Page 2 of 4

Re: Komodo - is it worth it.

Posted: Tue Mar 17, 2020 5:53 pm
by Dann Corbit
Occasionally, Komodo beats SF. In such a case, SF probably made better decisions.
I guess the real question here is do we want commercial chess engines?
Houdini has given it up.
Komodo is the last really strong commercial engine left in the fight.

The cost of a commercial chess engine is generally less than $100.
We spend thousands on hardware and nary a whimper is raised.
But pocket change for a chess program raises a formal hue and cry.
The irony being that many thousands of hours go into making a really strong chess engine.
If we were to pay someone to write one for us, it would cost a million dollars.

Personally, I want to support the commercial engines.
But it is a choice we all must choose for ourselves.

Re: Komodo - is it worth it.

Posted: Tue Mar 17, 2020 7:35 pm
by Ovyron
Master Om wrote:
Tue Mar 17, 2020 4:49 pm
Can u show me some situations where Komodo does and SF doesnot ?

I am ICCF player and I haven't found any yet.
If I posted examples you'd probably take a look at them and say that it's not clear if, or why is Komodo's line better, because this is in quiet positions where there's nothing to do, so there's many potential lines, so picking one is a matter of taste, and then Komodo has better taste.

In my game against Ozymandias I used Komodo as my main engine, and it general it provided better moves than Stockfish, but I couldn't have done better anyway, so it wasn't significant.

Re: Komodo - is it worth it.

Posted: Tue Mar 17, 2020 9:22 pm
by Ovyron
(replying to myself)
Ovyron wrote:
Tue Mar 17, 2020 7:35 pm
this is in quiet positions where there's nothing to do, so there's many potential lines, so picking one is a matter of taste, and then Komodo has better taste.
This sounds like a subjective statement, so let me elaborate: suppose that after a long analysis session the best line that you get from Stockfish as black is one where you move Rb8 and then that same Rook to Rc8 (it's unclear why not Rc8 already so the tempo isn't wasted), and later in the line black's queen moves to f6 and then back to d8 (wasting two tempos!)

What the hell is wrong with it?

So Komodo would suggest a line where black seems to be doing something instead of running in circles and wasting tempos. Evaluation-wise Stockfish thinks Komodo's line is worse by 2cp or so, but it's a worthwhile price to pay to avoid looking like a beginner learning how the pieces move.

In principle those Stockfish moves have nothing wrong about them, but visually having pieces improving their places would leave a better taste.

In my game against Master Om he plays 3...d6 4...d5 which is the kind of bad taste I'm talking about that you'd avoid by analyzing with Komodo, but truly there's nothing wrong with those kind of moves.

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 2:07 am
by Stephen Ham
Hello All,

I agree with Dann. Those of us who love computer chess need to support it, and that means supporting commercial chess engine developers. Since Komodo is the only one left standing, then it should receive our total support. We all benefit from having Komodo performing versus other engines, as it provides unique qualities, including MCTS. The quality of computer chess and our personal enjoyment would be diminished without Komodo's participation.

I generally purchase every new Komodo integer version, and I purchase directly from Komodo. However, after a hard-drive crash last October, requiring a total PC replacement and "upgrade" to Windows 10, I needed the latest ChessBase GUI. So, I purchased Komodo 13 from ChessBase. But when Komodo 14 arrives, I'll return to purchasing from Komodo.

Please support Komodo.

All the best,
-Steve-

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 3:29 am
by Ovyron
Having the monopoly on top chess engines where there's free alternatives that are superior and expecting full community support because of that makes no sense.

Mainly, if Komodo's project died due to lack of sales BUT it was open-sourced THEN the community as a whole would benefit better than if its sales increased, because we could be able to see what it does and how it does it and its secret ideas would help improve other engines and improve the whole field as a result.

We'd not be afraid that the project dies, but that they decide to keep the source closed, because the fact is that if Komodo was open sourced new people would join it and it'll improve faster and better than it does now, as currently they can't let people peek on it because they can't risk the secrets getting into competing projects.

All the commercial developers that failed and kept their sources closed didn't really care about the community, and only about money.

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 4:50 am
by Master Om
Ovyron wrote:
Tue Mar 17, 2020 7:35 pm
Master Om wrote:
Tue Mar 17, 2020 4:49 pm
Can u show me some situations where Komodo does and SF doesnot ?

I am ICCF player and I haven't found any yet.
If I posted examples you'd probably take a look at them and say that it's not clear if, or why is Komodo's line better, because this is in quiet positions where there's nothing to do, so there's many potential lines, so picking one is a matter of taste, and then Komodo has better taste.

In my game against Ozymandias I used Komodo as my main engine, and it general it provided better moves than Stockfish, but I couldn't have done better anyway, so it wasn't significant.

What u think I haven't tried? even in quiet positions there are other engines which can find moves. There are many SF flavors which can do it too.
Now whats better is a relative word. Even SF would have done that too. Komodo sucks in tactics and lacks understanding of king side attacks. key in chess.

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 12:46 pm
by Krzysztof Grzelak
All chess engines are original in the test .

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 1:10 pm
by swissman
I see nothing about the time control used in these matches.

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 1:12 pm
by Krzysztof Grzelak
Please read carefully because it is written.

Re: Komodo - is it worth it.

Posted: Wed Mar 18, 2020 6:41 pm
by Ovyron
Master Om wrote:
Wed Mar 18, 2020 4:50 am
What u think I haven't tried? even in quiet positions there are other engines which can find moves. There are many SF flavors which can do it too.
If Stockfish flavors can find the best line that means a best line exists and the position isn't quiet enough. I'm specifically talking about positions where several top moves are equal, so it makes no difference which one you pick. In principle those don't have tactics (the presence of tactics makes some lines better than others), so you can do whatever you want, and Komodo can help showing a line that you like how it looks.

But if you don't mind your lines having the same piece moving all over the place and coming back to its original square, or moving a pawn one step and then another instead of just moving it two squares in one go (which are fine because there's NOTHING to do in the position), then you don't need Komodo to find non-dumb looking plans.

(note: this isn't about buying Komodo, but about using Komodo, since I'm talking about K11 and it has become free)