Fritz 11 = Rybka3 at what ply?

[Kohflote]

Hi all,

OK, Rybka 3 is now the strongest. Just curious and wonder has anybody experiment it: 1 cpu, 32bit, say Fritz 11 is playing at ply 20, it would be about the same strength with Rybka 3 at what ply?

Best Regards,
Kah Huat, Koh
- from Singapore

[Dann Corbit]

Hi all,

OK, Rybka 3 is now the strongest. Just curious and wonder has anybody experiment it: 1 cpu, 32bit, say Fritz 11 is playing at ply 20, it would be about the same strength with Rybka 3 at what ply?

Best Regards,
Kah Huat, Koh
- from Singapore

This is a much harder question than you imagine.
We have Elo figures from:
http://www.computerchess.org.uk/ccrl/40 ... t_all.html
as:

Code: Select all

Name                Rating
Rybka 3 64-bit 4CPU 3229
Fritz 11            2962

That gives an Elo difference of 267. Using the crude assumption of 50 Elo per ply that is a 5 ply advantage.

That is only the crudest imaginable estimate, and actual testing would be needed to find the actual value.

[Dr.Ex]

Hi all,

OK, Rybka 3 is now the strongest. Just curious and wonder has anybody experiment it: 1 cpu, 32bit, say Fritz 11 is playing at ply 20, it would be about the same strength with Rybka 3 at what ply?

Best Regards,
Kah Huat, Koh
- from Singapore

Rybka reports wrong depths. I think depth 1 = depth 4 in reality. So you can test for example Fritz 11 at depth 4 against Rybka at depth 1 in the latest Chessbase GUIs.

[Dr.Ex]

Hi all,

OK, Rybka 3 is now the strongest. Just curious and wonder has anybody experiment it: 1 cpu, 32bit, say Fritz 11 is playing at ply 20, it would be about the same strength with Rybka 3 at what ply?

Best Regards,
Kah Huat, Koh
- from Singapore

This is a much harder question than you imagine.
We have Elo figures from:
http://www.computerchess.org.uk/ccrl/40 ... t_all.html
as:

Code: Select all

Name                Rating
Rybka 3 64-bit 4CPU 3229
Fritz 11            2962

That gives an Elo difference of 267. Using the crude assumption of 50 Elo per ply that is a 5 ply advantage.

That is only the crudest imaginable estimate, and actual testing would be needed to find the actual value.

He wrote 1 cpu,32-bit not 4 CPU, 64-bit.

[Dann Corbit]

Hi all,

OK, Rybka 3 is now the strongest. Just curious and wonder has anybody experiment it: 1 cpu, 32bit, say Fritz 11 is playing at ply 20, it would be about the same strength with Rybka 3 at what ply?

Best Regards,
Kah Huat, Koh
- from Singapore

This is a much harder question than you imagine.
We have Elo figures from:
http://www.computerchess.org.uk/ccrl/40 ... t_all.html
as:

Code: Select all

Name                Rating
Rybka 3 64-bit 4CPU 3229
Fritz 11            2962

That gives an Elo difference of 267. Using the crude assumption of 50 Elo per ply that is a 5 ply advantage.

That is only the crudest imaginable estimate, and actual testing would be needed to find the actual value.

He wrote 1 cpu,32-bit not 4 CPU, 64-bit.

I thought that this was one logical entity:
"1 cpu, 32bit, say Fritz 11"

[Kohflote]

Hi all,

Thank you for your quick reply. And my apologies for not being clear. I do mean comparing both Fritz 11 and Rybka3 on 1cpu, 32bits.


Best Regards,
Kah Huat, Koh
- from Singapore

[hgm]

Going by the depth that engines report or obey is really a meaningless exercise, as every engine measures depth differently. A meaningful question would be how much shorter time Rybka needs to reach an equivalent playing strength a Fritz 11.

This you can simply determine by playing a time-odds match, e.g. 40 moves/12 min, and specifying a time-odds factor of 12 for Rybka, so that it has only 1 min for 40 moves. (without ponder, of course). Keep increasing the time-odds factor as long a Rybka is still winning.

[bob]

Going by the depth that engines report or obey is really a meaningless exercise, as every engine measures depth differently. A meaningful question would be how much shorter time Rybka needs to reach an equivalent playing strength a Fritz 11.

This you can simply determine by playing a time-odds match, e.g. 40 moves/12 min, and specifying a time-odds factor of 12 for Rybka, so that it has only 1 min for 40 moves. (without ponder, of course). Keep increasing the time-odds factor as long a Rybka is still winning.

SInce you have thought about this, why don't you do it? It would be interesting to have such data, all run thru (say) BayesElo, to give a good estimate of 3:1 time odds (no pondering) is worth X +/- N Elo.

Actually, this might be something I could do on our cluster later on. It would definitely get rid of the old "doubling is 50 elo" or whatever rule-of-thumb that doesn't seem very accurate to me...

[hgm]

SInce you have thought about this, why don't you do it?

Because I have neither Fritz nor Rybka...

Neither am I particularly interested in how they play. I have done time-odds tourneys involving my own engines often enough. The bigger one I did for Gothic Chess, where I had 8 engines, some of them with 6 different handicap factors (1, 3, 9, 24, 54 and 90), is currently not on my webite. (I made a bit of a mess of the PGN file, replacing some engines half-way the tourney, and never got to cleaning it up.) A smaller Knightmate tourney, where I had handicapped the few available engines to make them close matches, can be found here. In that tourney I used handicaps factors 12 and 144, in 40/12' games.

Since in that tourney each engine participated only with one handicap factor, I could of course not derive the Elo-time relation, which I could for the Gothic Chess tourney. So perhaps I will dust off the results of that one.

[hgm]

Some discussion about the big time-odds tourney can be found here (not the full cross table or ratings, though):

http://64.68.157.89/forum/viewtopic.php ... t&start=34