also the rating where it's most important to have a fair and fun environment for players to stick with the game.

Note that I'm NOT saying there are no cheaters higher up, just much less in proportions. Though we all remember Mr Peepee in their Pampers.

I also often play in anonymous, and tbh I am surprised by the small number of cheaters (I would estimate maybe 10%? Less than 15% for sure). I'm also surprised by the number of good players who play on anonymous and I often get crushed by people who I am 99% sure are not cheaters. Also, in anonymous at least, the majority of cheaters are only using an opening explorer but no engine I think. It's allways weird when a players takes 10 seconds per move to play 20 moves of the main line of some rare subvariation, then starts to play like absolute garbage haha.

I want to join you on your last two sentences: everyone has the right to have a good game, and not everyone playing well or above their rating is cheating. We all do sometimes play above our ratings: that's how we gain rating points. Most players better than us are NOT cheaters.

@Lilsnottynose said in #14: > Guys I respectfully disagree with you. I have played here a lot for several years and I hardly ever encounter cheaters. There are just some very good players here and even lower rated players can play exceptional games occasionally. Learn to accept losing to better players or even weaker players. They are not all cheating. This also has to do with rating. Look at the discrepancy between your rating and OPs. I don't remember being suspicious of an opponent since when I crossed 1800 rating. Unfortunately, there are more cheaters at lower ratings. My guess is simply that cheating in normal random games on internet has close to zero appeal once you invested enough time to actually find the game interesting. So, yeah, more casual players are more likely to have cheaters at their rating than other players. This is very sad because this is also the rating where it's most important to have a fair and fun environment for players to stick with the game. Note that I'm NOT saying there are no cheaters higher up, just much less in proportions. Though we all remember Mr Peepee in their Pampers. I also often play in anonymous, and tbh I am surprised by the small number of cheaters (I would estimate maybe 10%? Less than 15% for sure). I'm also surprised by the number of good players who play on anonymous and I often get crushed by people who I am 99% sure are not cheaters. Also, in anonymous at least, the majority of cheaters are only using an opening explorer but no engine I think. It's allways weird when a players takes 10 seconds per move to play 20 moves of the main line of some rare subvariation, then starts to play like absolute garbage haha. I want to join you on your last two sentences: everyone has the right to have a good game, and not everyone playing well or above their rating is cheating. We all do sometimes play above our ratings: that's how we gain rating points. Most players better than us are NOT cheaters.

@ColonelMusgrave said in #17:

@DoctorFuu

"[A] team a few people can definitely tackle this ..."

A "few" people? There are millions of players, and tens of millions of games. A team of a "few" people isn't catching anything. If the algorithm isn't catching them, they're not being caught.

Nobody is sitting around at Lichess headquarters digging through hundreds of thousands of games in their spare time, seeing as how allegedly everyone working for this site is a volunteer.

You don't need millions of people to develop an algorithm. Manual reviews most likely only occur after the algorithm flagged a user, but the vast majority of the work is done by an algorithm. Manual reviewers are just last checks "just in case", and I'm sure that they are not called to review most of the cases, as most cheating attempts are likely very trivial.

@ColonelMusgrave said in #17: > @DoctorFuu > > "[A] team a few people can definitely tackle this ..." > > A "few" people? There are millions of players, and tens of millions of games. A team of a "few" people isn't catching anything. If the algorithm isn't catching them, they're not being caught. > > Nobody is sitting around at Lichess headquarters digging through hundreds of thousands of games in their spare time, seeing as how allegedly everyone working for this site is a volunteer. You don't need millions of people to develop an algorithm. Manual reviews most likely only occur after the algorithm flagged a user, but the vast majority of the work is done by an algorithm. Manual reviewers are just last checks "just in case", and I'm sure that they are not called to review most of the cases, as most cheating attempts are likely very trivial.

so much cheating in the 1200 rating range or if people are just paranoid

Paranoid. If someone cheats but somehow still has a 1200 (or 2000) rating, I can still beat him.

>so much cheating in the 1200 rating range or if people are just paranoid Paranoid. If someone cheats but somehow still has a 1200 (or 2000) rating, I can still beat him.

This topic has been archived and can no longer be replied to.