OliThink 5.2.7 released

[OliverBr]

As usual you can download it here:

http://home.arcor.de/dreamlike/chess/index.html

The point of this release is the good performamce against Crafty 20.14, playing 1000 fast blitz games, so we have a good samplesize.

Code: Select all

olithink527 - crafty20.14 : 402.0/1000 328-148-524 40%
olithink526 - crafty20.14 : 373.5/1000 303-141-556 37%
olithink525 - crafty20.14 : 345.0/1000 266-158-576 34%
olithink519 - crafty20.14 : 319.5/1000 255-129-616 31%
olithink517 - crafty20.14 : 273.5/1000 216-115-669 27%
olithink411 - crafty20.14 : 104.5/1000 68-73-859 10%

PS: For those who are interested, the difference between v526 and v527:

In v526 we have in line 1093:

Code: Select all

if (m) ppos += 8;

this became in v527:

Code: Select all

if (m) ppos += 8; else ppos -= 8;

That's all.

[Graham Banks]

Hi Oliver,

how much difference do you estimate it will make in terms of strength?
I thought that OliThink 5.2.3 was the strongest recent release. Did you try that version out under the same conditions? Could be interesting.

Cheers,
Graham.

[OliverBr]

Hi Oliver,

how much difference do you estimate it will make in terms of strength?
I thought that OliThink 5.2.3 was the strongest recent release. Did you try that version out under the same conditions? Could be interesting.

Cheers,
Graham.

I am not sure as I am mainly testing with Crafty 20.14. It's a little challenge for me to catch.
I know that in your tourney 5.2.3 performed much better than 5.2.5, but this can be just accidental as 30 games are not too much.
It could be as well, that 5.2.5 or above performas better at blitz, while 5.2.3 does better at slower controls... I don't know :/

[Graham Banks]

Hi Oliver,

how much difference do you estimate it will make in terms of strength?
I thought that OliThink 5.2.3 was the strongest recent release. Did you try that version out under the same conditions? Could be interesting.

Cheers,
Graham.

I am not sure as I am mainly testing with Crafty 20.14. It's a little challenge for me to catch.
I know that in your tourney 5.2.3 performed much better than 5.2.5, but this can be just accidental as 30 games are not too much.
It could be as well, that 5.2.5 or above performas better at blitz, while 5.2.3 does better at slower controls... I don't know :/

You're correct in that the sample of games was small. It was mainly just my own observations from looking through the games.
I'll stick with 5.2.6 in my current tourney (with only eight rounds left), but will use the latest version in any other testing.

Cheers,
Graham.