Chess Player Strip Searched

[Nick Burrows]

Ivanov's win against GM Grigoryan now online:

Event "XIX OPEN INT. NAVALMORAL DE LA MATA - XI PREMIO CE"] [Site "Navalmoral de la Mata"] [Date "2013.12.07"] [Round "4"] [Board "1"] [White "Grigoryan, Karen H."] [Black "Ivanov, Borislav"] [Result "0-1"] [ECO "A00"] [WhiteElo "2604"] [BlackElo "2318"] [PlyCount "100"] [EventDate "2013.12.05"] [EventRounds "7"] [EventCountry "ESP"] 1. Nf3 Nf6 2. g3 d5 3. Bg2 Bf5 4. O-O c6 5. c4 e6 6. cxd5 exd5 7. d3 Be7 8. Nh4 Be6 9. e4 O-O 10. Qc2 Na6 11. e5 Nd7 12. Nf3 Nb4 13. Qe2 Bf5 14. Rd1 Nc5 15. Ne1 d4 16. Na3 a5 17. Nac2 Qd7 18. b3 f6 19. Ba3 Rfe8 20. Bxb4 axb4 21. exf6 Bxf6 22. Qd2 Re7 23. Qxb4 Na6 24. Qd2 Rae8 25. Bf3 Bg4 26. Qf4 Bh3 27. b4 Nc7 28. Qd2 Bg4 29. Qf4 Bxf3 30. Qxf3 Nb5 31. Ng2 Nc3 32. Re1 Ne2+ 33. Kh1 Bg5 34. Rad1 Rf8 35. Qxe2 Rxe2 36. Rxe2 Re8 37. Rxe8+ Qxe8 38. Nxd4 Bf6 39. Re1 Qd7 40. Nc2 Qxd3 41. Nge3 Bd4 42. Kg1 Bxe3 43. Nxe3 Qd2 44. Rb1 Qxa2 45. Rc1 Qb2 46. Rc4 b5 47. Rxc6 Qb1+ 48. Nf1 h6 49. Rb6 Qxb4 50. Rb8+ Kh7 0-1

In this game, Mr Ivanov remarkably chooses one of Houdini's first 2 choices on every single move!!

[Paolo Casaschi]

Ivanov's win against GM Grigoryan now online:

Event "XIX OPEN INT. NAVALMORAL DE LA MATA - XI PREMIO CE"] [Site "Navalmoral de la Mata"] [Date "2013.12.07"] [Round "4"] [Board "1"] [White "Grigoryan, Karen H."] [Black "Ivanov, Borislav"] [Result "0-1"] [ECO "A00"] [WhiteElo "2604"] [BlackElo "2318"] [PlyCount "100"] [EventDate "2013.12.05"] [EventRounds "7"] [EventCountry "ESP"] 1. Nf3 Nf6 2. g3 d5 3. Bg2 Bf5 4. O-O c6 5. c4 e6 6. cxd5 exd5 7. d3 Be7 8. Nh4 Be6 9. e4 O-O 10. Qc2 Na6 11. e5 Nd7 12. Nf3 Nb4 13. Qe2 Bf5 14. Rd1 Nc5 15. Ne1 d4 16. Na3 a5 17. Nac2 Qd7 18. b3 f6 19. Ba3 Rfe8 20. Bxb4 axb4 21. exf6 Bxf6 22. Qd2 Re7 23. Qxb4 Na6 24. Qd2 Rae8 25. Bf3 Bg4 26. Qf4 Bh3 27. b4 Nc7 28. Qd2 Bg4 29. Qf4 Bxf3 30. Qxf3 Nb5 31. Ng2 Nc3 32. Re1 Ne2+ 33. Kh1 Bg5 34. Rad1 Rf8 35. Qxe2 Rxe2 36. Rxe2 Re8 37. Rxe8+ Qxe8 38. Nxd4 Bf6 39. Re1 Qd7 40. Nc2 Qxd3 41. Nge3 Bd4 42. Kg1 Bxe3 43. Nxe3 Qd2 44. Rb1 Qxa2 45. Rc1 Qb2 46. Rc4 b5 47. Rxc6 Qb1+ 48. Nf1 h6 49. Rb6 Qxb4 50. Rb8+ Kh7 0-1

In this game, Mr Ivanov remarkably chooses one of Houdini's first 2 choices on every single move!!

How odd, why not the first one every time?
Could you add some details how you tested this? I mean, which computer, how much time you did allow Houdini for each move and so on.

[Nick Burrows]

Hi Paolo,

I have a very fast computer, I only spent around 20 seconds per move - with which at 3400 elo Houdini would beat any GM.

It is impossible for any human being to unswervingly play the computers first 2 choices.

[Roger de Coverly]

How odd, why not the first one every time?

It's only conjecture that he's using an engine and no guarantee that it's even Houdini. The same engine can give different results depending on the machine used and the background settings.

[Paolo Casaschi]

How odd, why not the first one every time?

It's only conjecture that he's using an engine and no guarantee that it's even Houdini. The same engine can give different results depending on the machine used and the background settings.

Completely agree. This is why I believe that running through the game with a random computer for a random amount of time (or, even worse, waiting for the player move to show up) proves ABSOLUTELY NOTHING.

[Nick Burrows]

Completely agree. This is why I believe that running through the game with a random computer for a random amount of time (or, even worse, waiting for the player move to show up) proves ABSOLUTELY NOTHING.

It's not "proof" in the legal sense, however if you compare the analysis of the top few engines they are unanimous in their choices for virtually all critical positions, and give a choice of moves for non-critical positions.

Using an engine skillfully it is very easy to spot a human player vs an engine. Not one of the best Super GM's can consistently replicate an engines moves. The fact that Ivanov does do this is certainly not "nothing".

[Paolo Casaschi]

Completely agree. This is why I believe that running through the game with a random computer for a random amount of time (or, even worse, waiting for the player move to show up) proves ABSOLUTELY NOTHING.

It's not "proof" in the legal sense, however if you compare the analysis of the top few engines they are unanimous in their choices for virtually all critical positions, and give a choice of moves for non-critical positions.

For "critical" positions though it's not unlikely for a good player to find a good move either. The proof of comparing moves with an engine is more relevant to positions where there's no obvious best move; a match there with a given engine would indicate use of such engine. However, you need to be very specific with your engine use: for instance, there is research evidence that the engines change their selected best move with a probability of 15% on average every time the search depth goes one half move deeper; this means that, if you use houdini on a given computer for a given amount of time you only get a 100% match if Ivanov uses the same software, on the same computer, for the same amount of time; otherwise you should statistically get significant differences (that 15% per halfmove depth piles up quickly); unless of course you let the engine run until it shows the move you expect, then consider this a match and move on.

Using an engine skillfully it is very easy to spot a human player vs an engine. Not one of the best Super GM's can consistently replicate an engines moves. The fact that Ivanov does do this is certainly not "nothing".

This is because you already decided that Ivanov is cheating and how he's doing it.
It looks to me that you claim your detection method is good because easily spots Ivanov as a cheater and at the same time you claim that the detection of Ivanov as a cheater is correct because your "good" method spots him. After all, contrary to popular belief, Ivanov has not been proven a cheater yet (if he were then we would know how he's doing it and he would be banned from playing).

[Nick Burrows]

For "critical" positions though it's not unlikely for a good player to find a good move either. The proof of comparing moves with an engine is more relevant to positions where there's no obvious best move; a match there with a given engine would indicate use of such engine.

A good player will find 'good' moves. In critical positions when there is a single route to a clear evaluation it is impossible for a human to do this.

When the position is 'non-critical' there is a wide array of moves, for a human player to continuously choose one of (coincidentally the best) an engines top 2 choices is equally unlikely.

Ivanov has not been proven a cheater yet (if he were then we would know how he's doing it and he would be banned from playing).

This comes down to the minefield of "proving" the cheating legally. It is patently obvious he is in reality doing it.

[Paolo Casaschi]

It is patently obvious he is in reality doing it.

Then amuse me with your explanation about how he can possibly doing it.

[Matt Mackenzie]

The answer is - he MUST be.

It is IMPOSSIBLE for him to play as he does otherwise - as shown by the fact that *no* super-GM, from Carlsen downwards, can match it.

[Nick Burrows]

Then amuse me with your explanation about how he can possibly doing it.

I don't wish to amuse you thanks. Are you his manager?

*The analysis of his moves clearly shows cheating
*The analysis of his results clearly indicates cheating
*His peers (A 20 GM signed petition refusing to play him) clearly believe he is cheating
*His Behaviour clearly indicates cheating. (Refusing to be searched/Not presenting at the tribunal created to clear him)

Due to methodological problems of proving it, he cannot yet be prevented from playing. It will come!

[Paolo Casaschi]

Due to methodological problems of proving it, he cannot yet be prevented from playing. It will come!

At that time, then we would agree he's a cheater.

For now I place the sentence "Ivanov is certainly a cheater" next to "foreigners are benefit frauds", "bankers are greedy thiefs" and similar witchhunts.

[Nick Burrows]

So nothing is true until it is proven in law?

[Roger de Coverly]

So nothing is true until it is proven in law?

If you are intending to ban him from playing chess, then at least the "civil courts" test of "balance of probabilities" should be necessary. It's over the board chess with no apparent means that the position can be transmitted to an engine and the engine communicating back its move. In the absence of physical evidence to the contrary, matching moves to a computer engine is a demonstration of skill rather than cheating as is beating GMs whilst being a lower rated player.

The Germans have tried to ban players in circumstances with good evidence, only to find themselves in legal minefields to make it effective.

It might yet come to a legal action. Various people have made accusations which are defamatory unless true. What would happen if Ivanov decided to fight back?

[Matt Mackenzie]

Due to methodological problems of proving it, he cannot yet be prevented from playing. It will come!

At that time, then we would agree he's a cheater.

For now I place the sentence "Ivanov is certainly a cheater" next to "foreigners are benefit frauds", "bankers are greedy thiefs" and similar witchhunts.

The last two statements are generalisations about groups.

This matter concerns accusations against an individual which are, in my opinion and that of many much more knowledgeable about these things, very hard to explain by any other means than cheating. But continue with your misguided sense of moral superiority if you wish