i think houdini 1.5 is still better than the updated rybka, correct me if im wrong mr.yereslov
Of course you have to keep in mind that he would correct you if you were right as well.
Well-said!
No, that would be called "trolling."
I do that mostly with Pfren.
Way to emphasize my point perfectly.
i think houdini 1.5 is still better than the updated rybka, correct me if im wrong mr.yereslov
Of course you have to keep in mind that he would correct you if you were right as well.
Well-said!
No, that would be called "trolling."
I do that mostly with Pfren.
Way to emphasize my point perfectly.
What way?
I play such games all the time, Yereslov. Houdini 2 > Stockfish 2.2.2 > Deep Rybka 4, reliably so.
I play such games all the time, Yereslov. Houdini 2 > Stockfish 2.2.2 > Deep Rybka 4, reliably so.
Stockfish is nowhere near ahead of Rybka.
Interesting. Reading a good chunk of this thread (!) and from my own research I find it difficult to believe that any unassisted human could beat Houdini on a modern processor consistantly under any conditions. It's a new era.
Also I cannot tell any difference watching Houdini play against another engine vs. a grandmaster game. Its quite exciting to watch actually and very educational, particularly the evaluation function (computer reports strong white or black advantage in a position that to me looks like a tie). But of course I am a noob, more of a computer geek than serious chess player.
On a separate but related topic, my free Houdini 1.5a x64 easily beats both free Rybka 2.3.2a x32 AND (!) the commercial best-of-breed Deep Fritz 13 x64. I just spent $125 for a nice chess GUI. Groan.
I currently amuse myself playing against free Fritz 5.3 on my netbook at max settings, which I can win with a little help from Houdini on my main machine and a lot of take-backs. Unassisted I don't think I would have a chance. It's rather humbling.
Theoretically we are all undefeatable if we are given enough time to think things through.
Sadly, that's not how chess works.
computers constantly overlook things all the time :), that is until the next computer engine is created
It all depends on your computer and the time it takes the software to process information.
Engines like X3D Fritz or Deep Junior will never lose if given enough time.
computers constantly overlook things all the time :), that is until the next computer engine is created
It all depends on your computer and the time it takes the software to process information.
Engines like X3D Fritz or Deep Junior will never lose if given enough time.
Neither would I, but I would need A LOT of time, and probably a lot of paper too.
One thing Houdini is bad at is the time it takes mop up a completely won position especially when you get it to play itself. It will faff about for 20 or 30 moves gradually improving it's position while any human knows you can usually speed things up by grabbing material quickly at any cost as long as you don't give away the game.
Also sometimes you can speed up it's analysis especially in closed positions or sacrificial lines by "telling" it the first few moves and then backtracking. The hash table will then be filled with preloaded information it needs and empty lines will be pruned from its search.
Theoretically we are all undefeatable if we are given enough time to think things through.
I don't think that's true. If you can't evaluate a position after playing through continuations then you will still lose to a stronger player. Not everyone knows every element of positional chess.
Hi, I managed to make a draw with Houdini 1.5a x64. Here is the game.
I would be very happy if anyone would comment this game.
For once, pfren is well off target.
Humans have no chance against good software/good hardware combinations, at any time control.
All this talk that engines don't understand positional play is simply out of date nonsense.
p.s. I'm willing to back up my assertion.... how about this: I play you (pfren) two games of centaur chess, in which we both use computers. I'm a patzer compared to you OTB but suspect my rig (hardware/software) is superior. Bet you don't even come close to winning a game. Not even close.
I agree.
I think the only way to prove either theory is to hold an experiment, such as the worlds greatest players against the top chess engines.
Hmm, but does strategy matter much when you can see 64 moves ahead in every single line? It seems to me that strategy is essentially a bunch of thinking shortcuts that humans use because we're so intellectually limited.
As an engine fan, I can assure you that chess engines do not look 64 moves ahead. Houdini, Stockfish and Critter are among the farthest-searching programs that I can find, and the maximum depth I've seen in endgames is about 35 half-moves. In midgame and opening positions, the average is usually about 16-23 half-moves on a 64-bit, Windows 4-core PC, in non-correspondence time controls. At correspondence time controls, they usually reach 30-32 half-moves and then get bogged down by the sheer amount of calculations necessary to look farther.
Hi, I managed to make a draw with Houdini 1.5a x64. Here is the game.
I would be very happy if anyone would comment this game.
not to take anything away from your work here, really not my intention.
but I think the official elo's that the various engines sport are their elos/strengths when running on multiple cores, not just an ordinary pc. that is to say, you could probably pop in any of the more modern engines on to your pc and they will all perform at more or less the same strength regardless of their official elos (that were estimated by running them on advanced hardware in computer chess tournaments) so, I guess like around 2300? or something like this. is this right? I'm pretty sure it is.
I would also like to add that most "ordinary PCs" have multiple cores. :)
Hi, I managed to make a draw with Houdini 1.5a x64. Here is the game.
I would be very happy if anyone would comment this game.
not to take anything away from your work here, really not my intention.
but I think the official elo's that the various engines sport are their elos/strengths when running on multiple cores, not just an ordinary pc. that is to say, you could probably pop in any of the more modern engines on to your pc and they will all perform at more or less the same strength regardless of their official elos (that were estimated by running them on advanced hardware in computer chess tournaments) so, I guess like around 2300? or something like this. is this right? I'm pretty sure it is.
I can assure you that engines are stronger than 2300 (I'm assuming your comparing to FIDE ratings) on a normal pc without multiple cores. They would probably be much stronger than that even on a pc from 1995.
Your claim that engines are all about equal on single core machines doesn't seem logical to me at all. What is your justification?
Hi, I managed to make a draw with Houdini 1.5a x64. Here is the game.
I would be very happy if anyone would comment this game.
not to take anything away from your work here, really not my intention.
but I think the official elo's that the various engines sport are their elos/strengths when running on multiple cores, not just an ordinary pc. that is to say, you could probably pop in any of the more modern engines on to your pc and they will all perform at more or less the same strength regardless of their official elos (that were estimated by running them on advanced hardware in computer chess tournaments) so, I guess like around 2300? or something like this. is this right? I'm pretty sure it is.
I would also like to add that most "ordinary PCs" have multiple cores. :)
but they don't have more than 5 (at most) right? they don't have 16 do they? and aren't these just the newer computers anyway? if you have a 2010 computer with an i5 core or something, this isn't equivalent to 7 or 10 or 16 cores is it? I somehow doubt this.
Deep Blue in the late 1990s had 32 cores. The infidel Rybka had over a hundred cores for its supercomputer, and other competitors for the World Computer Chess Championship -- Johnny in particular -- have far more than the normal 4-core computers.
Also, there is an extreme difference between, say, Houdini running on a single-core PC and Dragon 4.6, SOS 5.1, etc. Chess engines are not created equal.
EDIT: And, finally, the Japanese have a super-computer that has over a hundred thousand cores. In answer to your question, going beyond 4 cores is not all that difficult.
ANOTHER EDIT: I just realized you were asking about ordinary computers. These days, 4 cores is the norm, but I have heard of six and eight core computers on the market.
2300? I've played houdini before and it seems somewhat more relentless than that. I'd be curious, though, to know just how much rating goes away when you only use ordinary hardware. Kenpo seems to be claiming that it takes off about 800 points (3100 to 2300) but that seems like a really extreme estimate. If the difference is anywhere near this significant, this is certainly news to me! 
If someone can have Houdini play ten games against Rybka, that would be great.