CB: Feedback on the ICGA/Rybka disqualification scandal

[K I Hyams]

My point is that we should be careful when we make observations that impute motives when we have no evidence for it other than our belief that he is guilty. That is the kind of thing flat earthers do. I'm not too comfortable using my imagination to fill in the gaps. I think we have plenty of facts and real evidence.

In the month before Rybka 1 was released, an influencial member of Rajlich’s team was posting on CCC under the pseudo “Stean”. Part of his contribution was analysis of a game in which he had defeated the, at the time, invincible Shredder 7, a game in which I now suspect that he used the then unreleased Rybka as an analysis tool.

In the weeks after the release of Rybka 1, he transferred his attention to CTF, writing under the name “Ed Murak”. At that time, copy protection was very much in vogue, the ridiculous measures that Letouzey used are worth reflecting upon. However Murak proclaimed very loudly that copy protection was pointless because it would be cracked immediately. The way in which he growled at me when I asked a harmless question led me to wonder whether there was dissent in the team about the issue.

Either way, Rybka 1 was released without protection. I puzzled about Rajlich’s motives when it was later discovered that, in spite of the fact that Rajlich had not seen fit to protect his code, keeping his output protected was so important to him that he would even try to mislead his users by obfuscating it. Bob has pointed out in this thread that such measures are very unusual indeed.

However, I take your point, perhaps I have too much imagination. Julian has said that his initial efforts to compare Fruit and Rybka output have shown no similarity between them. I suggested that there was a similarity in a post of mine earlier in this thread and so, until I have time to check, I am quite happy to withdraw that suggestion.

[Rebel]

If R1-R3 are lost, can you come up with an explanation for the reason why Cimiotti and Rajlich were discussing whether or not they should hand over the "non existent" code to the ICGA?

- N/- By Lukas Cimiotti (*****) [de] Date 2011-08-14 10:28
Vas and I discussed whether or not he should give source code to the ICGA. He really didn't like that idea. My idea was removing all comments and maybe changing all names of variables to make the code harder to understand. But as the guys that disassembled Rybka hadn't understood several parts of the code, we agreed it's safer to not give anything to our competitors.
So Vas only defended himself by saying: I did nothing wrong.
http://rybkaforum.net/cgi-bin/rybkaforu ... ;hl=source

If you can come up with an explanation, I suggest that you offer it to Rajlich because he appeared to me not to have one when I asked him to explain the conundrum.

A strange post indeed.

# Claims that the Fruit clone Strelka contains so much code that is also in Rybka that he was considering claiming ownership of Strelka.

I think Vas claimed ownership at first but later came to some kind of agreement with Osipov. I don't know precisely.

Whether or not Vas and Osipov reached an agreement is irrelevant.
# Strelka is, according to Osipov, a bitboarded Fruit
# Strelka is, according to Rajlich, so close to Rybka 1 that he (Rajlich) could claim ownership.
Put those two statements together and there is a very strong implication indeed that there is Fruit code in Rybka 1.

Mysterious Osipov has hs own agenda, one of the reasons why Zach & Mark started a painful RE job from the R1 executable. If Rybka=Fruit could be proven via Strelka they certainly would have done so.

Completely false. We went to the R1 binary because SEVERAL were saying "we don't know what strelka really is. We know it looks very fruit like. We have Vas' statement that strelka is Rybka 1.0 beta. But can we really trust that?" So there was a demand that we actually compare fruit and rybka 1.0 beta, most likely because those making the demand thought it would not happen. It did, and the rest is history. Only way to show that Rybka == fruit, taking Vas OUT of the equation since his honesty is not to be taken for granted, one HAS to look at Rybka directly. Our only choice was the ASM.

You practically repeat what I was saying, how can that be false ?

RE is the only right way if you don't have the source code.

[Rebel]

Look Keith, it's not something for me to answer, but Lukas Cimiotti. I already said it is odd.

Sorry Ed, I misinterpreted; I thought that you were saying that my comment was odd rather than that of Cimiotti. Had I realised that you had taken the point and were referring to his post, I would not have repeated it.
Keith.

These things happen, no problem.

[Rolf]

Completely false. We went to the R1 binary because SEVERAL were saying "we don't know what strelka really is. We know it looks very fruit like. We have Vas' statement that strelka is Rybka 1.0 beta. But can we really trust that?" So there was a demand that we actually compare fruit and rybka 1.0 beta, most likely because those making the demand thought it would not happen. It did, and the rest is history. Only way to show that Rybka == fruit, taking Vas OUT of the equation since his honesty is not to be taken for granted, one HAS to look at Rybka directly. Our only choice was the ASM.

In one of his many answers this week, Vas replied on that alleged remark "oh, strelka? that's my own rybka" - or similar remark, he said "I dont remember having said that."

So, what was your judicially correct point, Bob?


What?

So there was a demand that we actually compare fruit and rybka 1.0 beta, most likely because those making the demand thought it would not happen. It did, and the rest is history.

Actually what has been proven now? fruit = rybka 1 ?? Excuse me if I laugh. Exactly this has NOT been proven, Sir Bob.

[Don]

My point is that we should be careful when we make observations that impute motives when we have no evidence for it other than our belief that he is guilty. That is the kind of thing flat earthers do. I'm not too comfortable using my imagination to fill in the gaps. I think we have plenty of facts and real evidence.

In the month before Rybka 1 was released, an influencial member of Rajlich’s team was posting on CCC under the pseudo “Stean”. Part of his contribution was analysis of a game in which he had defeated the, at the time, invincible Shredder 7, a game in which I now suspect that he used the then unreleased Rybka as an analysis tool.

In the weeks after the release of Rybka 1, he transferred his attention to CTF, writing under the name “Ed Murak”. At that time, copy protection was very much in vogue, the ridiculous measures that Letouzey used are worth reflecting upon. However Murak proclaimed very loudly that copy protection was pointless because it would be cracked immediately. The way in which he growled at me when I asked a harmless question led me to wonder whether there was dissent in the team about the issue.

Either way, Rybka 1 was released without protection. I puzzled about Rajlich’s motives when it was later discovered that, in spite of the fact that Rajlich had not seen fit to protect his code, keeping his output protected was so important to him that he would even try to mislead his users by obfuscating it. Bob has pointed out in this thread that such measures are very unusual indeed.

However, I take your point, perhaps I have too much imagination. Julian has said that his initial efforts to compare Fruit and Rybka output have shown no similarity between them. I suggested that there was a similarity in a post of mine earlier in this thread and so, until I have time to check, I am quite happy to withdraw that suggestion.

Right. I'm not saying that I don't think he tried to obfuscate the output or that he didn't lose the source code, those things are both very suspicious but I'm uncomfortable with making such an indirect inference.

[Rolf]

I'm not saying that I don't think he tried to obfuscate the output or that he didn't lose the source code, those things are both very suspicious but I'm uncomfortable with making such an indirect inference.

Please try to argue this way about the evil agenda of Bob when he always metions a continual lying or such against Ed. How does he know anything about Ed's character or that of Vas?

I am certain that actually we hadnt this total split of two parties if from the beginning this debate about suspicions about copycoding would have been processed without any character assassination. I fight against it for many years. But apparently many here are being misleaden by Bob's overall fame, so any idea or suspicion from him must be the truth per se.

The truth is somewhere else but not with Bob alone!

[Don]

I'm not saying that I don't think he tried to obfuscate the output or that he didn't lose the source code, those things are both very suspicious but I'm uncomfortable with making such an indirect inference.

Please try to argue this way about the evil agenda of Bob when he always metions a continual lying or such against Ed. How does he know anything about Ed's character or that of Vas?

I am certain that actually we hadnt this total split of two parties if from the beginning this debate about suspicions about copycoding would have been processed without any character assassination. I fight against it for many years. But apparently many here are being misleaden by Bob's overall fame, so any idea or suspicion from him must be the truth per se.

The truth is somewhere else but not with Bob alone!

I just don't see that Bob has an "evil" agenda. Sometimes Bob is too forceful or opinionated (as I can sometimes be too) which offends people who lack humility but he is basically a good guy. What would be his motivation for going after Vas? If it was for competitive reasons he would not publish his source code and be so helpful to anyone who asks questions and he doesn't make money from computer chess that I know of and even if he did how is Vas any kind of threat to him?

[JuLieN]

Julian has said that his initial efforts to compare Fruit and Rybka output have shown no similarity between them. I suggested that there was a similarity in a post of mine earlier in this thread and so, until I have time to check, I am quite happy to withdraw that suggestion.

Well, my intention is to put to check by myself any assertion by one camp or the other, as far as I can do it by myself. Time to trust people without checking by myself is finished.

The current assertion to put to test is this one :

Rick Fadden, Barre, Vermont, USA
(...) People noted that Rybka search looked like Fruit search once the obfuscation was removed.

So here is what I did (I give all the links and procedure so everyone can reproduce everything by himself).

1- First, I downloaded Rybka 1.0 beta there: http://www.superchessengine.com/Rybka%20v1.0betaw32.rar
2- Then I downloaded Fruit 2.1 there: http://wbec-ridderkerk.nl/html/download ... uit_21.zip
3- Then I downloaded Rick Fadden's patch there: http://rybkaforum.net/mwf/rybkaattach/9 ... h9_NEW.zip
4- According to this forum, the patch couldn't be applied starting from Windows Vista, because of some incompatibility withthe real-time patcher. That wasn't a problem, as the pascal source code was included. I just compiled it with FreePascal and applied it to my Rybka binary. Everything went fine and I now had an unobfuscating Rybka 1.0 beta. Ready to test.

I decided to add a third engine to the benchmark, as a control. I chose Glaurung 2.0b, as it's strong and old, although not as much as Rybka and Fruit.

Release dates:
Fruit 2.1: June 17, 2005
Rybka 1.0 beta: November 30, 2005
Glaurung 2.0b : December 2, 2007

First, let's see those engines output after

ucinewgame
go movetime 1000

Rybka:

Code: Select all

info depth 1
info depth 1 score cp -2 time 17 nodes 22 nps 1325 pv a2a4
info depth 1 score cp -1 time 32 nodes 24 nps 768 pv a2a3
info depth 1 score cp 0 time 48 nodes 27 nps 576 pv b2b3
info depth 1 score cp 4 time 64 nodes 31 nps 496 pv d2d4
info depth 1 score cp 7 time 79 nodes 33 nps 427 pv d2d3
info depth 1 score cp 8 time 95 nodes 43 nps 463 pv b1a3
info depth 1 score cp 18 time 110 nodes 45 nps 418 pv b1c3
info depth 1 time 142 nodes 47 nps 338
info depth 2
info depth 2 score cp 3 time 157 nodes 69 nps 450 pv b1c3
info depth 2 time 157 nodes 107 nps 697
info depth 3
info depth 3 score cp 18 time 157 nodes 170 nps 1108 pv b1c3
info depth 3 time 173 nodes 212 nps 1254
info depth 4
info depth 4 score cp 3 time 204 nodes 326 nps 1636 pv b1c3
info depth 4 time 235 nodes 1107 nps 4823
info depth 5
info depth 5 score cp 7 time 266 nodes 1466 nps 5643 pv b1c3
info depth 5 time 298 nodes 1783 nps 6126
info depth 6
info depth 6 score cp 3 time 313 nodes 2847 nps 9314 pv b1c3
info depth 6 time 344 nodes 4331 nps 12892
info depth 7
info depth 7 score cp 8 time 376 nodes 5911 nps 16098 pv b1c3 b8c6
info depth 7 time 391 nodes 6935 nps 18162
info depth 8
info depth 8 score cp 3 time 422 nodes 9921 nps 24073 pv b1c3 b8c6 g1f3
info depth 8 time 485 nodes 20407 nps 43086
info depth 9
info depth 9 score cp 5 time 532 nodes 32684 nps 62910 pv b1c3 b8c6 g1f3 g8f6
info depth 9 time 610 nodes 48119 nps 80776
info depth 10
info depth 10 score cp 7 time 672 nodes 72526 nps 110515 pv b1c3 b8c6 g1f3 g8f6 d2d4
info depth 10 time 719 nodes 94561 nps 134673
info depth 11
info depth 11 score cp 5 time 937 nodes 274699 nps 300204 pv b1c3 g8f6 d2d4 d7d5 c1f4 c8f5
info depth 11 time 1031 nodes 332973 nps 330712
info depth 12
info depth 12 score cp 3 time 1358 nodes 605616 nps 456664 pv b1c3 g8f6 d2d4 d7d5 g1f3 b8c6 d1d3
info currmove b1a3 currmovenumber 2
info currmove d2d3 currmovenumber 3
info currmove d2d4 currmovenumber 4
info currmove b2b3 currmovenumber 5
info currmove a2a3 currmovenumber 6
info currmove a2a4 currmovenumber 7
info currmove b2b4 currmovenumber 8
info currmove c2c4 currmovenumber 9
info currmove c2c3 currmovenumber 10
info currmove e2e4 currmovenumber 11
info currmove e2e3 currmovenumber 12
info currmove f2f4 currmovenumber 13
info currmove f2f3 currmovenumber 14
info currmove g2g4 currmovenumber 15
info currmove g2g3 currmovenumber 16
info currmove h2h4 currmovenumber 17
info currmove h2h3 currmovenumber 18
info currmove g1f3 currmovenumber 19
info currmove g1h3 currmovenumber 20
info depth 12 time 1826 nodes 795680 nps 446208
info depth 13
info currmove b1c3 currmovenumber 1
info depth 13 score cp 7 time 2996 nodes 1834153 nps 626893 pv b1c3 g8f6 d2d4 d7d5 g1f3 b8c6 e2e3 d8d6
info currmove b1a3 currmovenumber 2
info currmove d2d3 currmovenumber 3
info currmove d2d4 currmovenumber 4
info currmove b2b3 currmovenumber 5
info currmove a2a3 currmovenumber 6
info currmove a2a4 currmovenumber 7
info currmove b2b4 currmovenumber 8
info currmove c2c4 currmovenumber 9
info currmove c2c3 currmovenumber 10
info currmove e2e4 currmovenumber 11
info currmove e2e3 currmovenumber 12
info currmove f2f4 currmovenumber 13
info currmove f2f3 currmovenumber 14
info currmove g2g4 currmovenumber 15
info currmove g2g3 currmovenumber 16
info currmove h2h4 currmovenumber 17
info currmove h2h3 currmovenumber 18
info currmove g1f3 currmovenumber 19
info currmove g1h3 currmovenumber 20
info depth 13 time 3449 nodes 1985653 nps 589535
info time 3464 nodes 1985653 nps 586982
bestmove b1c3 ponder g8f6

Fruit:

Code: Select all

info depth 1
info depth 1 seldepth 1 score cp 26 time 0 nodes 2 pv b1a3
info depth 1 seldepth 1 score cp 54 time 16 nodes 3 pv b1c3
info depth 1 seldepth 1 time 47 nodes 21 nps 0
info depth 2
info depth 2 seldepth 2 score cp 0 time 63 nodes 44 pv b1c3 b8c6
info depth 2 seldepth 2 time 94 nodes 82 nps 0
info depth 3
info depth 3 seldepth 3 score cp 54 time 109 nodes 148 pv b1c3 b8c6 g1f3
info depth 3 seldepth 3 time 125 nodes 186 nps 0
info depth 4
info depth 4 seldepth 6 score cp 0 time 141 nodes 300 pv b1c3 b8c6 g1f3 g8f6
info depth 4 seldepth 6 time 172 nodes 976 nps 0
info depth 5
info depth 5 seldepth 9 score cp 48 time 203 nodes 1729 pv b1c3 b8c6 g1f3 g8f6 d2d4
info depth 5 seldepth 9 time 234 nodes 1933 nps 0
info depth 6
info depth 6 seldepth 12 score cp 0 time 265 nodes 3331 pv b1c3 b8c6 g1f3 g8f6 d2d4 d7d5
info depth 6 seldepth 12 time 297 nodes 9447 nps 0
info depth 7
info depth 7 seldepth 14 score cp 42 time 328 nodes 15332 pv b1c3 b8c6 g1f3 g8f6 d2d4 d7d5 c1f4
info depth 7 seldepth 14 time 375 nodes 16243 nps 0
info depth 8
info depth 8 seldepth 17 score cp 0 time 421 nodes 35078 pv b1c3 g8f6 g1f3 b8c6 d2d4 d7d5 c1f4 c8f5
info depth 8 seldepth 20 time 468 nodes 72286 nps 0
info depth 9
info depth 9 seldepth 20 score cp 15 time 531 nodes 125215 pv b1c3 g8f6 g1f3 b8c6 d2d4 d7d5 d1d3 c6b4 d3b5 b4c6
info depth 9 seldepth 21 time 609 nodes 151298 nps 0
info depth 10
info depth 10 seldepth 22 score cp 12 time 827 nodes 330356 pv b1c3 d7d5 d2d4 c8f5 g1f3 g8f6 f3h4 f5g4 h2h3 g4d7
info depth 10 seldepth 22 score cp 15 time 952 nodes 427686 pv g1f3 b8c6 d2d4 d7d5 b1c3 g8f6 d1d3 g7g6 c1f4 c8f5
info time 1014 nodes 440000 nps 433925 cpuload 999
info hashfull 76
info time 1030 nodes 440000 nps 427184 cpuload 996
info hashfull 76
bestmove g1f3 ponder b8c6

Glaurung:

Code: Select all

info depth 2
info depth 2 score cp -9 time 19 nodes 45 nps 2368 pv g1f3 b8c6
info depth 3
info depth 3 score cp 29 time 46 nodes 193 nps 4195 pv g1f3 b8c6 b1c3
info depth 4
info depth 4 score cp -9 time 76 nodes 353 nps 4644 pv g1f3 b8c6 b1c3 g8f6
info depth 5
info depth 5 score cp -11 time 111 nodes 653 nps 5882 pv g1f3 b8c6 b1c3 g8f6 h2h4
info depth 6
info depth 6 score cp -11 time 158 nodes 1258 nps 7962 pv g1f3 b8c6 b1c3 g8f6 h2h4 d7d6
info depth 7
info depth 7 score cp -17 time 195 nodes 3132 nps 16061 pv g1f3 b8c6 b1c3 g8f6 h2h4 d7d6 e2e3
info depth 8
info depth 8 score cp -13 time 241 nodes 5763 nps 23912 pv g1f3 b8c6 b1c3 g8f6 d2d3 f6g4 c3d5 h7h5
info depth 9
info depth 9 score cp -19 time 287 nodes 14500 nps 50522 pv g1f3 b8c6 b1c3 g8f6 e2e3 h7h5 f1b5 a7a6 b5c6 b7c6
info depth 10
info depth 10 score cp -11 time 397 nodes 40040 nps 100856 pv g1f3 b8c6 b1c3 g8f6 e2e3 e7e6 h2h4 h7h5 a2a3 f6g4
info depth 11
info depth 11 score cp 5 time 556 nodes 169205 nps 304325 pv g1f3 b8c6 e2e3 g8f6 f1b5 a7a6 b5c6 d7c6 e1g1 f6g4 b1c3 d8d6
info depth 12
info depth 12 score cp -1 time 709 nodes 277551 nps 391468 pv g1f3 b8c6 e2e3 g8f6 f1b5 d7d5 b1c3 d8d6 b5c6 b7c6 d2d4 c8b7 e1g1 f6g4
info depth 13
info nodes 480000 nps 474308 time 1012 hashfull 20
bestmove g1f3 ponder b8c6

As you can see, Rybka and Fruit don't chose the same move, report different informations during the search (for example Fruit doesn't report the current move to be searched), and, incidently, Rybka doesn't obey the order to stay under 1000 milliseconds...). I can't see a single common line between the two engines in this starting position.

Let's try with other positions, like this one:
[d]r1bqkbnr/2pp1ppp/p7/np2p3/4P3/1B3N2/PPPP1PPP/RNBQK2R w KQkq - 2 6

Code: Select all

ucinewgame
position fen r1bqkbnr/2pp1ppp/p7/np2p3/4P3/1B3N2/PPPP1PPP/RNBQK2R w KQkq - 2 6
go movetime 1000

Rybka:

Code: Select all

info depth 1
info currmove f3e5 currmovenumber 1
info depth 1 score cp 54 time 17 nodes 32 nps 1927 pv f3e5
info depth 1 time 33 nodes 62 nps 1923
info depth 2
info depth 2 score cp 49 time 48 nodes 151 nps 3221 pv f3e5
info depth 2 time 79 nodes 283 nps 3668
info depth 3
info depth 3 score cp 69 time 95 nodes 552 nps 5949 pv f3e5
info depth 3 time 111 nodes 827 nps 7629
info depth 4
info depth 4 score cp 32 time 126 nodes 2545 nps 20683 pv f3e5
info depth 4 time 157 nodes 4023 nps 26239
info depth 5
info depth 5 score cp 31 time 173 nodes 5902 nps 34934 pv f3e5
info depth 5 time 189 nodes 18927 nps 102546
info depth 6
info depth 6 score cp 8 time 235 nodes 25719 nps 112069 pv f3e5
info depth 6 score cp 18 time 282 nodes 61100 nps 221866 pv d2d3
info depth 6 score cp 19 time 329 nodes 66839 nps 208033 pv b1c3
info depth 6 score cp 25 time 360 nodes 79982 nps 227504 pv e1g1
info depth 6 time 376 nodes 84556 nps 230280
info depth 7
info depth 7 score cp 22 time 407 nodes 99962 nps 251501 pv e1g1 g8f6
info depth 7 score cp 28 time 469 nodes 119326 nps 260532 pv b1c3 b5b4
info depth 7 time 501 nodes 137109 nps 280238
info depth 8
info depth 8 score cp 30 time 563 nodes 184263 nps 335142 pv b1c3 b5b4 c3d5
info depth 8 time 672 nodes 263040 nps 400822
info depth 9
info depth 9 score cp 35 time 781 nodes 363983 nps 477232 pv b1c3 b5b4 c3d5 c7c6
info depth 9 time 1000 nodes 586846 nps 600930
info depth 10
info depth 10 score cp 35 time 1593 nodes 1115751 nps 717218 pv b1c3 f8d6 d2d4 e5d4 f3d4
info depth 10 time 2263 nodes 1859088 nps 841231
info depth 11
info depth 11 score cp 32 time 3262 nodes 2886017 nps 905972 pv b1c3 b5b4 c3d5 c7c6 d5e3 g8f6
info depth 11 time 5134 nodes 5109354 nps 182510
info time 5150 nodes 5109354 nps 181943
bestmove b1c3 ponder b5b4

Fruit:

Code: Select all

info depth 1
info depth 1 seldepth 3 score cp 149 time 15 nodes 5 pv f3e5 a5b3 a2b3
info depth 1 seldepth 9 time 31 nodes 381 nps 0
info depth 2
info depth 2 seldepth 9 score cp 149 time 31 nodes 425 pv f3e5 a5b3 a2b3
info depth 2 seldepth 9 time 62 nodes 585 nps 0
info depth 3
info depth 3 seldepth 9 score cp 97 time 78 nodes 917 pv f3e5 a5b3 a2b3 d7d5
info depth 3 seldepth 10 time 125 nodes 1849 nps 0
info depth 4
info depth 4 seldepth 10 score cp 94 time 156 nodes 3003 pv f3e5 a5b3 a2b3 d8g5 e5g4
info depth 4 seldepth 11 time 171 nodes 5035 nps 0
info depth 5
info depth 5 seldepth 14 score cp 61 time 218 nodes 8732 pv f3e5 a5b3 a2b3 d8g5 d2d4 g5g2
info depth 5 seldepth 26 time 281 nodes 51701 nps 0
info depth 6
info depth 6 seldepth 26 score cp 29 time 296 nodes 61735 pv f3e5 a5b3 a2b3 d8g5 d2d4 g5g2 d1f3 g2f3 e5f3
info depth 6 seldepth 26 score cp 48 time 359 nodes 74453 pv b1c3 b5b4 c3d5 c7c6 d5e3 g8f6
info depth 6 seldepth 26 time 405 nodes 83041 nps 0
info depth 7
info depth 7 seldepth 26 score cp 19 time 468 nodes 116546 pv b1c3 b5b4 c3a4 a5b3 a2b3 g8f6 e1g1 d7d5 f3e5 f6e4
info depth 7 seldepth 26 score cp 29 time 483 nodes 138393 pv f3e5 a5b3 a2b3 d8g5 d2d4 g5g2 d1f3 g2f3 e5f3
info depth 7 seldepth 26 score cp 47 time 561 nodes 184803 pv e1g1 a5b3 a2b3 g8f6 d2d4 e5d4 d1d4 d7d5 e4e5
info depth 7 seldepth 29 score cp 57 time 749 nodes 340135 pv d2d4 a5b3 a2b3 d7d5 f3e5 d5e4 e1g1 g8f6
info depth 7 seldepth 29 time 858 nodes 410168 nps 0
info depth 8
info depth 8 seldepth 29 score cp 40 time 998 nodes 570439 pv d2d4 a5b3 a2b3 e5d4 d1d4 g8e7 e1g1 e7c6 d4c3 d8f6
info currmove e1g1 currmovenumber 2
info time 1029 nodes 580000 nps 563654 cpuload 1000
info hashfull 61
info time 1061 nodes 580000 nps 546654 cpuload 990
info hashfull 61
bestmove d2d4 ponder a5b3

Glaurung:

Code: Select all

info depth 2
info depth 2 score cp 135 time 0 nodes 244 nps 0 pv f3e5 a5b3 a2b3
info depth 3
info depth 3 score cp 96 time 31 nodes 534 nps 17225 pv f3e5 a5b3 a2b3 c8b7
info depth 4
info depth 4 score cp 143 time 62 nodes 1354 nps 21838 pv f3e5 a5b3 a2b3 g8f6 b1c3
info depth 5
info depth 5 score cp 23 time 78 nodes 2973 nps 38115 pv f3e5 a5b3 a2b3 d8g5 d2d4 g5g2
info depth 5 score cp 101 time 125 nodes 5596 nps 44768 pv b1c3 b5b4 c3d5 c7c6 d5e3 a5b3 a2b3
info depth 6
info depth 6 score cp 62 time 140 nodes 9937 nps 70978 pv b1c3 f8d6 e1g1 a5b3 a2b3 c8b7 d2d4
info depth 7
info depth 7 score cp 76 time 203 nodes 29717 nps 146389 pv b1c3 g8f6 f3e5 a5b3 a2b3 f8d6 e5g4 e8g8 g4f6 d8f6
info depth 7 score cp 78 time 249 nodes 35500 nps 142570 pv e1g1 a5b3 a2b3 g8f6 b1c3 c8b7 f3e5 f6e4
info depth 8
info depth 8 score cp 74 time 312 nodes 49437 nps 158451 pv e1g1 a5b3 a2b3 g8f6 b1c3 c8b7 f3e5 f8d6 e5f3 f6e4 c3e4 b7e4
info depth 9
info depth 9 score cp 64 time 405 nodes 82494 nps 203688 pv e1g1 a5b3 a2b3 f8d6 b1c3 c8b7 d2d4 e5d4 f3d4 g8f6 c1g5
info depth 10
info depth 10 score cp 68 time 624 nodes 210721 nps 337693 pv e1g1 f8d6 d2d4 e5d4 d1d4 g8f6 c1g5 a5b3 a2b3 c8b7 b1c3 d6e7
info depth 11
info depth 11 score cp 64 time 998 nodes 525950 nps 527004 pv e1g1 a5b3 a2b3 d8e7 b1c3 c8b7 d2d3 g8f6 c1e3 e8c8 b3b4 d7d6
info currmove f3e5 currmovenumber 2
info nodes 540000 nps 501858 time 1076 hashfull 17
bestmove e1g1 ponder a5b3

Three engines, three very different lines. (And Rybka still doesn't obey the order to use less than 1000 ms).

Now let's see with one of my favorite test positions (a deep and difficult checkmate sequence):
[d]1q2r1k1/5pb1/2bppnp1/p1p5/Nr2PPP1/1BBP4/1PP4Q/2KR3R w - - 0 1

Code: Select all

ucinewgame
position fen 1q2r1k1/5pb1/2bppnp1/p1p5/Nr2PPP1/1BBP4/1PP4Q/2KR3R w - - 0 1
go movetime 1000

Rybka:

Code: Select all

info depth 1
info depth 1 score cp 48 time 1 nodes 44 nps 45056 pv c3b4
info depth 1 time 1 nodes 97 nps 99328
info depth 2
info depth 2 score cp 48 time 17 nodes 137 nps 8252 pv c3b4
info depth 2 time 17 nodes 1100 nps 66258
info depth 3
info depth 3 score cp 42 time 17 nodes 1937 nps 116675 pv c3b4
info depth 3 score cp 47 time 32 nodes 2773 nps 88736 pv g4g5
info depth 3 time 48 nodes 3035 nps 64746
info depth 4
info depth 4 score cp 28 time 64 nodes 4536 nps 72576 pv g4g5
info depth 4 score cp 42 time 64 nodes 5075 nps 81200 pv c3b4
info depth 4 time 95 nodes 11832 nps 127536
info depth 5
info depth 5 score cp 28 time 126 nodes 13463 nps 109413 pv c3b4
info depth 5 score cp 66 time 157 nodes 34564 nps 225436 pv h2h4
info depth 5 time 173 nodes 36297 nps 214844
info depth 6
info depth 6 score cp 80 time 188 nodes 47290 nps 257579 pv h2h4
info depth 6 time 204 nodes 55027 nps 276213
info depth 7
info depth 7 score cp 70 time 251 nodes 80722 nps 329320 pv h2h4 b8d8
info depth 7 time 282 nodes 101278 nps 367761
info depth 8
info depth 8 score cp 39 time 391 nodes 179673 nps 470550 pv h2h4 b8d8 c3b4
info depth 8 time 516 nodes 296232 nps 587871
info depth 9
info depth 9 score cp 39 time 641 nodes 399391 nps 638028 pv h2h4 b8d8 c3b4 a5b4 f4f5
info depth 9 time 968 nodes 747342 nps 790576
info depth 10
info depth 10 score cp 54 time 1218 nodes 1011731 nps 850585 pv h2h4 b8d8 c3b4 a5b4 f4f5 e6f5
info depth 10 time 1530 nodes 1267884 nps 848570
info depth 11
info depth 11 score cp 54 time 1998 nodes 1794205 nps 919552 pv h2h4 b8d8 c3b4 a5b4 f4f5 e6f5 g4f5
info depth 11 time 3511 nodes 3504322 nps 1022052
info time 3527 nodes 3504322 nps 1017415
bestmove h2h4 ponder b8d8

Fruit:

Code: Select all

info depth 1
info depth 1 seldepth 2 score cp 126 time 0 nodes 7 pv c3b4 a5b4
info depth 1 seldepth 13 time 32 nodes 1905 nps 0
info depth 2
info depth 2 seldepth 14 score cp 84 time 47 nodes 2570 pv c3b4 b8b4 f4f5 c6a4 b3a4 b4a4 h2d6 f6g4 f5g6
info depth 2 seldepth 14 score cp 103 time 78 nodes 3161 pv g4g5 b4b3 c2b3
info depth 2 seldepth 14 time 125 nodes 4122 nps 0
info depth 3
info depth 3 seldepth 14 score cp 126 time 141 nodes 5036 pv g4g5 c6a4 b3a4 b4a4 g5f6
info depth 3 seldepth 14 time 188 nodes 7285 nps 0
info depth 4
info depth 4 seldepth 16 score cp 115 time 219 nodes 11534 pv g4g5 f6h5 c3b4 b8b4 a4c3 g7c3 b2c3 b4c3
info depth 4 seldepth 16 time 281 nodes 22649 nps 0
info depth 5
info depth 5 seldepth 16 score cp 115 time 312 nodes 30395 pv g4g5 f6h5 c3b4 b8b4 a4c3 g7c3 b2c3 b4c3
info depth 5 seldepth 19 time 359 nodes 59289 nps 0
info depth 6
info depth 6 seldepth 22 score cp 116 time 390 nodes 78660 pv g4g5 f6h5 c3b4 b8b4 a4c3 a5a4 c3a2
info depth 6 seldepth 26 score cp 135 time 593 nodes 214212 pv h2h4 d6d5 c3b4 a5b4 g4g5 f6h5 h1h2 b8f4 h4f4 h5f4 a4c5
info depth 6 seldepth 26 time 656 nodes 259197 nps 0
info depth 7
info depth 7 seldepth 26 score cp 139 time 843 nodes 445432 pv h2h4 b8d8 c3b4 a5b4 g4g5 f6h5 h4f2 c5c4
info depth 7 seldepth 26 time 968 nodes 527015 nps 0
info depth 8
info time 1014 nodes 560000 nps 552268 cpuload 1000
info hashfull 51
info time 1030 nodes 560000 nps 543689 cpuload 1000
info hashfull 51
bestmove h2h4 ponder b8d8

Glaurung:

Code: Select all

info depth 2
info depth 2 score cp 150 time 0 nodes 1963 nps 0 pv c3b4 a5b4
info depth 3
info depth 3 score cp 129 time 31 nodes 3191 nps 102935 pv c3b4 b8b4 g4g5 c6a4 g5f6 g7f6 b3a4 b4a4
info depth 4
info depth 4 score cp 129 time 94 nodes 4122 nps 43851 pv c3b4 b8b4 g4g5 c6a4 g5f6 g7f6 b3a4 b4a4
info depth 5
info depth 5 score cp 129 time 140 nodes 5497 nps 39264 pv c3b4 b8b4 g4g5 c6a4 g5f6 g7f6 b3a4 b4a4
info depth 6
info depth 6 score cp 133 time 187 nodes 8483 nps 45363 pv c3b4 b8b4 g4g5 f6e4 h2h7 g8f8 d3e4 c6a4 b3a4 b4a4
info depth 6 score cp 188 time 265 nodes 25154 nps 94920 pv g4g5 f6h5 c3b4 b8b4 a4c3 g7c3 b2c3 b4c3
info depth 7
info depth 7 score cp 188 time 312 nodes 31654 nps 101455 pv g4g5 f6h5 c3b4 b8b4 a4c3 g7c3 b2c3 b4c3
info depth 8
info depth 8 score cp 139 time 374 nodes 50687 nps 135526 pv g4g5 f6h5 c3b4 b8b4 a4c3 a5a4 c3a2 b4a5
info depth 8 score cp 201 time 530 nodes 148592 nps 280362 pv h2h4 b8d8 c3b4 a5b4 g4g5 f6h5 c1b1 e6e5 f4e5 e8e5
info depth 9
info depth 9 score cp 201 time 608 nodes 181142 nps 297930 pv h2h4 b8d8 c3b4 a5b4 g4g5 f6h5 c1b1 e6e5 f4e5 e8e5
info depth 10
info depth 10 score cp 127 time 780 nodes 306310 nps 392705 pv h2h4 b8d8 c3b4 a5b4 f4f5 c5c4 d3c4 c6e4 f5e6 e4h1 e6f7 g8f7 d1h1
info depth 11
info nodes 510000 nps 495145 time 1030 hashfull 13
bestmove h2h4 ponder b8d8

Again, lines are different.

So, sorry, but I can't replicate Rick's assertion. And I must say I feel bitrayed by those who make those fake assertions. I also read Ed's Rybka page, and I'm not sure anymore that Vas is guilty.

If I make an history of my opinion regarding Rybka, it goes that way:
1- before Fabien's open letter: 1% maybe he's guilty, 99% not guilty. (despite some of my best chess computing camarades being sure it was a Fruit clone, but I just couldn't see how a clone could be 300 Elo points stronger...).
2- After Fabien's letter: 50% guilty, 50% not guilty.
3- After ICGA's report: 90% guilty, 10% not guilty.
4- today: 50% guilty, 50% not guilty (= I don't know, which means "not guilty" if I'm in a jury).

[Albert Silver]

I'm not saying that I don't think he tried to obfuscate the output or that he didn't lose the source code, those things are both very suspicious but I'm uncomfortable with making such an indirect inference.

Please try to argue this way about the evil agenda of Bob when he always metions a continual lying or such against Ed. How does he know anything about Ed's character or that of Vas?

I am certain that actually we hadnt this total split of two parties if from the beginning this debate about suspicions about copycoding would have been processed without any character assassination. I fight against it for many years. But apparently many here are being misleaden by Bob's overall fame, so any idea or suspicion from him must be the truth per se.

The truth is somewhere else but not with Bob alone!

I just don't see that Bob has an "evil" agenda. Sometimes Bob is too forceful or opinionated (as I can sometimes be too) which offends people who lack humility but he is basically a good guy. What would be his motivation for going after Vas? If it was for competitive reasons he would not publish his source code and be so helpful to anyone who asks questions and he doesn't make money from computer chess that I know of and even if he did how is Vas any kind of threat to him?

Well, over four years ago, he let it known that he considered Vas a hooligan and supported reverse-engineering Rybka to take his secrets to use in other programs.

Here is a quote, regarding Strelka, exactly 4 years to the day:

"I am completely unconcerned about the reverse-engineering that has been done. Seems like a fair way to "even the playing field" by forcing a secretive author to expose secrets... I'm not going to lose any sleep over this at all."

Robert Hyatt (Jan 14, 2008)

[Don]

I'm not saying that I don't think he tried to obfuscate the output or that he didn't lose the source code, those things are both very suspicious but I'm uncomfortable with making such an indirect inference.

Please try to argue this way about the evil agenda of Bob when he always metions a continual lying or such against Ed. How does he know anything about Ed's character or that of Vas?

I am certain that actually we hadnt this total split of two parties if from the beginning this debate about suspicions about copycoding would have been processed without any character assassination. I fight against it for many years. But apparently many here are being misleaden by Bob's overall fame, so any idea or suspicion from him must be the truth per se.

The truth is somewhere else but not with Bob alone!

I just don't see that Bob has an "evil" agenda. Sometimes Bob is too forceful or opinionated (as I can sometimes be too) which offends people who lack humility but he is basically a good guy. What would be his motivation for going after Vas? If it was for competitive reasons he would not publish his source code and be so helpful to anyone who asks questions and he doesn't make money from computer chess that I know of and even if he did how is Vas any kind of threat to him?

Well, over four years ago, he let it known that he considered Vas a hooligan and supported reverse-engineering Rybka to take his secrets to use in other programs.

I also support reverse engineering. How many times do we have to listen to the "Vas is innocent" camp try to make this silly juxtaposition that equates using an idea with outright code theft?

The truth of the matter is that computer chess ideas are NOT protected, they belong to everyone and nobody is claiming any differently.

You probably also realize that Bob is an open source guy and doesn't believe ideas should be held "secret" and I don't see any inconsistency here.

I think you need to figure out why Bob called Vas a hooligan - I'm sure it was not for using ideas from other programs since Crafty itself was designed as a vehicle to share ideas with the world.

Here is a quote, regarding Strelka, exactly 4 years to the day:

"I am completely unconcerned about the reverse-engineering that has been done. Seems like a fair way to "even the playing field" by forcing a secretive author to expose secrets... I'm not going to lose any sleep over this at all."


Robert Hyatt (Jan 14, 2008)