S.Taylor wrote:The way it looks, from Houdini 3 vs Rybka cluster, now, well into game 2, it seems like Houdini 3 is WELL strong vs Rybka (esp that it is at much greater power), and likely more than Houdini 1.5. is, in straight games between it and Rybka.
Earlier tests may well have not shown the truth yet.
But no one said this yet.
[I hope no one wants to argue that Rybka plays worse when it thinks for too long/or too much power].
Houdini is not that brilliant as many think Shimon.....
It's just that Rybka is making horrible positional blunders and is getting squashed by Houdini.....
Something is wrong here,I don't know.....
Is it a bad special settings for the match
Is it a bad scaling and not getting the optimal communication with all the 64 processors that Rybka is using
Again,after watching the second game,I am pretty much convinced that Houdini is not that ingenius chess entity as much as there is definitely something wrong with cluster Rybka.......
Dr.D
My argument would be, that Houdini 3, as far as i know, is doing better than any other engine on the planet, except, some said, vs Rybka, which was dominated more, ironically, by Houdini 1.5a.
This did not make much sense anyway, and now we see that Houdini 3 IS dominating Rybka just as well, even at cluster x64 vs 16 cores.
So how do you conclude that Houdini is not all that great, if there has never yet been anything better, (in any way)?
By watching Rybka's terrible blunders even in the very first 4-5 moves.....
Sure,Houdini did good but look at the games again....
An experienced chess player will notice that there is something definitely wrong with Rybka's evaluation of the positions......
Frankly speaking,Rybka served these 2 points to Houdini on a silver plate.....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
yes rybka did hand the game over on a silver plate because that is all Rybka has to offer and the cluster was all hype to con people to believe it can give you the edge...
I say Robert should build a Houdini cluster that will work for real...
Dr.Wael Deeb wrote:By watching Rybka's terrible blunders even in the very first 4-5 moves.....
Sure,Houdini did good but look at the games again....
An experienced chess player will notice that there is something definitely wrong with Rybka's evaluation of the positions......
Frankly speaking,Rybka served these 2 points to Houdini on a silver plate.....
Dr.D
If so, that's a shame. you are not explainig why Houdini isn't the greatest in history to date, but perhaps that Rybka isn't good in the opening without a book (if it didn't use a book). If so, they should try opening moves pre-played, or book, to give a better test after that.
So are you saying that Rybka wasn't even transcibed well to cluster at all, and that it simply hadn't come to the attention of everyone as not many people reported back or tried it (or bothered spending expensive time testing it)? So Rybka 4.1 on a 16-core xeon would be alot better?
gotogo wrote:yes rybka did hand the game over on a silver plate because that is all Rybka has to offer and the cluster was all hype to con people to believe it can give you the edge...
I say Robert should build a Houdini cluster that will work for real...
I totally agree.......
Robert will go the cluster way as long as it will hatch golden eggs for him....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
Dr.Wael Deeb wrote:By watching Rybka's terrible blunders even in the very first 4-5 moves.....
Sure,Houdini did good but look at the games again....
An experienced chess player will notice that there is something definitely wrong with Rybka's evaluation of the positions......
Frankly speaking,Rybka served these 2 points to Houdini on a silver plate.....
Dr.D
If so, that's a shame. you are not explainig why Houdini isn't the greatest in history to date, but perhaps that Rybka isn't good in the opening without a book (if it didn't use a book). If so, they should try opening moves pre-played, or book, to give a better test after that.
So are you saying that Rybka wasn't even transcibed well to cluster at all, and that it simply hadn't come to the attention of everyone as not many people reported back or tried it (or bothered spending expensive time testing it)? So Rybka 4.1 on a 16-core xeon would be alot better?
I can't speculate about all this stuff Shimon,but just take a closer look at Rybka's third move in the second game.....
3....h6 ?!
Come on,give me a break regards,
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
Dr.Wael Deeb wrote:By watching Rybka's terrible blunders even in the very first 4-5 moves.....
Sure,Houdini did good but look at the games again....
An experienced chess player will notice that there is something definitely wrong with Rybka's evaluation of the positions......
Frankly speaking,Rybka served these 2 points to Houdini on a silver plate.....
Dr.D
If so, that's a shame. you are not explainig why Houdini isn't the greatest in history to date, but perhaps that Rybka isn't good in the opening without a book (if it didn't use a book). If so, they should try opening moves pre-played, or book, to give a better test after that.
So are you saying that Rybka wasn't even transcibed well to cluster at all, and that it simply hadn't come to the attention of everyone as not many people reported back or tried it (or bothered spending expensive time testing it)? So Rybka 4.1 on a 16-core xeon would be alot better?
I can't speculate about all this stuff Shimon,but just take a closer look at Rybka's third move in the second game.....
3....h6 ?!
Come on,give me a break regards,
Dr.D
But apart from the first few moves, which it does not usually do without a book, do you know for sure that it played worse than Rybka 4.1?
3...h6, ok, this is obvious. But that is a one-time hapenning, or maybe another one or two times in the opening. (not that it loses something directly either). But it IS a Rybka, and ANYTHING can have a bug somewhere, affecting only one thing or even one time.
Dr.Wael Deeb wrote:By watching Rybka's terrible blunders even in the very first 4-5 moves.....
Sure,Houdini did good but look at the games again....
An experienced chess player will notice that there is something definitely wrong with Rybka's evaluation of the positions......
Frankly speaking,Rybka served these 2 points to Houdini on a silver plate.....
Dr.D
If so, that's a shame. you are not explainig why Houdini isn't the greatest in history to date, but perhaps that Rybka isn't good in the opening without a book (if it didn't use a book). If so, they should try opening moves pre-played, or book, to give a better test after that.
So are you saying that Rybka wasn't even transcibed well to cluster at all, and that it simply hadn't come to the attention of everyone as not many people reported back or tried it (or bothered spending expensive time testing it)? So Rybka 4.1 on a 16-core xeon would be alot better?
I can't speculate about all this stuff Shimon,but just take a closer look at Rybka's third move in the second game.....
3....h6 ?!
Come on,give me a break regards,
Dr.D
But apart from the first few moves, which it does not usually do without a book, do you know for sure that it played worse than Rybka 4.1?
3...h6, ok, this is obvious. But that is a one-time hapenning, or maybe another one or two times in the opening. (not that it loses something directly either). But it IS a Rybka, and ANYTHING can have a bug somewhere, affecting only one thing or even one time.
No one knows for sure Shimon........
Maybe 2-3 people on this planet know and one of them is certainly Vasik.....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
well Houdini 3 made me repair a lot of inaccuracies in my ChessPartner userbook, but the repertoire hasnt changed. And just like it used to be, a good book can make about +/- 100 points difference; and you need quite some bucks to compensate for that with stronger hardware.
ok its comparing apples and oranges, but Houdini3 is sold without
book in most cases i think, and 'only' gives about 50 Elo improvement.
Not so spectacular imho, and a bit of a hype..
regards
jef www.superchess.blogspot.com
H3 is a truly remarkable analysis engine & this is what 99% of people are paying for...i do think houdart seems to understand this & it's why he's introduced a feature such as tactical mode.
thie HIARCS book model is probably best, but i would prefer a book where moves are more than color coded or given exclaims. a book is one area where the GUI+information can easily interact to really give the user some valuable additional info. w/ an up to date hiarcs book, you may have the best moves, but you're not given the reasons why & must discover those yourself. similar to analyzing with houdini.
