GM and Rybka vs. Stockfish

[Uri Blass]

Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Exactly.

Perhaps GM has a negative added value. That's my guess.

You mean the specific GM because a different GM may have a non negative added value and the simplest way to do it is simply to let rybka to play all the moves.

I do not believe that letting rybka to play is the best strategy that humans can use.

[S.Taylor]

Why not? Even with Rybka 3 vs the best engine around, i don't see why a human with good chess understanding would not win a match, because human understanding will allow him to think of very clever and subtle ideas whilst constantly looking forward several moves to check it out, before the human making his move. Therefore there will be no big mistakes plus great human brilliance and depth and wisdom.

Well, it seemed the other way around, the GM (having Rybka 3) wondered at Stockfish 5 "brilliance and depth and wisdom".

Nice test, anyway, it seems the engines are so far ahead, that Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Thanks for pointing this out to me. I admit i hadn't realized this.

It is hard to believe it should have been like this, because i thought that some humans can add huge amounts to the elo/iQ of a program. But maybe 1).that is no longer true 2).Rybka 3 is different 3). This GM is different.

[Laskos]

Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Exactly.

Perhaps GM has a negative added value. That's my guess.

You mean the specific GM because a different GM may have a non negative added value and the simplest way to do it is simply to let rybka to play all the moves.

I do not believe that letting rybka to play is the best strategy that humans can use.

I am becoming a bit skeptical, after seeing in ICCF rating list some multi-millionaires (with money unrelated to chess) like Joop van Oosterom (top of the list) ascend rapidly. It is rumored that they spend hundreds of thousands on computer hardware. All this ICCF competition seems like computer playground. Do you really believe in significant Joop van Oosterom's personal chess input behind his mountain of CPUs?

[Laskos]

Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Exactly.

Perhaps GM has a negative added value. That's my guess.

You mean the specific GM because a different GM may have a non negative added value and the simplest way to do it is simply to let rybka to play all the moves.

I do not believe that letting rybka to play is the best strategy that humans can use.

I am becoming a bit skeptical, after seeing in ICCF rating list some multi-millionaires (with money unrelated to chess) like Joop van Oosterom (top of the list) ascend rapidly. It is rumored that they spend hundreds of thousands on computer hardware. All this ICCF competition seems like computer playground. Do you really believe in significant Joop van Oosterom's personal chess input behind his mountain of CPUs?

To support the claims that ICCF ratings follow more the hardware use than human abilities in CC, I just downloaded the latest ICCF rating list, plotted it here, and calculated some statistical quantities. The upper half in ratings is especially dubious, and its shape resembles that of the distribution of used hardware:


Skewness: -0.772
Kurtosis: 0.274

These are not of particular human cognitive abilities in most cases.

[Uri Blass]

Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Exactly.

Perhaps GM has a negative added value. That's my guess.

You mean the specific GM because a different GM may have a non negative added value and the simplest way to do it is simply to let rybka to play all the moves.

I do not believe that letting rybka to play is the best strategy that humans can use.

I am becoming a bit skeptical, after seeing in ICCF rating list some multi-millionaires (with money unrelated to chess) like Joop van Oosterom (top of the list) ascend rapidly. It is rumored that they spend hundreds of thousands on computer hardware. All this ICCF competition seems like computer playground. Do you really believe in significant Joop van Oosterom's personal chess input behind his mountain of CPUs?

I do not know and millionaires can also buy GM's help and not only computer's help.

Note that I never claimed that hardware does not help so it is clearly possible that having a better hardware can be enough to get into the top of the ICCF rating list without human input and it does not contradict the opinion that human input can help.

I did not say how much human input can help and 10 elo is also something when of course having hardware that is twice faster give more than 10 elo.

Uri

[Dr.Wael Deeb]

Why not? Even with Rybka 3 vs the best engine around, i don't see why a human with good chess understanding would not win a match, because human understanding will allow him to think of very clever and subtle ideas whilst constantly looking forward several moves to check it out, before the human making his move. Therefore there will be no big mistakes plus great human brilliance and depth and wisdom.

Well, it seemed the other way around, the GM (having Rybka 3) wondered at Stockfish 5 "brilliance and depth and wisdom".

Nice test, anyway, it seems the engines are so far ahead, that Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Thanks for pointing this out to me. I admit i hadn't realized this.

It is hard to believe it should have been like this, because i thought that some humans can add huge amounts to the elo/iQ of a program. But maybe 1).that is no longer true 2).Rybka 3 is different 3). This GM is different.

Most of the GMs add nada to the chess engines if any Shimon....

This is because the nowadays chess engines play in a completely different league of their own regards,
Dr.D

[S.Taylor]

Why not? Even with Rybka 3 vs the best engine around, i don't see why a human with good chess understanding would not win a match, because human understanding will allow him to think of very clever and subtle ideas whilst constantly looking forward several moves to check it out, before the human making his move. Therefore there will be no big mistakes plus great human brilliance and depth and wisdom.

Well, it seemed the other way around, the GM (having Rybka 3) wondered at Stockfish 5 "brilliance and depth and wisdom".

Nice test, anyway, it seems the engines are so far ahead, that Rybka 3 scored no better with a GM than it would have scored without a GM. Pretty amazing and educative.

Thanks for pointing this out to me. I admit i hadn't realized this.

It is hard to believe it should have been like this, because i thought that some humans can add huge amounts to the elo/iQ of a program. But maybe 1).that is no longer true 2).Rybka 3 is different 3). This GM is different.

Most of the GMs add nada to the chess engines if any Shimon....

This is because the nowadays chess engines play in a completely different league of their own regards,
Dr.D

I would imagine that I myself, would add a lot to a program alone, if i was in the mood, because i have extreme intuition and imagination, and it's only my lack of time and opportunity, and maybe concentration and training, that i didn't actualize great chess achievement. But if i am in the mood, i feel i would reach high up with the computer backing.

But then again, i might not have enough time during an over the board game, to look down every ally way in great depth and wait until i see if the computer increases evaluation at the end of my sequences or not.

I used to do this a lot with much older computers, and i sometimes had much greater results than i or the computer would have had alone.

I don't know if it is a sign of getting older or not, but right now, the craze over creating a beautiful work of art in chess, doesn't mean so much to me now as it used to. Though i would be in cuckoo land if it still did!

(wow! that was too many I's It's just ideas which might be interesting food for thought to anyone. But still, I'm very happy if Stockfish 5+ is so much greater than rybkas 3 and 4 etc. Although it is a bit too late for me now. I would have LOVED to have todays SF, 10 years ago or more. If it was as user friendly as Hiarcs 7 was, that is.)

[Nitro]

Hi! I'm Jesse, one of the organizers of this match.

Since some of you asked: Daniel was assisted by Rybka 3 on a late-2008 MacBook. I made the moves on the board on behalf of my current 8-core Mac Pro running the July 18th dev version of Stockfish (which I downloaded from github and compiled on OS X -- by the way, the Haswell extensions don't seem to work on OS X for me).

I gave Stockfish no opening book to give Daniel a bit of an extra advantage (I was quite sure he'd need any extra help he could get; Stockfish is just too good, plus the hardware difference is roughly equally significant here). No TBs were used on either engine. I gave Stockfish 8 threads and 8GB of hash.

If you have other questions about the match, let me know. Huge thanks and congratulations to all of the Stockfish developers and FishCooking CPU contributors for an amazing engine.

[Uri Blass]

Hi! I'm Jesse, one of the organizers of this match.

Since some of you asked: Daniel was assisted by Rybka 3 on a late-2008 MacBook. I made the moves on the board on behalf of my current 8-core Mac Pro running the July 18th dev version of Stockfish (which I downloaded from github and compiled on OS X -- by the way, the Haswell extensions don't seem to work on OS X for me).

I gave Stockfish no opening book to give Daniel a bit of an extra advantage (I was quite sure he'd need any extra help he could get; Stockfish is just too good, plus the hardware difference is roughly equally significant here). No TBs were used on either engine. I gave Stockfish 8 threads and 8GB of hash.

If you have other questions about the match, let me know. Huge thanks and congratulations to all of the Stockfish developers and FishCooking CPU contributors for an amazing engine.

I wonder what was the speed advantage of stockfish.
I did not know that stockfish had hardware advantage and based on this knowledge I suspect that rybka3 could probably score not better in similar conditions so all the claims that rybka3 could perform better than rybka3+GM are wrong.

[Nitro]

The 8-core Mac Pro would be somewhere around 16x as powerful as the MacBook (Rybka 3 ran on one core). That's a rough estimate based on the Mac Pro having 8x as many active cores, and ~2x the performance per core (Ivy Bridge Xeon @ 3Ghz vs Core2 @ 2Ghz).

That performance difference plus the engine strength difference indicates that a score of 3.5/4 is exactly in line with what you'd expect just based on the two computers alone. So in this case, there's not much evidence that the GM helped or hurt Rybka. Of course, with only 4 games, it's very hard to draw any conclusions with confidence -- except that a human GM plus the best chess engine from 2008 on 2008 hardware is almost certainly much inferior to the best chess engine from today on modern hardware.

That conclusion should be unsurprising to those who closely follow computer chess. (Indeed, I am of the opinion that rating lists typically understate engine strength relative to their human counterparts; that is, an engine rated at 2500 Elo on CCRL is probably superior to a 2500 Elo human.) And with the latest version of Stockfish on 8 modern Xeon cores, we're likely pushing ~3600 Elo -- a level of skill which is hard to fathom in human terms. So it was not surprising to me that the match turned out the way it did, but some IMs and GMs I spoke to before the match really thought Daniel not only had a chance, but even an advantage!