- Blind mode tutorial
lichess.org
Donate

Lies, damned lies, and statistics

#2

Where can I see this game or the discussion related to it?

Where can I see this game or the discussion related to it?
#5

I was thinking about cheating the other day. Mostly because there is a lot of chatter about it. I rarely actually play chess, and I do it badly, so I don't even know if someone is cheating or not. I guess ignorance is bliss. But I am a technical person, so I was wondering what could be done against it.

For a few hours I thought about how Counter Strike solved this issue: by asking you to install a third party software that checks your computer for autoaim software. But then I realized that no one would have the resources or motivation to maintain such software, not to mention that you could just as well cheat with another device or someone next to you telling you the moves.

Being a dev, creating some extension to allow me to cheat, even undetectably, is relatively trivial, as well.

So the only solution is actually a fuzzy solution. You have to calculate the probability (take that, Twain!) of someone cheating over time. This means someone who cheats relatively rarely or using some sort of softer assistance would be able to get away with it. No one single game should be accepted as evidence of cheating, even if some are pretty obvious.

But what if we focus on the actual problem? It is NOT cheating that is the problem, but the emotional impact that cheating has on the other players. Well, and financial fraud, one might say, if it's in a paid competition, but I don't care about that at the moment. It also hints on the motivation for cheating: the same as trolling, an emotionally insecure person who lashes out on other people in order to feel powerful. But I digress.

If the problem is that some players negatively affect others, then the solution, as disgusting as it sounds, is simple: a social score. It doesn't even have to be a rating control, it's enough to have a like/dislike button after each game. It doesn't even matter why you didn't like those people. A person that gets a lot of dislikes then will just have a harder and harder time finding opponents.

It's basically UberChess. No one cares at Uber why you gave one star or five stars to a driver. It only matters that you had a reasonably pleasant experience as a client.

Now, I know why such a solution would cause controversy, but it is the simplest, most direct one to the actual reason people are complaining about cheaters (and sandbagging and smurfing and insulting and all of these random words that do nothing but cover up the actual issue).

just my two cents.

I was thinking about cheating the other day. Mostly because there is a lot of chatter about it. I rarely actually play chess, and I do it badly, so I don't even know if someone is cheating or not. I guess ignorance is bliss. But I am a technical person, so I was wondering what could be done against it. For a few hours I thought about how Counter Strike solved this issue: by asking you to install a third party software that checks your computer for autoaim software. But then I realized that no one would have the resources or motivation to maintain such software, not to mention that you could just as well cheat with another device or someone next to you telling you the moves. Being a dev, creating some extension to allow me to cheat, even undetectably, is relatively trivial, as well. So the only solution is actually a fuzzy solution. You have to calculate the probability (take that, Twain!) of someone cheating over time. This means someone who cheats relatively rarely or using some sort of softer assistance would be able to get away with it. No one single game should be accepted as evidence of cheating, even if some are pretty obvious. But what if we focus on the actual problem? It is NOT cheating that is the problem, but the emotional impact that cheating has on the other players. Well, and financial fraud, one might say, if it's in a paid competition, but I don't care about that at the moment. It also hints on the motivation for cheating: the same as trolling, an emotionally insecure person who lashes out on other people in order to feel powerful. But I digress. If the problem is that some players negatively affect others, then the solution, as disgusting as it sounds, is simple: a social score. It doesn't even have to be a rating control, it's enough to have a like/dislike button after each game. It doesn't even matter why you didn't like those people. A person that gets a lot of dislikes then will just have a harder and harder time finding opponents. It's basically UberChess. No one cares at Uber why you gave one star or five stars to a driver. It only matters that you had a reasonably pleasant experience as a client. Now, I know why such a solution would cause controversy, but it is the simplest, most direct one to the actual reason people are complaining about cheaters (and sandbagging and smurfing and insulting and all of these random words that do nothing but cover up the actual issue). just my two cents.
#6

@TotalNoob69 said in #5:

If the problem is that some players negatively affect others, then the solution, as disgusting as it sounds, is simple: a social score. It doesn't even have to be a rating control, it's enough to have a like/dislike button after each game. It doesn't even matter why you didn't like those people. A person that gets a lot of dislikes then will just have a harder and harder time finding opponents.

This was proposed in the linked video (a karma score) and 81Dojo (shogi) uses such a feature. But players still cheat. Personally, I think it's a reasonable solution, since only so much can be done to prevent cheating.

@TotalNoob69 said in #5:

But what if we focus on the actual problem? It is NOT cheating that is the problem, but the emotional impact that cheating has on the other players.

Curiously, Shogi Wars has a paid cheat feature built in to the platform. Players can still be banned for cheating through external means, but as long as you're paying for it, the site allows it. Kramnik suggested something similar though: https://youtu.be/TIBDqf_3NxQ?t=895

@TotalNoob69 said in #5: > If the problem is that some players negatively affect others, then the solution, as disgusting as it sounds, is simple: a social score. It doesn't even have to be a rating control, it's enough to have a like/dislike button after each game. It doesn't even matter why you didn't like those people. A person that gets a lot of dislikes then will just have a harder and harder time finding opponents. This was proposed in the linked video (a karma score) and 81Dojo (shogi) uses such a feature. But players still cheat. Personally, I think it's a reasonable solution, since only so much can be done to prevent cheating. @TotalNoob69 said in #5: > But what if we focus on the actual problem? It is NOT cheating that is the problem, but the emotional impact that cheating has on the other players. Curiously, Shogi Wars has a paid cheat feature built in to the platform. Players can still be banned for cheating through external means, but as long as you're paying for it, the site allows it. Kramnik suggested something similar though: https://youtu.be/TIBDqf_3NxQ?t=895
#7

@Toadofsky said in #6:

Curiously, Shogi Wars has a paid cheat feature built in to the platform

Pay to Win is nothing new :)

@Toadofsky said in #6: > Curiously, Shogi Wars has a paid cheat feature built in to the platform Pay to Win is nothing new :)
#8

@Toadofsky said in #6:

This was proposed in the linked video (a karma score) and 81Dojo (shogi) uses such a feature. But players still cheat. Personally, I think it's a reasonable solution, since only so much can be done to prevent cheating.

The ethical outrage fades when you realize any anti-cheat/moderating system is basically automation of this karma system, several steps removed, to make us feel better.

@Toadofsky said in #6: > This was proposed in the linked video (a karma score) and 81Dojo (shogi) uses such a feature. But players still cheat. Personally, I think it's a reasonable solution, since only so much can be done to prevent cheating. The ethical outrage fades when you realize any anti-cheat/moderating system is basically automation of this karma system, several steps removed, to make us feel better.
#9

@TotalNoob69 said in #8:

The ethical outrage fades when you realize any anti-cheat/moderating system is basically automation of this karma system, several steps removed, to make us feel better.

Eh, I contributed patches to the Lichess rating system and leaderboard to help our automated cheat detection (as well as leaderboards) function more effectively. What is outrageous is Chess.com requiring tournament participants to install spyware to participate in their tournaments: https://www.chess.com/proctor

@TotalNoob69 said in #8: > The ethical outrage fades when you realize any anti-cheat/moderating system is basically automation of this karma system, several steps removed, to make us feel better. Eh, I contributed patches to the Lichess rating system and leaderboard to help our automated cheat detection (as well as leaderboards) function more effectively. What is outrageous is Chess.com requiring tournament participants to install spyware to participate in their tournaments: https://www.chess.com/proctor
#10

Re #5:
While a social score could solve some problems, it could also induce others.

First, some players could have a low social score for their political or other views, even though those are irrelevant during the games. Say, I am unwilling to play about 4 players because of their political views and related actions, and some people might be unwilling to play against me because of my political views. Some players could have a low social score just because their flag is not very popular. (When writing this, I do not want to indulge into fruitless political discussions here.)

Next, some players might have bad reputation just because someone influential publicly accuses them. (To avoid misunderstandings, GM Kramnik was NOT accusing me in his one-year old tweet, but a small fraction of the readers might have interpreted it in that way, and a small fraction is still a lot when there are so many readers.)
Say, there are players A and B, both playing fair. While B looks a bit more suspect according to objective criteria (if there are any), considerably more people suspect A, who had been accused publicly. If you use social score, player A will be paired against many dubious players and some opponents will abort games to him/her. Some heavily underrated youngsters could face too many cheaters just because of performing much better than their (outdated) FIDE ratings would suggest. Sadly, there are still relatively few FIDE-rated rapid and blitz events, especially in some countries. Having more rated events and more data could help us to understand better who is playing fair and who is not.

Finally, one of my friends is not the most pleasant person, yet a welcome opponent in online games. He plays there under his name, does not cheat, does not flag me in deadly drawn positions and so on. That said, he could get many downvotes for not being the most pleasant person, and face worse opponents than he deserves.

That said, there is no perfect solution and some version of social score could be used for pairings if it reduces the side-effects. I guess that chess.com tries to pair official accounts primarily against other official accounts and secret accounts against other secret accounts, which makes sense. (Maybe it even works for trustworthy and dubious accounts as well, though not always well.)

Re #5: While a social score could solve some problems, it could also induce others. First, some players could have a low social score for their political or other views, even though those are irrelevant during the games. Say, I am unwilling to play about 4 players because of their political views and related actions, and some people might be unwilling to play against me because of my political views. Some players could have a low social score just because their flag is not very popular. (When writing this, I do not want to indulge into fruitless political discussions here.) Next, some players might have bad reputation just because someone influential publicly accuses them. (To avoid misunderstandings, GM Kramnik was NOT accusing me in his one-year old tweet, but a small fraction of the readers might have interpreted it in that way, and a small fraction is still a lot when there are so many readers.) Say, there are players A and B, both playing fair. While B looks a bit more suspect according to objective criteria (if there are any), considerably more people suspect A, who had been accused publicly. If you use social score, player A will be paired against many dubious players and some opponents will abort games to him/her. Some heavily underrated youngsters could face too many cheaters just because of performing much better than their (outdated) FIDE ratings would suggest. Sadly, there are still relatively few FIDE-rated rapid and blitz events, especially in some countries. Having more rated events and more data could help us to understand better who is playing fair and who is not. Finally, one of my friends is not the most pleasant person, yet a welcome opponent in online games. He plays there under his name, does not cheat, does not flag me in deadly drawn positions and so on. That said, he could get many downvotes for not being the most pleasant person, and face worse opponents than he deserves. That said, there is no perfect solution and some version of social score could be used for pairings if it reduces the side-effects. I guess that chess.com tries to pair official accounts primarily against other official accounts and secret accounts against other secret accounts, which makes sense. (Maybe it even works for trustworthy and dubious accounts as well, though not always well.)