SCCT Rating List (22.06.2012)

[Sedat Canbaz]

Hello Chess Friends,

New games are added...

And very soon each participant will be based on 1000 games per player

Rating List:
http://www.sedatcanbaz.com/chess/scct-rating/

Conditions:
http://www.sedatcanbaz.com/chess/2838-2/

Games:
http://www.sedatcanbaz.com/chess/games/scct_3m2s.rar

Best Regards,
Sedat Canbaz

[Izak Pretorius]

Hi Sedat

Interesting to see a mod of the top engine scoring 20 ELO higher than original after around a 1000 games.This shows there is still much room for improvement in the top engines if only changing some values can add 20 elo's.So i would not be surprised if Houdini 3 will be 40 elo or more stronger than Houdini 2.

Anyway,i would like to thank you for you rating list and your efforts and i love visiting it

Keep up the good work.

Best Regards
Izak

[Houdini]

I've tested the "z" settings with 27,000 games and haven't found any strength improvement compared to the standard settings.
In the tests of others I haven't seen any statistically significant result either, the difference always lies within the error margins. This is also the case in the SCCT rating list.

I'd love to gain some easy points this way, but so far it doesn't look convincing. Please prove me wrong .

Robert

[Sedat Canbaz]

Hi Sedat

Interesting to see a mod of the top engine scoring 20 ELO higher than original after around a 1000 games.This shows there is still much room for improvement in the top engines if only changing some values can add 20 elo's.So i would not be surprised if Houdini 3 will be 40 elo or more stronger than Houdini 2.

Anyway,i would like to thank you for you rating list and your efforts and i love visiting it

Keep up the good work.

Best Regards
Izak

Dear Izak,

Many thanks for your interest and kind words

And for me its a big pleasure of testing your Ivanhoe complies

About Houdini 2.0z performance,
It seems Houdini is a very potential engine
To be honest,i did no expect v2.0z to be ranked higher than v2.0c

But however,(for more serious conclusion) we need more testings under different conditions

Best,
Sedat

[Sedat Canbaz]

I've tested the "z" settings with 27,000 games and haven't found any strength improvement compared to the standard settings.
In the tests of others I haven't seen any statistically significant result either, the difference always lies within the error margins. This is also the case in the SCCT rating list.

I'd love to gain some easy points this way, but so far it doesn't look convincing. Please prove me wrong .

Robert

Dear Robert,

I guess you tested "z" settings with ultra fast time control, right ?!
There is no doubt that you used different openings than mine, right ?!
I mean that your conditions are not same as mine, right ?!

In other words,its quite normal to see different standings due to different conditions

Btw,more results by Houdini v2.0z in TEB Rating List:
http://www.talkchess.com/forum/viewtopic.php?t=44131

Best,
Sedat

[Houdini]

Dear Robert,

I guess you tested "z" settings with ultra fast time control, right ?!
I mean that your conditions are not same as mine,right ?!

In other words,its quite normal to see different standings due different conditions

Btw,more results by Houdini v2.0z in TEB Rating List:
http://www.talkchess.com/forum/viewtopic.php?t=44131

Best,
Sedat

Hello Sedat, both SCCT and TEB show that the versions are equally strong within the respective uncertainty margins of the two lists.
There is a suspicion that the "z" settings could be better, but so far no conclusive evidence has been presented.
Believe me, I'd be the first to welcome some additional, free Elo points for Houdini, but it's usually not as simple as that.

Robert

P.S. In the past people have reported up to 30 Elo points difference between Houdini 2.0/2.0b or 2.0c, when it is a fact that these versions have perfectly identical strength.

[Sedat Canbaz]

Dear Robert,

I guess you tested "z" settings with ultra fast time control, right ?!
I mean that your conditions are not same as mine,right ?!

In other words,its quite normal to see different standings due different conditions

Btw,more results by Houdini v2.0z in TEB Rating List:
http://www.talkchess.com/forum/viewtopic.php?t=44131

Best,
Sedat

Hello Sedat, both SCCT and TEB show that the versions are equally strong within the respective uncertainty margins of the two lists.
There is a suspicion that the "z" settings could be better, but so far no conclusive evidence has been presented.
Believe me, I'd be the first to welcome some additional, free Elo points for Houdini, but it's usually not as simple as that.

Robert

P.S. In the past people have reported up to 30 Elo points difference between Houdini 2.0/2.0b or 2.0c, when it is a fact that these versions have perfectly identical strength.

Hello Robert,

In my testings,Houdini 2.0c performed better than Houdini 2.0b
And in my opinion,the Elo difference is appeared due to Houdini 2.0b endgame bug

Best Regards,
Sedat

[velmarin]

Hello, Sedat,

Very intesetante your rating, how all the things you do for Chess.
After participating in one of your Books tournaments, a great time.

I wanted to express my opinion on the inclusion of a Mod on the list,
I think it's OK to test, but I think the inclusion of a MOD devalues ​​the overall rating.

What happened when they move all MODs that are preparing the same author and has outstanding other supposedly better.

I think it is not serious, and what better stop sooner rather than later.

Just an opinion.
I will keep eye on all things with admiration.
Greetings.

[Sedat Canbaz]

Hello, Sedat,

Very intesetante your rating, how all the things you do for Chess.
After participating in one of your Books tournaments, a great time.

I wanted to express my opinion on the inclusion of a Mod on the list,
I think it's OK to test, but I think the inclusion of a MOD devalues ​​the overall rating.

What happened when they move all MODs that are preparing the same author and has outstanding other supposedly better.

I think it is not serious, and what better stop sooner rather than later.

Just an opinion.
I will keep eye on all things with admiration.
Greetings.

Dear Jose,

Thank you for nice words and interesting comments

Yes...i still remember your Top book performance in SCCT,nice job !

Some notes about the current situation,
In my current testing (also in TEB),we noticed that Houdini 2.0z performed better than Houdini 2.0c
Robert Houdart claims that v2.0z is not stronger than v2.0c
The creator of "z" settings (Salvador) claims just opposite

Actually they both can be right,but only for their testing conditions
I mean,any engine Elo performance is highly depending on different conditions

One thing more about the current SCCT rating,
I noticed that the results are in favor for v2.0c (Houdini 2.0c vs Houdini 2.0z)
But it seems, Houdini 2.0z has better overall performance

Individual statistics by EloStat 1.3:

Code: Select all

1 Houdini 2.0z Pro x64 6c   : 3394  1050 (+625,=341,- 84), 75.8 %


Houdini 2.0c Pro x64 6c       :  50 (+ 11,= 24,- 15), 46.0 %
Stockfish 120430P x64 6c      :  50 (+ 23,= 20,-  7), 66.0 %
Zappa Mexico II x64 6c        :  50 (+ 38,= 11,-  1), 87.0 %
Deep Junior 13 x64 6c         :  50 (+ 35,= 10,-  5), 80.0 %
Spike 1.4 Leiden w32 6c       :  50 (+ 39,= 10,-  1), 88.0 %
Spark 1.0 x64 6c              :  50 (+ 40,= 10,-  0), 90.0 %
Ivanhoe B46fC x64 6c          :  10 (+  4,=  6,-  0), 70.0 %
Vitruvius 1.11C x64 6c        :  50 (+ 18,= 28,-  4), 64.0 %
Deep Rybka 4.1 x64 6c         :  50 (+ 15,= 28,-  7), 58.0 %
Naum 4.2 x64 6c               :  50 (+ 38,=  9,-  3), 85.0 %
Critter 1.4 x64 6c            :  50 (+ 15,= 28,-  7), 58.0 %
Fire 2.2 xTreme x64 6c        :  50 (+ 19,= 24,-  7), 62.0 %
Protector 1.4.0 x64 6c        :  50 (+ 37,= 12,-  1), 86.0 %
Chiron 1.1a x64 6c            :  50 (+ 44,=  4,-  2), 92.0 %
Deep Fritz 12 w32 6c          :  50 (+ 39,= 10,-  1), 88.0 %
Hiarcs 13.2 w32 6c            :  50 (+ 45,=  4,-  1), 94.0 %
Komodo 4.0 x64 1c             :  50 (+ 31,= 14,-  5), 76.0 %
Deep Shredder 12 x64 6c       :  50 (+ 38,= 10,-  2), 86.0 %
Equinox 1.35 x64 6c           :  40 (+ 27,= 12,-  1), 82.5 %
Ivanhoe B46fE.02 x64 6c       :  50 (+ 20,= 25,-  5), 65.0 %
Deep Fritz 13 w32 6c          :  50 (+ 34,= 16,-  0), 84.0 %
Critter 1.6 x64 6c            :  50 (+ 15,= 26,-  9), 56.0 %

2 Houdini 2.0c Pro x64 6c   : 3374  1051 (+584,=388,- 79), 74.0 %

Houdini 2.0z Pro x64 6c       :  50 (+ 15,= 24,- 11), 54.0 %
Stockfish 120430P x64 6c      :  45 (+ 12,= 23,- 10), 52.2 %
Zappa Mexico II x64 6c        :  46 (+ 38,=  8,-  0), 91.3 %
Deep Junior 13 x64 6c         :  46 (+ 33,= 11,-  2), 83.7 %
Spike 1.4 Leiden w32 6c       :  45 (+ 31,= 14,-  0), 84.4 %
Spark 1.0 x64 6c              :  46 (+ 31,= 12,-  3), 80.4 %
Ivanhoe B46fC x64 6c          :  46 (+ 16,= 24,-  6), 60.9 %
Vitruvius 1.11C x64 6c        :  46 (+ 18,= 24,-  4), 65.2 %
Deep Rybka 4.1 x64 6c         :  46 (+ 14,= 25,-  7), 57.6 %
Naum 4.2 x64 6c               :  46 (+ 27,= 16,-  3), 76.1 %
Critter 1.4 x64 6c            :  46 (+ 20,= 19,-  7), 64.1 %
Fire 2.2 xTreme x64 6c        :  45 (+ 11,= 29,-  5), 56.7 %
Protector 1.4.0 x64 6c        :  46 (+ 33,= 12,-  1), 84.8 %
Fruit 090705 x64 6c           :  44 (+ 41,=  3,-  0), 96.6 %
Chiron 1.1a x64 6c            :  46 (+ 35,= 10,-  1), 87.0 %
Deep Fritz 12 w32 6c          :  40 (+ 26,= 12,-  2), 80.0 %
Hiarcs 13.2 w32 6c            :  45 (+ 30,= 14,-  1), 82.2 %
Komodo 4.0 x64 1c             :  46 (+ 25,= 20,-  1), 76.1 %
Deep Shredder 12 x64 6c       :  45 (+ 36,=  9,-  0), 90.0 %
Equinox 1.35 x64 6c           :  46 (+ 32,= 13,-  1), 83.7 %
Ivanhoe B46fE.02 x64 6c       :  40 (+ 14,= 24,-  2), 65.0 %
Deep Fritz 13 w32 6c          :  50 (+ 37,= 10,-  3), 84.0 %
Critter 1.6 x64 6c            :  50 (+  9,= 32,-  9), 50.0 %

Best,
Sedat