CyclonexTremeII: Stockfish1.4 40/10 14,5:19,5 (35/70)

[rainhaus]

current standing:
CyclonexTremeII(3) : Stockfish1,4x64
15.5 : 19.5 (44%:56%);
lost-won-remis: 12-8-15

well, it's not easy testing a cyclone engine at all. One reason is that the author spits out one update after other and you can't finish any tournament with the same engine. Hello Norman, please slow down, collect your new bugs and inspirations and do only one update for three!
I've started a match series with CycloneXtremeII (time stamp Sep 20 2009 06:25:57). Opponents : Stockfish 1.4 , Thinker 5.4Di, Bright 0.4a, Rybka 2.2n2
The author promises an improvement of 50-100 Elo! If he isn't boasting CyclonexTreme II might belong to the 5 best free engines and score better than all the other free (not private!) Fruit derivatives.
System: core i7 220 , oc 3600-3800 MHz, 6GB RAM. Windows Vista 64Bit. Arena Mark 33.73. Fritz11 Mark: 22.61/10.852.000 n/sec.
BIOS setting: Hyperthreading off, Speed Step off, Turbo Boost off.
GUI=Arena2.01. Hash=1024, time control=40/10 min, ponder=off. threads=3
Nalimov 3,4,5 on, BitBases 3,4,5 on.
Stockfish don't use any tablebases at all, Rybka does, Bright does well with the Bitbases, CyclonexTreme II has the bitbase option but I (it) don't succeed to load it. The use of endgame tables often has been discussed. They seem to be essential in official tournaments for example, where one game could decide the ranking but they probably have no significant statistical value for the rating lists. But there are no sufficient experiments or even scientific studies about it. Here you see a position from a game Rybka 2.2n : CycloneXtreme. Cyclone looses with Kd7-c7 ??? because he does not use EGTBs.
[D] b6R/P2k4/4n3/1K6/8/8/8/8 b - - 8 72
(I don' t know if the FEN works, because it is never shown in the preview)
Book: none. 35 own Starting positions (ECO Mainline), to repeat with switched colours. I want to see the naked engine before it is well dressed with an opening book:)

Rainer Neuhäusler

Game 1-35 (70)
http://www.file-upload.net/download-190 ... 5.pgn.html

[Graham Banks]

Looking at the Cyclone xTreme II games in my Division 1 tournament, the evaluations towards the end of the games against Twisted Logic and Scorpio would suggest that the egbbs are loading okay for me.

[rainhaus]

Looking at the Cyclone xTreme II games in my Division 1 tournament, the evaluations towards the end of the games against Twisted Logic and Scorpio would suggest that the egbbs are loading okay for me.

Maybe you have a better bitbase weather in Auckland than I have in my little German village:) We need a clear test position, take the one of my post above. I've tested it in Arena 2.01, Fritz 11 and Shredder 11 and with the engines Brigt0.4a, Cyclone(xtreme, xtremeII), Fritz11, Rybka2.2n2, Shredder11. The result is always the same.
First, unload EGTBs and make the move Kd7-c7. The engines have no idea the game is won (evaluation not over 2.0).
Second, load the Nalimovs and you see at once #29, load the Bitbases and you see at once 50.99 (Bright). Enjoy it and please tell me what's the matter.
Rainer

[Graham Banks]

Looking at the Cyclone xTreme II games in my Division 1 tournament, the evaluations towards the end of the games against Twisted Logic and Scorpio would suggest that the egbbs are loading okay for me.

Maybe you have a better bitbase weather in Auckland than I have in my little German village:) We need a clear test position, take the one of my post above. I've tested it in Arena 2.01, Fritz 11 and Shredder 11 and with the engines Brigt0.4a, Cyclone(xtreme, xtremeII), Fritz11, Rybka2.2n2, Shredder11. The result is always the same.
First, unload EGTBs and make the move Kd7-c7. The engines have no idea the game is won (evaluation not over 2.0).
Second, load the Nalimovs and you see at once #29, load the Bitbases and you see at once 50.99 (Bright). Enjoy it and please tell me what's the matter.
Rainer

When Cyclone gives evaluations like 50.33, I'm taking that as an indication that it is accessing the egbbs.

Cheers,
Graham.

[rainhaus]

Looking at the Cyclone xTreme II games in my Division 1 tournament, the evaluations towards the end of the games against Twisted Logic and Scorpio would suggest that the egbbs are loading okay for me.

Maybe you have a better bitbase weather in Auckland than I have in my little German village:) We need a clear test position, take the one of my post above. I've tested it in Arena 2.01, Fritz 11 and Shredder 11 and with the engines Brigt0.4a, Cyclone(xtreme, xtremeII), Fritz11, Rybka2.2n2, Shredder11. The result is always the same.
First, unload EGTBs and make the move Kd7-c7. The engines have no idea the game is won (evaluation not over 2.0).
Second, load the Nalimovs and you see at once #29, load the Bitbases and you see at once 50.99 (Bright). Enjoy it and please tell me what's the matter.
Rainer

When Cyclone gives evaluations like 50.33, I'm taking that as an indication that it is accessing the egbbs.

Cheers,
Graham.

I'm not quite sure it's so easy. Maybe, but such evaluations and higher you can see with and without bitbases. A good indication is the sudden jump of evaluation. We should find it out. Please give me the number of a Cyclone game of your division1 which shows your "evaluation indication" and I will download it. Meanwhile you could prove the bitbase loading by cyclone at once and 100% in checking my position above. For you only a few clicks. Don't forsake me, as you see, we are the only lonely Cyclone testers here;) If you don't do it, maybe you have a nightmare and you see a horrible cyclone and thousands of tablebase are crawling from under your bed.
cheeerio
Rainer

[Graham Banks]

When Cyclone gives evaluations like 50.33, I'm taking that as an indication that it is accessing the egbbs.

Cheers,
Graham.

I'm not quite sure it's so easy. Maybe, but such evaluations and higher you can see with and without bitbases. A good indication is the sudden jump of evaluation. We should find it out. Please give me the number of a Cyclone game of your division1 which shows your "evaluation indication" and I will download it. Meanwhile you could prove the bitbase loading by cyclone at once and 100% in checking my position above. For you only a few clicks. Don't forsake me, as you see, we are the only lonely Cyclone testers here;) If you don't do it, maybe you have a nightmare and you see a horrible cyclone and thousands of tablebase are crawling from under your bed.
cheeerio
Rainer

There is indeed a sudden jump in evaluation in the cases I've seen.
Here is the link to my Division 1 pgns. I know that there were games against Twisted Logic and Scorpio that showed such behaviour:
http://kirill-kryukov.com/chess/discuss ... f=7&t=4570
I'll have more time to check out your position at the weekend. but perhaps others might like to comment.

Cheers,
Graham.