IvanHoe matches

[kingliveson]

4'2"/ i7-940 @ 3.75 /3 cores ea / 512 mb / ponder off / klo openings

IvanHoe 9.57b x64..................76.5...............+60,=33,-7

Shredder 12..........................23.5................+7,=33,-60

Thanks for the results; are tablebases in use?

[JackStraw]

IvanHoe 9.57b x64..................76.5...............+60,=33,-7

Shredder 12..........................23.5................+7,=33,-60


Yes , using 3-4-5 egtb for Rybka / Shredder / Naum , and RobboBases for IvanHoe.

I sometimes wonder about RobboBases......... I use TripleBases + TotalBases , and occasionally they will show +100 , or +200 , and then a couple moves later it will be +1 , or +2........... What's up with that ??

[kingliveson]

.
.
.

I sometimes wonder about RobboBases......... I use TripleBases + TotalBases , and occasionally they will show +100 , or +200 , and then a couple moves later it will be +1 , or +2........... What's up with that ??

If there is a game to use to analyze the behavior, it would help.

[Charles]

Arena +3s 256hash Ivan=64ph 4 cpu ponder=off Nunn 20 position suite 32 bit

Code: Select all

Ivan57cx32SSE3 - Rybka 4 w32 : 22.5/40 13-8-19 (==1=1===00==1011=0=101=1===10==10=11=10=)  56%  

Same conditions but using a different suite mfl 30 positions:

Code: Select all

Ivan57cx32SSE3 - Rybka 4 w32   : 26.0/60 9-17-34 (00=1====0===110=11===10===0===1000==1=0==00=00=======000===1)  43%  
IvanHoe T63B w32 - Rybka 4 w32 : 24.5/60 11-22-27 (01=1==01==01==0===0=01=0===00=0==1000=001=0=00==01=10=010==1)  41%  

Total 100 game vs R4 result for Ivan57cx32SSE3: 22-25-53


So looks like Rybka 4 is still slightly better.


Note: I think the 0+3s makes R4 stronger because it has less time management issues than with 5/0 4/0 type controls.
Would be interesting to find out if R4 at default is best at this time control or if a bit more tweaking gives R4 more strength.

[kingliveson]

I am actually fine tuning IvanHoe time management settings that counters Rybka recommended settings and beats Rybka. Will post my findings when done.