Amazing How This Engine Plows Thru the Competition!

[geots]

The great majority of these matches I posted as they were played- with one match in each thread. But that way you do not get the full benefit of exactly how devastating this engine can be. I posted, I believe, something similar to this- but a few more matches have been added. And it is a version that I and one other- maybe 2 other- people have been allowed to test.

Most of these matches were either at 40/3 repeating or 4m+2s blitz. The one 40/40 match is labeled as such. I have match results with Komodo beta 4536- but I am only dealing here with official releases. It, being Strelka 5.6- has yet to meet Komodo CCT 64-bit, but have no fear- the match is coming.

Also, please note the vast improvement from Stockfish 2.3.1 as compared to the latest official release, 3.0 SSE4.2, as compiled by Jim.

I am posting this as is, because some did not see the first post before more were added- and it gives you a whole different outlook when you see it all on one page. I thought it might be interesting to some. And if not for Houdini................................. (Unless denoted, all matches were 100 games- the only losses being to Houdini!)

Code: Select all

Strelka 5.6 x64: (Only Wins & Losses)


Fire 2.2b xTreme GH x64      31-13    (+63)
Komodo 5 64-bit              38-16    (+78)
Ivanhoe 46h x64              31-5     (+92)
IvanHoe 50kQ x64 SSE4.2      27-15    (+42)
Rybka 4.1 x64                26-19    (+24)
Stockfish 2.3.1 pcnt x64     37-7     (+108)
Critter 1.6a 64-bit          32-25    (+12)---200 games
Houdini 2.0c x64             37-48    (-19)---200 games***
DeepSaros 3.3 pcnt x64       31-11    (+70)
Gull II x64                  39-8     (+111)
Bouquet 1.5 pop64            41-11    (+108)
IvanHoe 46ei x64 SSE4.2      26-17    (+31)
IvanHoe 46e x64 SSE4.2       25-10    (+53)
RobboLito 0.21Q x64 SSE4.2   28-13    (+53)
Vitruvius 1.19 pcnt x64      29-12    (+60)
Stockfish 3 JA 64bit SSE4.2  30-15    (+53)
Houdini 3 x64                17-65    (-85)---200 games***
Houdini 3 x64                 3-12    (-63)---50 games @ 40/40***

As you can see, except for one engine, all the matches Strelka won were basically routs. At any rate, it is time for Strelka 5.6 and Komodo CCT. My prediction- Komodo CCT is now just a bit too strong for Strelka.

I need to do a bit of thinking- either 100 games at 40/3 repeating OR 50 games at 40/40 repeating. I am leaning toward the 40/40. It will start tonight- and this match is definitely and without doubt for No.2 in the world.




Never have your back to the door-

george

[geots]

I call it the "Super Match"- being played for No.2 in the world: Komodo CCT 64-bit vs. Strelka 5.6 x64. It is being played at 40/40 repeating, benchmarked to CCRL and CEGT controls, except that CEGT no longer uses 40/40.

For anyone who thinks 50 games is not enough, let's take every trophy and title away from all the WCCC world champions in history- because none of them in winning played close to 50 games all total.

Next to last- Critter is not in this struggle for the No.2 spot. Simply put, IMO and in my testing- it is not strong enough to fight for No.2

And lastly, 8 games have been played. Updates will come every 10 games.


Best,

[geots]

We have an update in the "Super Match" for No. 2 in the world. With the extremely bad weather heading this way, hopefully they will continue to play if we happen to be in a bathtub with a mattress over us. At any rate 10 of 50 games have completed:



Alienware AURORA_R4
Intel i7 w/6 True Cores
Fritz 11 gui
1CPU/64bit
128MB hash
Bases=NONE
Ponder_Learning=OFF
Perfect 2012b.ctg w/12-move limit
40/19 Repeating (Benched to adapt to 40/40)
Match=50 games

Code: Select all

SP8-i7, 19m/40+19m/40+19m/40  5/21/2013  

                                         
Komodo CCT 64-bit    +4/-1/=5   65.00%   6.5/10
Strelka 5.6 x64      +1/-4/=5   35.00%   3.5/10  

At the 20% mark, nothing to be gained by applying an elo figure to the difference. Save that for later.




Best,

gts

[gerold]

I call it the "Super Match"- being played for No.2 in the world: Komodo CCT 64-bit vs. Strelka 5.6 x64. It is being played at 40/40 repeating, benchmarked to CCRL and CEGT controls, except that CEGT no longer uses 40/40.

For anyone who thinks 50 games is not enough, let's take every trophy and title away from all the WCCC world champions in history- because none of them in winning played close to 50 games all total.

Next to last- Critter is not in this struggle for the No.2 spot. Simply put, IMO and in my testing- it is not strong enough to fight for No.2

And lastly, 8 games have been played. Updates will come every 10 games.


Best,

I don't have Strelka 5.6 to test. Good luck and hoping to get Strelka 5.6 to test as it is one of my favorite engines.
Have you run a test Critter 1.6 vs. Komodo cct. I think Critter is much stronger than the cct.

Best,
Gerold.

[geots]

I call it the "Super Match"- being played for No.2 in the world: Komodo CCT 64-bit vs. Strelka 5.6 x64. It is being played at 40/40 repeating, benchmarked to CCRL and CEGT controls, except that CEGT no longer uses 40/40.

For anyone who thinks 50 games is not enough, let's take every trophy and title away from all the WCCC world champions in history- because none of them in winning played close to 50 games all total.

Next to last- Critter is not in this struggle for the No.2 spot. Simply put, IMO and in my testing- it is not strong enough to fight for No.2

And lastly, 8 games have been played. Updates will come every 10 games.


Best,

I don't have Strelka 5.6 to test. Good luck and hoping to get Strelka 5.6 to test as it is one of my favorite engines.
Have you run a test Critter 1.6 vs. Komodo cct. I think Critter is much stronger than the cct.

Best,
Gerold.

Critter stronger than CCT? Not in a million years- trust me.


Best,

[geots]

Thought I would report this update before heading to sleep. Komodo CCT has increased his lead from 3 games to 4 games.




Alienware AURORA_R4
Intel i7 w/6 True Cores
Fritz 11 gui
1CPU/64bit
128MB hash
Bases=NONE
Ponder_Learning=OFF
Perfect 2012b.ctg w/12-move limit
40/19 Repeating (Benched to adapt to 40/40)
Match=50 games

Code: Select all

SP8-i7, 19'/40+19'/40+19'/40  5/21/2013  

Komodo CCT 64-bit    +5/-1/=8   64.29%   9.0/14                 
Strelka 5.6 x64      +1/-5/=8   35.71%   5.0/14

Off to sleep-

Updates later

[geots]

The only way you could get any more "official" than this would be to have each programmer running his own engine. And as I said, in the history of the WCCC, no engine that was crowned "world champion" and had that title plastered all over his next release (more likely than not a sick joke IMO) played 50 games in the whole tournament.

Even with that, I am doubling the number of games from 50 to 100! In all my testing (I checked)- over 90% of the time the engine leading after 50 games was also leading after 100 games.

But no matter- 100 games it is. Update soon.


Best,

[geots]

One game from the quarter mark- and it is becoming more interesting as we go. Strelka goes on a 2 game mini-run, and then Komodo CCT stops it with a victory of his own. Actually, it doesn't get much better than this.





Alienware AURORA_R4
Intel i7 w/6 True Cores
Fritz 11 gui
1CPU/64bit
128MB hash
Bases=NONE
Ponder_Learning=OFF
Perfect 2012b.ctg w/12-move limit
40/19 Repeating (Benched to adapt to 40/40)
Match=100 games

Code: Select all

SP8-i7, 19'/40+19'/40+19'/40  5/21/2013    

Komodo CCT 64-bit   +44    +6/-3/=15   56.25%   13.5/24                 
Strelka 5.6 x64     -44    +3/-6/=15   43.75%   10.5/24

Depending on the length of the games, there may be another update later tonight.


Enjoy,

[Jim Collins]

Hi George. Will you be posting the games at the end of the match?

[geots]

Hi George. Will you be posting the games at the end of the match?

Posting the PGNs I "think" would be ok. Right now I would say "yes". But if Yuri says no, with Strelka, then it would be out of my hands. But as for now, take my answer as a "yes".



Best,

george