Cheating suspicion at the Zadar Open in Croatia

[Don]

I am not 100% sure that this was a cheating case. I have seen cases of people who improved in their 20's, when they start to take chess seriously. I find the article of Chessbase, throwing mud to a guy with no evidence whatsoever, very yellowish and irresponsible. What if he is innocent?

Miguel

I'm jumping into this late and I have not read the numerous posts on it so forgive me if I'm covering old ground.

Did you watch the video where a strong player compares Ivanov's moves to Houdini's? He observes a number of instances where where an unnatural move is made, a move that a strong player is very unlikely to make and yet is very strong. It was pretty overwhelming.

Here is the problem with prosecuting someone based on statistical evidence. When not done with due scientific rigor it can be misinterpreted, such as in the way you mentioned from the link and others. So great care must be taken. It starts when there is an initial suspicion, then facts are collected. Which facts are collected? Whatever facts you NEED to make your case! If you are trying to convict you can search for all the facts that seem to imply guilt. If you are trying exonerate you can easily find lots of facts that will support innocence.

A lot of the "facts" are not facts but observations subject to interpretation and of course the interpretation can be discounted by the other side or hailed as strong evidence.

Sometimes all you have is common sense and unfortunately that is not very common, especially when people get all emotional over things. You cannot prove cheating with overwhelming scientific rigor so you might as well just say nobody cheats and nobody can every be accused of such unless the entire room sees it directly. And even then 1 guy will try to make the case that you didn't see what you thought you saw and will produce at least a "plausible" explanation.

Some people are predisposed one way or the other and their stated reasons are what they believe, but probably not accurate. It is more about what they have witnessed in their lives and things and people who have influenced them. If you have consistently witnessed injustices of the nature that innocent people are persecuted, or it has happened to you, you are probably going to be biased for life against thinking anyone is guilty of anything (unless it is against your own persecutors.) I have seen that from people who have witnessed many injustices or lived in countries where persecution is common, they tend to assume that if ANYONE with any kind of perceived authority makes an accusation against someone it is something corrupt and evil. They are 100% distrustful of authority. There are those that go the other way too, they have good feelings towards people with wealth and power and they are always right, especially when there is a dispute against someone with peon status. Who would take the word of some "lowlife" over a senator?

The scariest thing to me though is that people often cannot tell the difference between whether someone is guilty and whether they should be punished. Deep down inside they don't think someone was wrong but it manifests itself as a belief that they didn't do it at all! And I think they honestly don't know the difference. You have to separate "what happened" from "what should be done about it."

Humans are notoriously bad judges when it comes to discerning the guilt of people they either admire, or are prejudiced against and as such we get the "motive" arguments which resonates with people because we intuitively understand the power of this.

A recent and powerful example was the Lance Armstrong scandal. It illustrated everything I have been saying in the above paragraphs. Lance is extremely popular and charismatic and people love him. Therefore, in the minds of many people this was all about his enemies trying to discredit him. It didn't matter to them that he is admired even by him competitors because it's just not possible that Lance could have been guilty. Arguments were constructed about why the "evil and corrupt" cycling organizations were out to get him and of course pettiness and greed must have been involved.

In like fashion there were many who knew or strongly suspected that he was guilty, but argued against his guilt. The dead giveaway is statements like, "even if he DID take steroids, which he didn't, everyone else was doing it too." This is where tacit approval of something impacts your judgement on ever whether something was done or not.

We have a shameful history of that in this country and probably most other countries too. In past decades it was possible to get away with murder, if it was against a black person. Even when everyone KNEW who did it, there were not enough people with strong enough moral fortitude to stand up against what was done, since in the minds of many at the time it was as best considered a minor crime and in a lot of peoples minds not a crime at all! Convicting such a person was a long uphill battle.

I see a lot of parallels with things that have happened in computer chess. Motive is talked about more than fact, and tacit approval or just plain disagreement about what is acceptable and what is not plays into this in a big way - to the extent that people cannot even agree on what was actually done.

I believe it IS possible to be objective enough to come to rational conclusions about things like the Ivanov cheating but one must heavily discount emotional arguments and your own leanings and your own prior assumptions. In US courts motive plays a big role, if you have overwhelming evidence you also have to show motive. But I have never been comfortable with that because that is the most emotional argument of all. It's easy to impute motive for almost anything. And what is disconcerting to me is that when it gets emotional the argument of motive tends to resonate with people a lot more than actual facts.

I did watch the utube video and I'm completely convinced that this man was cheating. The master kept trying to give the benefit of the doubt but even while making generous assumptions it was impossible to do. It's sort of like a crime where you walk in and find a gun in someone's hand and blood all over the place, but the accused has some plausible explanations and says he walked in and picked up the gun without thinking and was bloody because he was trying to see if he could help the victim. You give him the benefit of the doubt but then you discover that he swore to kill them man. He says, "I said that in anger, I didn't really mean it." But then you discover he made an attempt on his life earlier, and the evidence just keeps rolling in. At some point if you continue to believe his innocence then you are gullible and probably you are emotionally disposed to wanting to believe he is innocent. It's always good to give someone the benefit of the doubt but at some point you have to accept the overwhelming empirical evidence.

Why is he guilty? It isn't because he showed rapid improvement, there is no crime in that. It was because he played the same moves as Houdini an overwhelming number of times including in some situations that seemed really unnatural. It's true that what is natural and what is not is subjective, but I saw those positions and was in agreement. You cannot construct a perfectly rigorous scientific argument but at some point what you should be able to see what clearly stretches the boundaries of credulity.

[stevenaaus]

Anyway - i have the solution. Just hire him as a cheat detector for FIDE tournaments. This is how many security/law-enforcement systems work i think.

But seriously... i kind-of agree with Don's sentiments, but imho the only solution is to catch him red-handed.

[michiguel]

I am not 100% sure that this was a cheating case. I have seen cases of people who improved in their 20's, when they start to take chess seriously. I find the article of Chessbase, throwing mud to a guy with no evidence whatsoever, very yellowish and irresponsible. What if he is innocent?

Miguel

I'm jumping into this late and I have not read the numerous posts on it so forgive me if I'm covering old ground.

Did you watch the video where a strong player compares Ivanov's moves to Houdini's? He observes a number of instances where where an unnatural move is made, a move that a strong player is very unlikely to make and yet is very strong. It was pretty overwhelming.

The video you mentioned was posted after the original chessbase article. So, I still believe that the chessbase article was irresponsible.

Miguel

Here is the problem with prosecuting someone based on statistical evidence. When not done with due scientific rigor it can be misinterpreted, such as in the way you mentioned from the link and others. So great care must be taken. It starts when there is an initial suspicion, then facts are collected. Which facts are collected? Whatever facts you NEED to make your case! If you are trying to convict you can search for all the facts that seem to imply guilt. If you are trying exonerate you can easily find lots of facts that will support innocence.

A lot of the "facts" are not facts but observations subject to interpretation and of course the interpretation can be discounted by the other side or hailed as strong evidence.

Sometimes all you have is common sense and unfortunately that is not very common, especially when people get all emotional over things. You cannot prove cheating with overwhelming scientific rigor so you might as well just say nobody cheats and nobody can every be accused of such unless the entire room sees it directly. And even then 1 guy will try to make the case that you didn't see what you thought you saw and will produce at least a "plausible" explanation.

Some people are predisposed one way or the other and their stated reasons are what they believe, but probably not accurate. It is more about what they have witnessed in their lives and things and people who have influenced them. If you have consistently witnessed injustices of the nature that innocent people are persecuted, or it has happened to you, you are probably going to be biased for life against thinking anyone is guilty of anything (unless it is against your own persecutors.) I have seen that from people who have witnessed many injustices or lived in countries where persecution is common, they tend to assume that if ANYONE with any kind of perceived authority makes an accusation against someone it is something corrupt and evil. They are 100% distrustful of authority. There are those that go the other way too, they have good feelings towards people with wealth and power and they are always right, especially when there is a dispute against someone with peon status. Who would take the word of some "lowlife" over a senator?

The scariest thing to me though is that people often cannot tell the difference between whether someone is guilty and whether they should be punished. Deep down inside they don't think someone was wrong but it manifests itself as a belief that they didn't do it at all! And I think they honestly don't know the difference. You have to separate "what happened" from "what should be done about it."

Humans are notoriously bad judges when it comes to discerning the guilt of people they either admire, or are prejudiced against and as such we get the "motive" arguments which resonates with people because we intuitively understand the power of this.

A recent and powerful example was the Lance Armstrong scandal. It illustrated everything I have been saying in the above paragraphs. Lance is extremely popular and charismatic and people love him. Therefore, in the minds of many people this was all about his enemies trying to discredit him. It didn't matter to them that he is admired even by him competitors because it's just not possible that Lance could have been guilty. Arguments were constructed about why the "evil and corrupt" cycling organizations were out to get him and of course pettiness and greed must have been involved.

In like fashion there were many who knew or strongly suspected that he was guilty, but argued against his guilt. The dead giveaway is statements like, "even if he DID take steroids, which he didn't, everyone else was doing it too." This is where tacit approval of something impacts your judgement on ever whether something was done or not.

We have a shameful history of that in this country and probably most other countries too. In past decades it was possible to get away with murder, if it was against a black person. Even when everyone KNEW who did it, there were not enough people with strong enough moral fortitude to stand up against what was done, since in the minds of many at the time it was as best considered a minor crime and in a lot of peoples minds not a crime at all! Convicting such a person was a long uphill battle.

I see a lot of parallels with things that have happened in computer chess. Motive is talked about more than fact, and tacit approval or just plain disagreement about what is acceptable and what is not plays into this in a big way - to the extent that people cannot even agree on what was actually done.

I believe it IS possible to be objective enough to come to rational conclusions about things like the Ivanov cheating but one must heavily discount emotional arguments and your own leanings and your own prior assumptions. In US courts motive plays a big role, if you have overwhelming evidence you also have to show motive. But I have never been comfortable with that because that is the most emotional argument of all. It's easy to impute motive for almost anything. And what is disconcerting to me is that when it gets emotional the argument of motive tends to resonate with people a lot more than actual facts.

I did watch the utube video and I'm completely convinced that this man was cheating. The master kept trying to give the benefit of the doubt but even while making generous assumptions it was impossible to do. It's sort of like a crime where you walk in and find a gun in someone's hand and blood all over the place, but the accused has some plausible explanations and says he walked in and picked up the gun without thinking and was bloody because he was trying to see if he could help the victim. You give him the benefit of the doubt but then you discover that he swore to kill them man. He says, "I said that in anger, I didn't really mean it." But then you discover he made an attempt on his life earlier, and the evidence just keeps rolling in. At some point if you continue to believe his innocence then you are gullible and probably you are emotionally disposed to wanting to believe he is innocent. It's always good to give someone the benefit of the doubt but at some point you have to accept the overwhelming empirical evidence.

Why is he guilty? It isn't because he showed rapid improvement, there is no crime in that. It was because he played the same moves as Houdini an overwhelming number of times including in some situations that seemed really unnatural. It's true that what is natural and what is not is subjective, but I saw those positions and was in agreement. You cannot construct a perfectly rigorous scientific argument but at some point what you should be able to see what clearly stretches the boundaries of credulity.

[Modern Times]

I am very uneasy about this. I see no evidence whatsoever, but just observations about some moves being moves that Houdini would make and that are unnatural to humans. Do other GMs not also do this sometimes ?

With no evidence and furthermore no theories as to HOW he might be cheating, HOW it is possible for him to do it, how can you say he is guilty at this point in time ? The evidence may well come later, but at the moment I see no factual basis for a guilty verdict.

In Don's analogy, it would perhaps be like walking in to see the dead man with the bullet wound, but he has no gun in his hand.

[Uri Blass]

With chess it is pretty difficult to test, but maybe some new methods should be used ... maybe a lie detector test.

In the post-game press conference he's supposed to reveal what his thoughts were during the game, and if the moves were not his own how can he do that?

This part is not a problem for cheaters who are also strong players.

Often humans hesitate between 2 moves without knowing which move is better and the computer choose one of them that is better.

The human can explain everyone of them.

Even if the computer choose a move that the human did not think that it is a good move then the human can understand later during the game why the move is good and give some explanation.

[Don]

I am very uneasy about this. I see no evidence whatsoever, but just observations about some moves being moves that Houdini would make and that are unnatural to humans. Do other GMs not also do this sometimes ?

It's not just the unnatural moves, it's the fact that he consistently plays the same moves as Houdini. It happens over and over again and then there is a number of these odd moves too.

The way probability and statistics work in this case is that no single move he played means much and you can argue just as you have, that it proves nothing. You cannot look at one move and say, "see, no human would play that!" and it's no crime playing the same move Houdini would play. But if something has only a 20% probability of happening and you keep doing it, the evidence is overwhelming and probably sums up to something very close to certainty.

A lot of people don't understand that. In actual court cases where the evidence really was overwhelming the defense will always try to show that there is another possible explanation, even if it's far fetched for any individual piece of evidence. But when the defense has to argue that way to explain away many strong pieces of evidence, statistically it becomes unfeasible. And yet some jurists don't have a statistical sense and do not know how to combine the evidence and even swayed by the single weakest piece, thinking it's representative.

That is what happened with the Lance Armstrong case in the court of public opinion. You could explain away any individual piece of evidence but together it formed a clear picture. And guess who was guilty? A lot of people said, "I don't really see any evidence here ....."

It works in this case because you can sample ALL the moves he played, not just pick and choose one or two to make your case.

With no evidence and furthermore no theories as to HOW he might be cheating, HOW it is possible for him to do it, how can you say he is guilty at this point in time ? The evidence may well come later, but at the moment I see no factual basis for a guilty verdict.

In Don's analogy, it would perhaps be like walking in to see the dead man with the bullet wound, but he has no gun in his hand.

[Laskos]

I am very uneasy about this. I see no evidence whatsoever, but just observations about some moves being moves that Houdini would make and that are unnatural to humans. Do other GMs not also do this sometimes ?

It's not just the unnatural moves, it's the fact that he consistently plays the same moves as Houdini. It happens over and over again and then there is a number of these odd moves too.

I didn't followed all the discussion, but here I saw only some small probabilities to perform as well as he performed. Is the fact that he plays better than expected something so outrageous? Did someone actually blunder checked or ponder hit his games with Houdini and compared to other humans? Just saying that he plays Houdini moves "again and again" without any serious proof is a bit of stretching the evidence. Could someone actually prove statistically that he is helped by Houdini? For example, that he is some standard deviations closer to Houdini in long enough sequences than other humans?

The way probability and statistics work in this case is that no single move he played means much and you can argue just as you have, that it proves nothing. You cannot look at one move and say, "see, no human would play that!" and it's no crime playing the same move Houdini would play. But if something has only a 20% probability of happening and you keep doing it, the evidence is overwhelming and probably sums up to something very close to certainty.

A lot of people don't understand that. In actual court cases where the evidence really was overwhelming the defense will always try to show that there is another possible explanation, even if it's far fetched for any individual piece of evidence. But when the defense has to argue that way to explain away many strong pieces of evidence, statistically it becomes unfeasible. And yet some jurists don't have a statistical sense and do not know how to combine the evidence and even swayed by the single weakest piece, thinking it's representative.

That is what happened with the Lance Armstrong case in the court of public opinion. You could explain away any individual piece of evidence but together it formed a clear picture. And guess who was guilty? A lot of people said, "I don't really see any evidence here ....."

It works in this case because you can sample ALL the moves he played, not just pick and choose one or two to make your case.

With no evidence and furthermore no theories as to HOW he might be cheating, HOW it is possible for him to do it, how can you say he is guilty at this point in time ? The evidence may well come later, but at the moment I see no factual basis for a guilty verdict.

In Don's analogy, it would perhaps be like walking in to see the dead man with the bullet wound, but he has no gun in his hand.

[Don]

I am very uneasy about this. I see no evidence whatsoever, but just observations about some moves being moves that Houdini would make and that are unnatural to humans. Do other GMs not also do this sometimes ?

It's not just the unnatural moves, it's the fact that he consistently plays the same moves as Houdini. It happens over and over again and then there is a number of these odd moves too.

I didn't followed all the discussion, but here I saw only some small probabilities to perform as well as he performed. Is the fact that he plays better than expected something so outrageous?

No, that is only what caused people to take notice. I don't think any reasonable person considered that as proof of anything.

Did someone actually blunder checked or ponder hit his games with Houdini and compared to other humans? Just saying that he plays Houdini moves "again and again" without any serious proof is a bit of stretching the evidence. Could someone actually prove statistically that he is helped by Houdini? For example, that he is some standard deviations closer to Houdini in long enough sequences than other humans?

The way probability and statistics work in this case is that no single move he played means much and you can argue just as you have, that it proves nothing. You cannot look at one move and say, "see, no human would play that!" and it's no crime playing the same move Houdini would play. But if something has only a 20% probability of happening and you keep doing it, the evidence is overwhelming and probably sums up to something very close to certainty.

A lot of people don't understand that. In actual court cases where the evidence really was overwhelming the defense will always try to show that there is another possible explanation, even if it's far fetched for any individual piece of evidence. But when the defense has to argue that way to explain away many strong pieces of evidence, statistically it becomes unfeasible. And yet some jurists don't have a statistical sense and do not know how to combine the evidence and even swayed by the single weakest piece, thinking it's representative.

That is what happened with the Lance Armstrong case in the court of public opinion. You could explain away any individual piece of evidence but together it formed a clear picture. And guess who was guilty? A lot of people said, "I don't really see any evidence here ....."

It works in this case because you can sample ALL the moves he played, not just pick and choose one or two to make your case.

With no evidence and furthermore no theories as to HOW he might be cheating, HOW it is possible for him to do it, how can you say he is guilty at this point in time ? The evidence may well come later, but at the moment I see no factual basis for a guilty verdict.

In Don's analogy, it would perhaps be like walking in to see the dead man with the bullet wound, but he has no gun in his hand.

[Modern Times]

That is what happened with the Lance Armstrong case in the court of public opinion. You could explain away any individual piece of evidence but together it formed a clear picture. And guess who was guilty? A lot of people said, "I don't really see any evidence here ....."

It works in this case because you can sample ALL the moves he played, not just pick and choose one or two to make your case.

There are significant differences from the Armstrong case.

Again, if you can't show HOW he is doing this, then there is no case to answer. You may be 99% sure he is cheatiing, but if you can't show how it is possible that he is doing it, then I'm afraid he is innocent until the cheating mechanism is discovered. It *was* discovered eventually with Armstrong, maybe it will be discovered here too. But until it is, if it is, then it is innocent until proven guilty.

[Houdini]

Again, if you can't show HOW he is doing this, then there is no case to answer.

I disagree with this.
No human being can consistently play the #1 or #2 choice of Houdini. Detecting this pattern exposes the cheating, no further evidence is required.
I'm surprised that even on a computer chess forum this is not self evident.

Robert