SedatChess

[Sedat Canbaz]

Meanwhile, just I'd like to add,
- Up to about 2 MB (uncompressed) in size are allowed...
- And as usual, all games will be shared after end of tour

Conditions:
2x Epyc 7B12, CuteChess, 1 Core, Ponder OFF, 3m+2s, 128 Hash, 4-MEN

Best,
Sedat

[Sedat Canbaz]

UPDATE

SCCT (3m+2s) Classic Championship's Latest Standings:

The Champion: Cfish - Congratulations to Cfish, SF team!
Via this title, Cfish engine is already 9-times as Champion!

Note also that,
The current CS is paused again...but planning to be resumed later...

Code: Select all

Rank Name               Elo    +    - games score oppo. draws 
   1 Cfish 030820      3597   23   22   740   89%  3255   21% 
   2 Stockfish 11      3558   21   20   740   86%  3256   26% 
   3 Houdini 6.03      3488   19   19   740   79%  3258   32% 
   4 Komodo 14.1       3479   19   18   740   78%  3258   38% 
   5 Ethereal 14       3464   19   18   740   76%  3259   35% 
   6 Berserk 5         3413   18   17   720   70%  3266   45% 
   7 SlowChess 2.3     3377   18   17   700   65%  3272   48% 
   8 Stash v35.0       3364   17   17   740   64%  3262   45% 
   9 Booot 6.5         3360   17   17   740   63%  3266   47% 
  10 Xiphos 0.6        3346   17   17   740   62%  3262   49% 
  11 RubiChess 180923  3335   17   17   720   59%  3267   42% 
  12 rofChade 2.3      3333   17   17   740   60%  3262   44% 
  13 Weiss 2           3331   17   17   740   60%  3262   50% 
  14 Fritz 18          3302   16   16   920   63%  3200   41% 
  15 DefenChess 2.2    3300   18   18   720   61%  3209   43% 
  16 Schooner 2.2      3295   18   18   720   60%  3209   43% 
  17 Fizbo 2           3291   19   19   720   60%  3209   32% 
  18 Andscacs 0.95     3271   18   18   720   57%  3209   38% 
  19 Willow 3.0        3237   19   18   720   55%  3192   33% 
  20 Wasp 4.50         3222   18   18   720   51%  3211   40% 
  21 Gull 150216       3200   18   18   720   47%  3219   36% 
  22 Arasan 24.1       3197   18   18   720   48%  3211   40% 
  23 Igel 2.6.0        3196   18   18   720   50%  3193   40% 
  24 Altair 5.0.0      3190   18   18   720   47%  3211   36% 
  25 Demolito 041021   3185   18   18   720   46%  3212   36% 
  26 ChessBrain 4.03   3177   18   18   720   45%  3212   34% 
  27 Ice 4.0.853       3165   18   18   720   46%  3194   38% 
  28 Nemorino 5.28     3161   18   18   720   46%  3194   35% 
  29 Vajolet2 2.8      3160   18   18   720   46%  3194   34% 
  30 Minic 2.5.3       3153   18   18   720   45%  3194   35% 
  31 Pedone 1.9        3150   18   18   720   45%  3194   33% 
  32 Hannibal 1.7      3147   18   19   720   44%  3195   33% 
  33 Protector 1.9.0   3132   18   18   740   43%  3192   37% 
  34 Texel 1.08        3126   18   18   740   42%  3192   35% 
  35 Senpai 2.0        3123   18   18   740   42%  3192   37% 
  36 Caissa 1.8        3086   18   18   760   37%  3192   30% 
  37 HakkaTCEC v2      3073   19   19   740   36%  3194   30% 
  38 Polaris 1.8.1     3057   19   19   720   37%  3171   33% 
  39 Midnight v6       3035   17   18   720   39%  3117   41% 
  40 Marvin 5.2.0      3030   18   18   720   38%  3117   39% 
  41 Tucano 9          3027   18   18   720   38%  3117   37% 
  42 Sjeng ct 2010     2982   18   18   720   32%  3118   30% 
  43 Naum 4.1          2977   18   19   700   31%  3120   34% 
  44 Bobcat v8.0       2962   19   19   700   29%  3120   32% 
  45 Spark 1.0         2958   19   19   700   29%  3121   32% 
  46 ZappaMexico II    2933   19   19   700   26%  3121   31% 
  47 Gaviota 1.0       2926   19   19   700   25%  3122   30% 
  48 Minko 1.3         2892   20   21   700   22%  3123   23% 
  49 Tornado 8         2860   21   21   700   19%  3123   20% 

[Homayoun]

Dear Sedat
Thank you very much.
Greetings

[Sedat Canbaz]

Dear Sedat
Thank you very much.
Greetings

Dear Homayoun,

Not at all..it is my pleasure

Greetings

[Sedat Canbaz]

Hello Friends,

Here is just a Blitz (2m+1s) test... played today...

Sorry for the games that not uploaded...I missed...
But in my next updates...I will try to share...

To be more clear,
Tetra x1 (by Tanick Ramz) is a new engine...which is
Using a tiny Evalfile... and its tuned, optimized for
Producing lower draw values...

Meanwhile, after checking.. current draws are little bit high: 92%
But for the test, Balsa openings are used... you may know...
Via strong openings: no much chances to see lower draw values...
But this is also true that via other Top NN engines,
We may see close to 98% or 99% draws...

In other words:
Another candidate NN engine for book tours.... nice..)
And many thanks to Mr. Tanick...

However, there are some important issues in Tetra x1 engine,
Which still needs to be improved...such as CTG book option is missing,
Plus no way to be played at Classic conditions (without NN assistance),
As I stated before, I am BIG fun playing chess at Classic cond. as well!!
As other important issue is that, Tetra is released with embedded net

E.g for running book tours...I prefer engines as without embedded net
Otherwise, too complicated...because for each book player I install other
Copy of engine...

Anyhow...here are the current played results,
About 5 Elo difference (in favor for Raubfish)

Code: Select all

Score of Tetra x1 vs Raubfish SC: 16 - 22 - 462 [0.494]
...      Tetra x1 playing White: 12 - 4 - 234  [0.516] 250
...      Tetra x1 playing Black: 4 - 18 - 228  [0.472] 250
...      White vs Black: 30 - 8 - 462  [0.522] 500
Elo difference: -4.2 +/- 8.4, LOS: 16.5 %, DrawRatio: 92.4 %
SPRT: llr 0 (0.0%), lbound -inf, ubound inf
500 of 500 games finished.

Conditions:
2x Epyc 7B12, CuteChess, 1 Core, Ponder OFF, Balsa, 2m+1s, 128 Hash, 4-MEN

Greetings )

[Homayoun]

Dear Sedat
Hi
First, I want to say my congratulations and appreciations to our dear friend Tanick Ramz for his decision to make strong engine (even derivatives) with tiny size nets. For me it’s very important if an author can make a super strong engine with very small size and sophisticated nets. As example, when I see Obsidian with 3 Mb size is stronger than Seer engine with 70Mb size, I enjoy. Now we don’t have enough information about this new engine, but the decision, itself is very valuable. I guess the future will belonged to the small size and sophisticated engines. Also thank you very much for your test sessions,which always includes new, interesting and innovative ideas.
Best regards

[Sedat Canbaz]

Dear Sedat
Hi
First, I want to say my congratulations and appreciations to our dear friend Tanick Ramz for his decision to make strong engine (even derivatives) with tiny size nets. For me it’s very important if an author can make a super strong engine with very small size and sophisticated nets. As example, when I see Obsidian with 3 Mb size is stronger than Seer engine with 70Mb size, I enjoy. Now we don’t have enough information about this new engine, but the decision, itself is very valuable. I guess the future will belonged to the small size and sophisticated engines. Also thank you very much for your test sessions,which always includes new, interesting and innovative ideas.
Best regards

Dear Homayoun,

Totally agreed with you...
And many thanks for your useful, interesting comments...

Yes...Mr. Tanick Ramz is doing really a good job...
E.g often he is trying to do something new...

Meanwhile,
Let Me say please a few more words over these issues )

Why I pay attention about engines to be small in sizes
Actually if the target is to test 3-5 books... it is all right..
But what about if the target testing over 50 or 100 books ?
In other words,
I see no reason to setup the same Evalfile for each book player...
Simply.. it is not so much logical...right ?)

And a few more words about Classic (NON-NN) testings,
Chess is born with that and will LIVE with that...
Just killing it...will be a BIG miss for many...

On other hand,
Playing with previous data (as NNUE) can be said still as a new future...
As we know, since 2020, SF is started to play via NN file...since 2023:
SF is playing only via NN... and sure, via NN file, engines are stronger,
But not so much exciting...reason, gaining from previous data...

Meanwhile, believe Me
If we will follow the sick mentality, to produce better quality games...
Then all testers have to use deeper opening lines...such as 40-50 moves..
Why in our eng testings, we have to use up to 4 or 8 deep opening moves ?
Or as other example, minimum 6-men endgames should be used....
Plus our testings should be with huge learned exp files... etc and etc..

Ah...one thing more...
Dual or Triple etc. NN files should be in test...
You know...all ...for better quality games...

In other words,
After all (with so many huge assistance...)
What will be the point then ... ?
Really computer chess is in danger, e.g
As we know also that.. if close to 3750 Elo eng testings...
Sure under more optimal cond. then almost 100% draws..
What is that ? a chess ? yes chess...but too boring for sure...
And no doubt via playing/gaining much with previous data:
I can not say much original too !

And yes...and for this reason, I say over:
Classic is Classic..just pure original chess !

Ok.. that's all for now )

Greetings )

[Sedat Canbaz]

Meanwhile,
I've realized to run some more Raubfisch and Tetra testings

And what is new:
In short, simply to see draw influences, regarding playing itself...

For Latest Standings, Games, Details:
https://sites.google.com/site/computers ... b-vs-tetra

Best,
Sedat

[OrgZ]

Results are close to what i got you see Raubf vs Neon Tetra
Greetings

[Sedat Canbaz]

Results are close to what i got you see Raubf vs Neon Tetra
Greetings

Yes... good news that both engines are very close in strength

And I hope to see a newer Tetra version, which will produce
Less draws... e.g Raubfish SC produced about 5% less draws..

Greetings