Komodo 7

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Komodo 7

Post by overlord »

I disagree. We played against Badur Jobava in one of freestyles. He didn´t have old laptop, he was using 8-core machine (at that time it was very good hardware). The problém of GM´s in Freestyle is that they tends to influence engines opinion. Moreover, they are not used to work with books, databases in a short time control (e.g. 60m+15s). For good Freestyle game you need to use more engines, you need to know their weaknesses. And the crucial sklil is effective analyzing. That is something what they don´t know and simply can´t because it is useless for human OTB play. I know personally many GM´s and IM´s, their preparation for tournaments and matches is far,far from preparation for Freestyle...
Uri Blass
Posts: 10298
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Komodo 7

Post by Uri Blass »

overlord wrote:I disagree. We played against Badur Jobava in one of freestyles. He didn´t have old laptop, he was using 8-core machine (at that time it was very good hardware). The problém of GM´s in Freestyle is that they tends to influence engines opinion. Moreover, they are not used to work with books, databases in a short time control (e.g. 60m+15s). For good Freestyle game you need to use more engines, you need to know their weaknesses. And the crucial sklil is effective analyzing. That is something what they don´t know and simply can´t because it is useless for human OTB play. I know personally many GM´s and IM´s, their preparation for tournaments and matches is far,far from preparation for Freestyle...
I am surprised to read that effective analysis is useless for human OTB play because I thought that it is better for strong humans to do effective analysis of their OTB games in order to learn from their mistakes.
User avatar
M ANSARI
Posts: 3707
Joined: Thu Mar 16, 2006 7:10 pm

Re: Komodo 7

Post by M ANSARI »

overlord wrote:I disagree. We played against Badur Jobava in one of freestyles. He didn´t have old laptop, he was using 8-core machine (at that time it was very good hardware). The problém of GM´s in Freestyle is that they tends to influence engines opinion. Moreover, they are not used to work with books, databases in a short time control (e.g. 60m+15s). For good Freestyle game you need to use more engines, you need to know their weaknesses. And the crucial sklil is effective analyzing. That is something what they don´t know and simply can´t because it is useless for human OTB play. I know personally many GM´s and IM´s, their preparation for tournaments and matches is far,far from preparation for Freestyle...

+1

Freestyle chess can be considered blitz correspondence chess. You need good engines and good hardware and you also need to know the weaknesses of each particular engine. For example if I have 3 engines running and am in a rook endgame where Houdini and Komodo are all showing drawing scores but SF seems to see something ... I will most likely look into the SF move as the strength of SF is ability to go incredibly deep in the endgame. The days where a strong human player will override an engine move quite often are gone ... of course you will see some moves where a human will override the engine's first choice. For example humans are quite good at deciding which line to avoid because in the resulting PV the human can see that although a pawn up, the rook endgame or the opposite bishop endgame is drawish and thus will avoid that line even if it looks really good for the engine. But even these overrides are less and less as engines today understand drawn endgames much better than before.
Uri Blass
Posts: 10298
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Komodo 7

Post by Uri Blass »

M ANSARI wrote:
overlord wrote:I disagree. We played against Badur Jobava in one of freestyles. He didn´t have old laptop, he was using 8-core machine (at that time it was very good hardware). The problém of GM´s in Freestyle is that they tends to influence engines opinion. Moreover, they are not used to work with books, databases in a short time control (e.g. 60m+15s). For good Freestyle game you need to use more engines, you need to know their weaknesses. And the crucial sklil is effective analyzing. That is something what they don´t know and simply can´t because it is useless for human OTB play. I know personally many GM´s and IM´s, their preparation for tournaments and matches is far,far from preparation for Freestyle...

+1

Freestyle chess can be considered blitz correspondence chess. You need good engines and good hardware and you also need to know the weaknesses of each particular engine. For example if I have 3 engines running and am in a rook endgame where Houdini and Komodo are all showing drawing scores but SF seems to see something ... I will most likely look into the SF move as the strength of SF is ability to go incredibly deep in the endgame. The days where a strong human player will override an engine move quite often are gone ... of course you will see some moves where a human will override the engine's first choice. For example humans are quite good at deciding which line to avoid because in the resulting PV the human can see that although a pawn up, the rook endgame or the opposite bishop endgame is drawish and thus will avoid that line even if it looks really good for the engine. But even these overrides are less and less as engines today understand drawn endgames much better than before.
The question is if today it is not better to use only stockfish with better hardware instead of using stockfish komodo and houdini.

I suspect that
giving 16 cores to stockfish may be more effective then giving 8 cores to stockfish and 4 cores to houdini and komodo.
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Komodo 7

Post by overlord »

If you play against human, you just don´t need so deep and serious preparation. You just check opponents games. You are looking for positions which does not like a check with engine some opening lines. I you are very motivated...I kkow even one GM who was very surprised when I told hin two years ago that there are much better engines than Firtz 6 :D (it is not a joke...)
Uri Blass
Posts: 10298
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Komodo 7

Post by Uri Blass »

overlord wrote:If you play against human, you just don´t need so deep and serious preparation. You just check opponents games. You are looking for positions which does not like a check with engine some opening lines. I you are very motivated...I kkow even one GM who was very surprised when I told hin two years ago that there are much better engines than Firtz 6 :D (it is not a joke...)

I think that having a better opening repertoire can help strong GM's because they can get advantage from the opening.

In other words if you have a good memory and you are a strong GM
then it is better for you to know to use chess engines in order to get better openings.

If you remember your preperation you can get advantage from the opening from most of the games that you play with white and usually at least equality with black.
overlord
Posts: 198
Joined: Sun Jun 03, 2012 6:46 pm
Location: Trinec, Czech Republic

Re: Komodo 7

Post by overlord »

Well, even GM´s don´t have super opening lines. In my experience opening level can be divided into these groups:

1. Correspondence games
2. Freestyle games
3. Playchess engine room games
4. GM´s
5. Remaining human beiings
User avatar
Leto
Posts: 2071
Joined: Thu May 04, 2006 3:40 am
Location: Dune

Re: Komodo 7

Post by Leto »

overlord wrote:Well, even GM´s don´t have super opening lines. In my experience opening level can be divided into these groups:

1. Correspondence games
2. Freestyle games
3. Playchess engine room games
4. GM´s
5. Remaining human beiings
Especially Carlsen, his opening prep is perhaps the worst amongst all previous champions. He makes up for it through his ability to play out games that other grandmasters considered as draw. He realized the fact that humans are prone to mistakes even when the game is considered drawish so he takes advantage of this fact by playing out the game knowing that eventually these Super Grandmasters will make a mistake.

Carlsen knows his opening game is weak but he thinks his middlegame and endgame ability is good enough to cover for it and so far he's been right. It would however be better if he improved his opening play and he has made efforts lately to improve that.