Ivanov speaks out!

Analysis here:

http://www.chess.com/forum/view/general/is-ivanov-a-cheater?page=3

For openers:

http://www.cse.buffalo.edu/~regan/chess/fidelity/ACPcover-and-report.pdf

Google "Ivanov interview" and you'll find it.

Yeah, his statements are pretty over the top.

It worries me that people insist on finding the smoking gun.

Years ago I worked at a casino, before the era of electronic slot machines, they were all electro-mechanical.

One particular machine had a +- 1/500 000 shot of paying the top jackpot. It paid R10 000 for the top prize, with 4 coins played. with 3 coins played, the prize was R2 500, which happened to be the limit the supervisor could authorize without calling the manager.

On one nightshift, this machine paid 4 such Jackpots, with 3 coins played.

Was there any "proof" that there was a scam by staff? No. Were they guilty as hell? Absolutely.

If you haven't read it, the article goldendog posted is excellent. 

Your argument is very strong for someone who didn't review the one hour bit of evidence that was posted. If you don't want to "sit in front of a computer" and review the short bit of evidence that many have linked to, that's fine, but maybe your statement shouldn't be as strong as it is. Those with opinions on the other end have at least backed it up with analysis. By calling people sore losers, you are taking a strong stance yourself -- without even considering the evidence that is out there.

The Kurajika game (round 3) convinced me:

http://www.chessgames.com/perl/chessgame?gid=1703505

Does anyone need the round-by-round links?

It's best you stay out conversations whose subject you clearly (and willfully) know nothing about.

And a thief.  And even worse is this fool of a player (his non-assisted games make me wonder if he's even 2000-2100, it seems not) endangers competitive chess everywhere for all of us who actually play chess.  Arrogant, ignorant, liar, yes, but also saboteur of the game, undermining all professional chess players, and in his games making a mockery of thier work and skill, and he does this  in multiple tournaments.  This is why I say a lifetime ban is the only acceptable response from FIDE.

And if judges in courts knew anything about chess he'd convicted in a criminal case too.  Circumstantial evidence is enough to convict in a court of law here, and the expert testimony by grandmasters and computer experts would be quite convincing, however whatever small amount of money he's won may not be worth a costly court case.  Sadly the offense against the game and all its players all over the world is much greater than whatever prize money he's stolen.

I have read the article, and I believe it could and would be used in a trial as evidence (showing that the guy with the Portugal flag, who believe to be a lawyer in England, would lose in front of the law). Of course there are some funny parts, like that Carlsen cannot match Ivanov's performance. Now we could have a new topic in who is the strongest player of all time? Ivanov, Fischer or Kasparov?

Yes, this game was not only brilliant and with many non-intuitive moves... but these moves had to be backed up by incredible calculation.  Even a very calculating player, who was deeply prepared in that line would find it nearly impossible to produce such a game.  It just doesn't look like a human game.  And he gets a mating attack in a middle game against a 2600 player rated 400 points above him.

And then when I saw his unassisted games I was outraged.  This fool doesn't even care to hide his cheating.  He'll gladly play like an idiot in one tournament then steal from hard working professional players the next.

Maybe an EMP device should be detonated before the start of each round? 

Lilov mentions that the technology is available and inexpensive in his country. A legacy from the Bulgarian State Security agency?

This calls for a SPORTS SWAP.

Lance Dopestrong takes up chess. Ivanov takes up pro cycling.

Imagine if they met up a few months later...

Lance: Holy hell Ivan, the secret lunchbox just isn't helping me get the checkmates.

Ivanov: Stop Whining Lance, for the last %^$ climb I turned on Rybka and the climb still had a $%# gradient of 10 %.

Disappearance of big money tournaments in chess is not necessarily evil idea itself. My opinion is exactly opposite: I don't like professionalism in sports, because it is just unfair pyramid scheme where a lot of amateurs pay and a few professionals get paid. If I want to be just a member of my chess club and play in local team league, I must pay fees to my national federation which in turn sends money to FIDE and Kirsan uses them to do whatever he wants.

The problem is these cheaters steal money from amateurs first place, because they focus mostly on open swiss tournaments.

A few sentences picked from Goldendog's link:

"With Rybka 3, the Single-PV test showed 68.8% matching...On the global list of chess performances, 68.8% is tied with Viswanathan Anand and Vinay Bhat for 29th all-time, sandwiched between Lubomir Kavalek at the 1984 Olympiad and Garry Kasparov in his 1985 match with Ulf Andersson."

"The IPR figure of 3089 is determined directly from the quality of the moves, rather than the results of games, and is completely independent of the opponents' ratings and performance. By comparison, since January 2010, Vladimir Kramnik has two IPRs above 3000, Levon Aronian one(3002 in the match against Kramnik), and Magnus Carlsen none. IPRs at the recent London Classic ranged from 2455 for Anand to 2991 for Kramnik, with similar error bars."

He also says that while 2300 players do have 2700 performances, the intrinsic quality of the moves is far higher than that which is normally needed for a 2700 performance.

Conclusion: "My model projects that for a 2300 player to achieve the high computer correspondence shown in the nine tested games, the odds against are almost a million-to-one."

Or, as Valeri Lilov says: "The evidence shown in the video is pretty clear. First of all, we have 98% of engine moves (Houdini 2.0c to be exact) as the other 2% are the opening moves and second, we can see this player's rating chart. It's impossible to lose or/and drew 1900 players with a very poor play and suddenly destroy 2600 rated GMs in 20 moves with absolute engine match moves. In this case, I believe physical evidence is not required."

Goldendog had a bias in his procedure. He is asserting what is the chance that a person in a tournament with a specific rating is playing moves equal to the suggested computer moves. The bias in the procedure is that he does not look at the strength of the opponent.

Suppose a 2250 player is playing against someone with a rating of 1850 or 2650. In the first game will he have a bigger chance to make a move consistent with moves suggested by the computer then in the game with the higher rated opponent. Simply, because the problems to handle are less complex and more straightforward. The chance that a 1850 player makes a mistake is bigger and it is easier to counter this mistake. Most high rated chess players will agree on how to make use of a mistake made by a player with a much lower rating. Hence, the chance that a 2250 player is making 'computer moves' against a 1850 player should be higher then the chance that a 2250 player is finding 'computer moves' against a 2650 player.

So, his chance of almost one in a million should be corrected. That chance should be significantly lower, make it at most one in ten million.

 That issue is addressed in the report linked when he says "the intrinsic quality of the moves is far higher than that which is normally needed for a 2700 performance"

I agree though, sometimes you make no 'mistakes' and feel like you haven't done anything special, sometimes a couple of blunders and you may still think of it as a good game.

It might be addressed although I understand something else when reading that particular sentence, but I think that it does not return in the conclusion of one in a million. I guess that the chance that Ivanov could have a 2700 performance against a 2600 rated player is much lower then the average chance of one in a million.

It is like the prediction 'will person A jump higher then person B'? Now person A is 1.80m, person B can range between 1.50m and 2.10m. It is obvious that for a good prediction the length of person B is needed and that there is a significant difference in the expected outcome.

Are there other engines of similar strength that Ivanov rarely

agreed with?