2 Komodo Games: Larry's and Mark's opinions welcome

[Pichard]

Houdini 4 is Ancient History.
Its like matching a Stone Age man with a club against a modern soldier with a Rocket-Launcher !
In case you're not aware, Komodo 10.1 just got blown away by Houdini 200716 at the TCEC.
None of Komodo's too cute for words "good positional play" or anything else you mentioned, helped it to withstand the Brilliance and Ferocity of the new Houdini Engine !
I hope for a Stockfish-Houdini Super-Final as I've always maintained that only the Houdini Engine is capable of beating the new Stockfish, which too is looking awesome, by the way.
This is the end of the reign of the Pretender.
All HAIL the Return of the True King of Computer Chess, HOUDINI !
HOUDART HOUDART HOUDART !

Umm, calm down I guess? Houdini 4.0 is currently number 3 on CCRL, hardly obsolete. And who's leading all the rating lists? Komdoo 10.1 is. You act like the TCEC proves it's inferior. You're judging that one engine is better than another after one game? It takes many games. I blame cato for the opening choice. I'm partially trying to be funny, but also a little bit true. Starting out with such a disadvantage after opening book is pretty ridiculous. 10.1 is the strongest engine out there right now, and it plays beautifully like these games show. Komodo's always on top, so it doesn't get liked by a lot of pepole is what I'm presuming.

I'm not planning to respond to you if you respond with another post like your first.

Tim, let's be real. The strongest engine right now, in the world, is unequivocally and undoubtedly Stockfish. Komodo 10.1 leads the rating lists because Graham chooses to only use Stockfish 7. Anyone who follows Stockfish development knows that Stockfish is already in excess of 60+ Elo over Stockfish 7. Komodo is a fantastic engine, and super strong - but we need to quit lying to ourselves. Even in the last few days there have been like 10 patches committed. In the end Komodo development will not be able to beat a community dedicated to the improvement of an engine. Komodo is just "unlucky" is that typical excuse; Houdini has been quite improved, and Stockfish is far stronger than last TCEC. What we are seeing is Komodo's superiority faltering. Komodo will be effectively superseded. Team Komodo should be fortunate that AsmFish is not used as its scaling with huge threads is out of this world.

PS Houdini 4 is old news. But hey, I agree, it is still the number three engine that we presently have access too.

If you want real results that can be trusted take a look at this: http://spcc.beepworld.de/long-thinkingtime.htm

Those are real results; clearly Stockfish is stronger whether you like it or not. Anyone trusting a rating list using Stockfish 7 needs to rethink what they think they know.

Note that CCRL does rate newer Stockfishes at 40/4, and the July 16 version (about same date as Komodo 10.1) was only 26 elo above SF7, nothing like 60+ claimed. Probably the gain would be even less at 40/40. If you add 26 elo to the IPON, CCRL 40/40, CEGT 40/20, and CEGT 5' + 3" ratings of SF 7, Komodo 10.1 would still generally be on top. It is clear that for whatever reason, Komodo does not do well directly against Stockfish, but almost always does better than Stockfish against other engines. Whether this makes Stockfish or Komodo stronger is a matter of opinion. It's like when Tigran Petrosian was World Champion (based on matches) but rarely won tournaments, was he really the best player in the world? There is no simple answer, it depends on your definition of best player.
We continue to make good steady progress, but so does Stockfish, so the race goes on, regardless of TCEC. The rating lists are the main standard of which engine is best. Now Houdini will also be in the race once version 5 is released, but let's not forget that Mr. Houdart is not the original author of Houdini, who is the still-unknown "Roberto Pescatori", a pseudonym for whoever really wrote the Ippolit program. Mr. Houdart does deserve credit for having improved Ippolit greatly, but he has never admitted that the first "Houdini" was one of the Ippolit versions despite clear and overwhelming evidence to that effect.

My question is, if you take away the opening database, which some people here insist that can give X engine an edge would the latest Komodo beat the latest Stockfish in Chess960? Who is testing engines in Chess960? Another question is why have the Chess960 Computer championship discontinued and when was the latest computers Chess960 championship held?

[lkaufman]

Houdini 4 is Ancient History.
Its like matching a Stone Age man with a club against a modern soldier with a Rocket-Launcher !
In case you're not aware, Komodo 10.1 just got blown away by Houdini 200716 at the TCEC.
None of Komodo's too cute for words "good positional play" or anything else you mentioned, helped it to withstand the Brilliance and Ferocity of the new Houdini Engine !
I hope for a Stockfish-Houdini Super-Final as I've always maintained that only the Houdini Engine is capable of beating the new Stockfish, which too is looking awesome, by the way.
This is the end of the reign of the Pretender.
All HAIL the Return of the True King of Computer Chess, HOUDINI !
HOUDART HOUDART HOUDART !

Umm, calm down I guess? Houdini 4.0 is currently number 3 on CCRL, hardly obsolete. And who's leading all the rating lists? Komdoo 10.1 is. You act like the TCEC proves it's inferior. You're judging that one engine is better than another after one game? It takes many games. I blame cato for the opening choice. I'm partially trying to be funny, but also a little bit true. Starting out with such a disadvantage after opening book is pretty ridiculous. 10.1 is the strongest engine out there right now, and it plays beautifully like these games show. Komodo's always on top, so it doesn't get liked by a lot of pepole is what I'm presuming.

I'm not planning to respond to you if you respond with another post like your first.

Tim, let's be real. The strongest engine right now, in the world, is unequivocally and undoubtedly Stockfish. Komodo 10.1 leads the rating lists because Graham chooses to only use Stockfish 7. Anyone who follows Stockfish development knows that Stockfish is already in excess of 60+ Elo over Stockfish 7. Komodo is a fantastic engine, and super strong - but we need to quit lying to ourselves. Even in the last few days there have been like 10 patches committed. In the end Komodo development will not be able to beat a community dedicated to the improvement of an engine. Komodo is just "unlucky" is that typical excuse; Houdini has been quite improved, and Stockfish is far stronger than last TCEC. What we are seeing is Komodo's superiority faltering. Komodo will be effectively superseded. Team Komodo should be fortunate that AsmFish is not used as its scaling with huge threads is out of this world.

PS Houdini 4 is old news. But hey, I agree, it is still the number three engine that we presently have access too.

If you want real results that can be trusted take a look at this: http://spcc.beepworld.de/long-thinkingtime.htm

Those are real results; clearly Stockfish is stronger whether you like it or not. Anyone trusting a rating list using Stockfish 7 needs to rethink what they think they know.

Note that CCRL does rate newer Stockfishes at 40/4, and the July 16 version (about same date as Komodo 10.1) was only 26 elo above SF7, nothing like 60+ claimed. Probably the gain would be even less at 40/40. If you add 26 elo to the IPON, CCRL 40/40, CEGT 40/20, and CEGT 5' + 3" ratings of SF 7, Komodo 10.1 would still generally be on top. It is clear that for whatever reason, Komodo does not do well directly against Stockfish, but almost always does better than Stockfish against other engines. Whether this makes Stockfish or Komodo stronger is a matter of opinion. It's like when Tigran Petrosian was World Champion (based on matches) but rarely won tournaments, was he really the best player in the world? There is no simple answer, it depends on your definition of best player.
We continue to make good steady progress, but so does Stockfish, so the race goes on, regardless of TCEC. The rating lists are the main standard of which engine is best. Now Houdini will also be in the race once version 5 is released, but let's not forget that Mr. Houdart is not the original author of Houdini, who is the still-unknown "Roberto Pescatori", a pseudonym for whoever really wrote the Ippolit program. Mr. Houdart does deserve credit for having improved Ippolit greatly, but he has never admitted that the first "Houdini" was one of the Ippolit versions despite clear and overwhelming evidence to that effect.

My question is, if you take away the opening database, which some people here insist that can give X engine an edge would the latest Komodo beat the latest Stockfish in Chess960? Who is testing engines in Chess960? Another question is why have the Chess960 Computer championship discontinued and when was the latest computers Chess960 championship held?

CCRL actually has a chess960 list, but it is a blitz (40/4') list. SF is in first, but SF usually is in first on blitz lists of regular chess. My guess is that results would be similar to results in normal chess at the same time control, since neither Komodo nor SF actually tests chess960 normally. Probably the upcoming TCEC rapid tournament with no books, just two games per each pairing, will give you some idea of relative strength when there is no book. You can say this is a chess960 test, in the sense that the normal start position is one of the chess960 positions, and with no books allowed play should reflect engine strength.