CyclonexTremeII : Stockfish1.4 (3 CPU) finished: 26,5 : 43,5

[rainhaus]

Cyclone xTremeII : Stockfish 1.4x64

games 1 -35 score: 15.5 : 19.5 (44% : 56%) lost-won-remis: 12-08-15
games 36-70 score: 11.0 : 24.0 (31% : 69%) lost-won-remis: 17-04-14
_____________________________________________________
total score:..............26.5 : 43.5 (38% : 62%) lost-won-remis: 29-12-29

Data Analysis by EloStat
Hypothetical starting Elo 2960
Elo Cyclone Xtreme II : Stockfish 1.4
2917 : 3003 (+64/-63 confidence interval)

Scores based on one match only, respectively on one opponent, scarcely hit the Elo level of engines within a large-scaled rating list. Such results rather should be understood as an indication if a difference in playing strength might be probable or not. This is true in general. But have a look on CEGT and you see cyclone 3.1 4CPU (no higher version available) with 2920 ELO and Stockfish 1.4 4CPU with 3007 ELO. Not a bad representation. And for good measure, the difference between the two engines is significant on the 95% level, even with this small sample. Cyclone was beaten so hard that even the statistics had to admit this! But of course, we expect from the xTrem version something more. My previous tests with cyclone 4.0 suggests that the authors parameter screwing for the xTremeII version don't fit stockfish. Cyclone 4.0 scored clearly better against Stockfish. But let's see how it goes on with the other opponents. Cyclone may have a breathing pause and should check out his bottom limit. I gave him nice Brigth0.4a and Normans's engine is about to take advantage of this opportunity by leading 33,5 : 22,5 (60%:40%). But shortly after comes Spartan Thinker... Have a look after this one if I'll still be in the mood for the final against free Rybka.

Rainer Neuhäusler
_______
System: core i7 220,oc3600-3800 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.
_______
Link: Game 36-70 (70)
http://www.file-upload.net/download-190 ... 0.pgn.html

[Sylwy]

Cyclone xTremeII : Stockfish 1.4x64

games 1 -35 score: 15.5 : 19.5 (44% : 56%) lost-won-remis: 12-08-15
games 36-70 score: 11.0 : 24.0 (31% : 69%) lost-won-remis: 17-04-14
_____________________________________________________
total score:..............26.5 : 43.5 (38% : 62%) lost-won-remis: 29-12-29

Data Analysis by EloStat
Hypothetical starting Elo 2960
Elo Cyclone Xtreme II : Stockfish 1.4
2917 : 3003 (+64/-63 confidence interval)

Scores based on one match only, respectively on one opponent, scarcely hit the Elo level of engines within a large-scaled rating list. Such results rather should be understood as an indication if a difference in playing strength might be probable or not. This is true in general. But have a look on CEGT and you see cyclone 3.1 4CPU (no higher version available) with 2920 ELO and Stockfish 1.4 4CPU with 3007 ELO. Not a bad representation. And for good measure, the difference between the two engines is significant on the 95% level, even with this small sample. Cyclone was beaten so hard that even the statistics had to admit this! But of course, we expect from the xTrem version something more. My previous tests with cyclone 4.0 suggests that the authors parameter screwing for the xTremeII version don't fit stockfish. Cyclone 4.0 scored clearly better against Stockfish. But let's see how it goes on with the other opponents. Cyclone may have a breathing pause and should check out his bottom limit. I gave him nice Brigth0.4a and Normans's engine is about to take advantage of this opportunity by leading 33,5 : 22,5 (60%:40%). But shortly after comes Spartan Thinker... Have a look after this one if I'll still be in the mood for the final against free Rybka.

Rainer Neuhäusler
_______
System: core i7 220,oc3600-3800 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.
_______
Link: Game 36-70 (70)
http://www.file-upload.net/download-190 ... 0.pgn.html

Hi !

They are two engines Cyclone xTreme II : timestamped 17 (Rage version )& 20 ( Fury version ) September 2009. The first version has razoring & history research options actives. This version performed much, much better in my tests ( 10'+5" time control ).

Regards,
Silvian

[rainhaus]

Hi ! They are two engines Cyclone xTreme II : timestamped 17 (Rage version )& 20 ( Fury version ) September 2009. The first [version has razoring & history research options actives. This version performed much, much better in my tests ( 10'+5" time control ). Regards,Silvian

Hello, I know, but of course I don't follow the almost daily instructions of the author:...these executables supersede all other releases...
As I said it before, if an author spies out every week a new update there is no serious and continuous test working possible. The fan enjoys it and can't resist : quick a few Blitz games with the new update and you will find soon an opponent against it scores better than the previous one. Norman uses a quite normal market strategy and it's very easy to see through: you have to be present an you have to change the ware in the windows faster as the tests could sit down. I don't like it and it and I'll remain solid on my time stamp, not me but the engine has to suffer:)
But never I will spoil your party. Publish your results in the same way as I do and describe it a little more precise: opponents, Gui, hardware, settings, statistics and downloads. Norman can use it well for Beta testing, perhaps for cyclone "gentle breeze":)

Cheers Rainer

[ernest]

They are two engines Cyclone xTreme II : timestamped 17 (Rage version )& 20 ( Fury version ) September 2009. The first version has razoring & history research options actives.

On the cyclone_xTreme_StormPack.rar I downloaded last night, Fear, Fury, Rage and Wrath are all time-stamped Sept 24, 21:42 to 21:47.
Difficult to follow...

[gerold]

They are two engines Cyclone xTreme II : timestamped 17 (Rage version )& 20 ( Fury version ) September 2009. The first version has razoring & history research options actives.

On the cyclone_xTreme_StormPack.rar I downloaded last night, Fear, Fury, Rage and Wrath are all time-stamped Sept 24, 21:42 to 21:47.
Difficult to follow...

They are all the same.

[Graham Banks]

xTreme II is performing worse than the original xTreme in my testing to date. Still early days though.

[ernest]

They are all the same.

Wrong! : I checked indeed (in console mode) that cyclone xTreme II.exe (17th Sept) and cyclone xTreme Rage.exe behave exactly the same, as indicated in the latest http://www.chesslogik.com/cyclone.htm

[rainhaus]

Norman Schmidt has promised with Cyclone IIxTreme an increase of 50-100 Elo! I'll take him at his word. Which Cyclone xTreme is working in your Division one (time stamp)? I see you're testing Cyclone Xtreme II too, which version, do you add it to CCRL?
Your "Divisions" are marching ahead diligently:) It's unterstandable you've replaced Twisted Logic 20090922 with version 20080620 and it was a good idea to replay all matches.
See you
Rainer

[Graham Banks]

Norman Schmidt has promised with Cyclone IIxTreme an increase of 50-100 Elo! I'll take him at his word. Which Cyclone xTreme is working in your Division one (time stamp)? I see you're testing Cyclone Xtreme II too, which version, do you add it to CCRL?
Your "Divisions" are marching ahead diligently:) It's unterstandable you've replaced Twisted Logic 20090922 with version 20080620 and it was a good idea to replay all matches.
See you
Rainer

Hi Rainer,

the Cyclone xTreme I'm using is timestamped 15/9 and Cyclone xTreme II is timestamped 20/9.
Both are on the CCRL 40/40 list.

Cheers,
Graham.

[Spock]

I was curious to see if the chess960 support actually worked or not on Cyclone Xtreme. It seems that is probably does. I played 100 blitz games and it played just fine. It also did rather well, Cyclone xTreme Fear 53.5 - 46.5 vs Hiarcs 12.1 (both single CPU). I'm going to run a few more games, sometimes it takes engines a while to hit positions/moves where chess960 related bugs surface