Just a small tournament

[Xaake]

Just ran a small tournament in arena. FB and Protector used end game tablebases. (Not sue about Rybka 2, if it is even capable of using them, but in case it could it had them too (general Nalimov path was set).)

All engines used the same opening book (Mainbook 7moves.abk)

Stockfish, Rybka 2 and Protector seems to be quite even, Firebird way ahead (it surprises me since I thought Stockfish would be closer to Firebird, maybee I did something wrong there... or maybe it is not the case when running on a 32bit OS)

Code: Select all

Engine                 Score          Fi         St         Ry         Pr         Fr         SP    S-B
1: FireBird_12_w32        41,0/50 ·········· 1=10011110 1111=011=1 011==111== 1111111111 =111111111  842,00
2: Stockfish-163-32-ja    29,5/50 0=01100001 ·········· 1===000=10 ====1===== 111=1=01== 111111=111  591,25
3: Rybka v2.2n2.mp.w32    28,5/50 0000=100=0 0===111=01 ·········· 01=0==0==0 1=111111=1 =1111=1=1=  556,50
4: Protector_Win32        28,0/50 100==000== ====0===== 10=1==1==1 ·········· 1==10==11= 11=1010111  598,00
5: Fruit-2-3-1            15,5/50 0000000000 000=0=10== 0=000000=0 0==01==00= ·········· 11=0111=11  275,00
6: SP-x86-Active--Thinker 7,5/50  =000000000 000000=000 =0000=0=0= 00=0101000 00=1000=00 ··········  193,25

Level: Tournament 40/10 Rest:5
Hardware: Intel(R) Core(TM)2 Duo CPU     T7700  @ 2.40GHzDual  2393 MHz with 3 056 MB Memory
Operating system: Microsoft Windows XP Professional Service Pack 3 (Build 2600)

[Carlos777]

Hi Rickard,

Nice tournament. Firebird is indeed a strong engine, so no surprises here.

I guess Protector was v1.34? Engines were able to use 2 cores?

OTOH, I recommend the use of Thinker Passive, it gets better results than the Active one.

Regards,
Carlos

[alpha123]

Hi Rickard,

Nice tournament. Firebird is indeed a strong engine, so no surprises here.

I guess Protector was v1.34? Engines were able to use 2 cores?

OTOH, I recommend the use of Thinker Passive, it gets better results than the Active one.

Regards,
Carlos

And Inert gets better results than Passive and Active.

Peter

[Carlos777]

Hi Rickard,

Nice tournament. Firebird is indeed a strong engine, so no surprises here.

I guess Protector was v1.34? Engines were able to use 2 cores?

OTOH, I recommend the use of Thinker Passive, it gets better results than the Active one.

Regards,
Carlos

And Inert gets better results than Passive and Active.

Peter

What I was thinking about?

Peter is right. Inert is the best.

[alpha123]

Hi Rickard,

Nice tournament. Firebird is indeed a strong engine, so no surprises here.

I guess Protector was v1.34? Engines were able to use 2 cores?

OTOH, I recommend the use of Thinker Passive, it gets better results than the Active one.

Regards,
Carlos

And Inert gets better results than Passive and Active.

Peter

What I was thinking about?

Peter is right. Inert is the best.

Eh, I just go by CCRL . I have never really tested anything other than Inert myself .

Peter

[Xaake]

Thanks, I will use Thinker Inert next time. Protector was 1.3.4 but I just realized it used only one core but it had ponder on, the others did not, I will correct that in next tournament.

Browsing thorugh the UCI settings for the engines I used I can see that there is some tweaking needed to make this tournament fair. I will rerun it maybee this weekend and then do some tweaking like:

Making sure Ponder is off
Making sure all engines that are capable uses tablebases
Making sure all engines that are capable uses 2 cores
Using Thinker Inert
Using Rybka 2.3.2

[Frank Quisinsky]

Hi there,

after all my experience it's not very interesting to test such a setting.

I played in the past with active also 500 games. The result is the same as for Inert. Same for all the Cyclone versions. Or Hiarcs Sharpen PV parameter. Can't not see clearly better results.

More interesting as to test such things is to test an other engine!

After all:
Thinker is very strong. Its a pity that I can't see the information about the analyses.

Best
Frank