S.Taylor wrote:Well, so it looks like you are very open to coming second!
Well, in case you are refering to the IPON test - no, or better: unlikly after 2600 games.
K10.0 finished the same opponents with the same openings on the same Hardware with 79.4%. Everything of K10.1 above that will result in an Elo increase over 10.0.
Do not check the online Elo calculation. The input values are ORDO, the live ELO calculation is pure Elo-formula ... that is ok to get an "impression" or for everything in the middle of the range but not the final ORDO rating.
Additionaly please do not check the individual results. That is just 220 games at the end which is not worth considering for a "rating".
On top of that I "fear" that the contempt is not helping against SF either. Go down a few lines on the IPON start page and check the contempt test. Interesting stuff
In case Stuart can't find Ingo's testrun, click on Komodo:
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
I see both images the same size, just 100 x 50 pixels, maybe you get to see an original image that is larger? Aesthetically I like the logo and the colors, it is very peaceful Very Zen.
The result from Komodo 10.1 against Stockfish 7 remains a bit disappointing, especially as I think in TCEC Komodo will do better, if it comes to a direct clash with stockfish. But to be fair, Komodo's results against all the other engines do compensate enough to pass Stockfish again.
The results from Stockfish dev from Jean Paul Vael on high end NUMA machines continue to worry, although people are enthusiastic about some games seen in Clemens Keck's rapid tournament.
Thanks for the list Ingo!
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Eelco de Groot wrote:...
The result from Komodo 10.1 against Stockfish 7 remains a bit disappointing...
Again, just 220 games max (MUCH to less) and check the contempt test. K10 without contempt looks different against SF7 (but thats only 220 games as well)!
Eelco de Groot wrote:I see both images the same size, just 100 x 50 pixels, maybe you get to see an original image that is larger? Aesthetically I like the logo and the colors, it is very peaceful Very Zen.
Yes the size is 100x50 but my browser shows i like this:
and yes, I 'could' have seen it but I did not paid attention until it was suddenly bigger than I was used to. It is my Komodo folder for ages ... obviously.
Anyhow, any other nice and more corrent K-logo available?
Ingo
PS: and the picture I included here looks even bigger
Here is a logo that could be used for personal use only, because it is copied from a commercial logo (so copyright applies) and I changed the colors a bit:
100 x 50 version:
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Mark,
If one has an SSD for Syzygy TBs, should one use "Smart Syzygy"?
As for the default 28 for "Null Scale", would that also be best for analysis at long times (30 minutes)?
BTW, I notice that "Magnify" has disappeared.
