IvanHoe V48 tests

!0 minute games

[You must be registered and logged in to see this link.]


1 Ivanhoe B48iX64 +16 +2/=19/-1 52.27% 11.5/22
2 Houdini 1.5 x64 -16 +1/=19/-2 47.73% 10.5/22

test conditons

Hash 512 megs each
LargePages = OFF
2-3-4-5-Z Bases loaded in memory
6 men (3+3 only used)
TripleWeakHeight=5
TripleHashSize=512
DynamicLoadTripleCacheSize=512
LoadonWeakProbe checked.
Perfect 2010 up to move 7

As Houdinis Large Pages are disabled NO large pages were used on IvanHoe
As Houdini has built in end game base a minimun of RobboBases was enabled for IvanHoe

fair is fair

the much vaunted power of Houdini seems to primerily be vey short time control games , though we do see great improvent in more than its speed and expect Ippolit to continue to improve in some aspects of search and evaluation that have taken a back seat to the building of 6 man bases and 2 annalise GUIs



That said we can see clearly that IvanHoe is still the world leader in computer chess , those fortunate enogh to have new bases should test at even longer time controls,,we arent so blessed as to be able to as we do not have them

[img][You must be registered and logged in to see this link.][/img]

ANYONE doubting our condemation of the last 2 source codes may like to know this is peter pans best errort at compileing the 999948n codes...it beats KLO compile but,,,

the test continued and was stopped at +13 = 20 -7 in favour of dr4 as it was simply to painfull to watch
ivan even lost the above game

[img][You must be registered and logged in to see this link.][/img]

Another one bites the dust

usuall test conditions

Because it is Ippolits Aniversry and izak was a friend once i gave this engine every chance in a longer test than was realy nessacry
some will say 100 games is insufficent a test , then when 100 are done they demand a 1000 , they are fools
one can judge an engine by the way it plays in less than even 50 games, this one is judged as "Unfit for purpose"

unless the purpose was to please Rybka and co

to even consider releasing an engine without testing it first is either complete arrogence , desperation for adoration, or simply a lack of care for end users , a simple test would have shown it has issues , Not only weaknesses , if you want an engine that might,,,might draw with critter or stockfish then by all means use this one

[img][You must be registered and logged in to see this link.][/img]

"B" wrote:

Because it is Ippolits Aniversry and izak was a friend once i gave this engine every chance in a longer test than was realy nessacry
some will say 100 games is insufficent a test , then when 100 are done they demand a 1000 , they are fools
one can judge an engine by the way it plays in less than even 50 games, this one is judged as "Unfit for purpose"

unless the purpose was to please Rybka and co

to even consider releasing an engine without testing it first is either complete arrogence , desperation for adoration, or simply a lack of care for end users , a simple test would have shown it has issues , Not only weaknesses , if you want an engine that might,,,might draw with critter or stockfish then by all means use this one

[img][You must be registered and logged in to see this link.][/img]

Great test "B"

We did our own . and followed your first to 10 wins idea ,obut only 1 min blitz but as you can see it shows why no one should release engines without testing them first and peter pants should know better , he just keeps embaressing himself and feeding rybka presents
TripleBases dont work on this compile either,but it would still be one for scrap yard even if they did

[img][You must be registered and logged in to see this link.][/img]

[img][You must be registered and logged in to see this link.][/img]



[img][You must be registered and logged in to see this link.][/img]


Both tests ...NO bases NO large pages Ponder Off

Thank you "B" , very good testings is safe from latest ivans