SedatChess

[Sedat Canbaz]

Breaking News (16.01.2024):
Human like tour is over, congrats and thanks to all and here final results...
Be aware that I realized to run two separate tourneys, one which is close
And above than 65% in Eng Similarity, where the other tour is up to 65%...
In other words, Ethereal is won the tour, based on more original engines,
Where Cfish is managed to win the tour, based on derivative etc engines
And as you may see too, no double standards..no unbalanced issues..e.g
Every engine is under test...sure which suits to the current new rules and
For anyone missed, as players only small in size (up to 555 KB) allowed..
And played the Top engines, which are close to GM level plus 3000+ Elo
Btw, to be honest, current competition reminded Me the old good days..)
Sure time will tell, but I hope to run more frequent Engine matches under
More real, serious conditions..e.g gaining from older/previous data, I mean
Playing via BIG help (via Evalfiles ) is already not much exciting as before..
E.g many of NEW engines are NOT capable to play at 'Classical' system
Only via NNUE assistance...otherwise, no way at all...really as sad news...
It's like, without compass: you can't find your home's way from 100 meters)
But the good news is that we have too many engines, which are capable..
Ok..that's all for now, see you later friends and thanks for your interest )


65%+ Champion: Cfish 030820 - My Congratulations to Cfish, Stockfish Team!
Note: Exactly same Cfish vers (via this tittle) is already 8-times Champion!

Up to 65% Champion: Ethereal 14 - Congrats to Andrew Grant, Alayan & Laldon!
Note: Ethereal engine is as 1st time a Winner, nice! but better late than never..


FOR FINAL STANDINGS:
https://sites.google.com/site/computers ... on-nn-cs-7

GAMES:
https://mega.nz/file/b5J0UALb#glsv9v8L2 ... HW9SyrC3dY


Best,
Sedat

[Homayoun]

Hi Mr. Sedat
Many thanks for great tour. Xiphos was a great absent. Unfortunately the size of Xiphos engine is more than 555kb.Also Klasman Engine (strong stockfish derivative) was another great absent. Totally it was a an interesting and nostalgic tournament . nice work.
Best regards

[Sedat Canbaz]

Hi Mr. Sedat
Many thanks for great tour. Xiphos was a great absent. Unfortunately the size of Xiphos engine is more than 555kb.Also Klasman Engine (strong stockfish derivative) was another great absent. Totally it was a an interesting and nostalgic tournament . nice work.
Best regards

Hello Mr. Homayoun

You are welcome )
And very interesting what you wrote...

Ok..here is my view.. and let's give a start... )

I see... but rule is rule, used up to 555 KB (uncompressed)
Not sure about next planning championship, but there is a
Possibility the size limit to be about up to 2 MB, and
Just in case, Xiphos plus many other strong ones will
Be as players too...e.g one of them will be Houdini 6.03
It is true also that by some people, Houdini is accused to
Be a SF clone..but my similarity tests indicate opposite!

Btw, why I limited the engine size up to 555 kb...e.g
Some of the chess engines, which are released
With huge sizes are out of interest.. as main reason:
Not sure what they include inside...in short:
My target is simply just pure real chess..played on
Own eng ideas (during active games, like Humans)
And as other good news is that,
Under these new cond. Ethereal is managed to be Champ too!

What I can say more,
Let's say we test Engines with Evalfiles,
Just one example, Stockfish includes also
NN data, which is generated/based on LCZERO (Leela)
Like a cocktail...everything is there...plus full
With draws...Elo is not as most important, at least
For my side...I'd prefer weaker, even old-dated engines...
But in same time...definitely we will see more fun !!
Meanwhile,
Did you notice the draw-ratio stats? close to 20%-25%
Really very low draw numbers..like GMs draw ratio ...

A few words about undefeated Champion (Cfish 030820),
Yes... Cfish can be as nostalgia engine, but still wins..
And I think that this is as most important, right ?)
As we know... not always.. newer are better...at least
Many Classical (without Evalfiles) tours indicate this...
I mean not only this tour...e.g in many previous tours,
Cfish's opponents were much much newer versions..but..

About the mentioned Klasman chess engine,
I think that it is based/derivative of asmFish...
Note also that Klasman is not available on my PCs,
But if you have it..just share it please, so later
I may try test it in the next new tours..


Greetings )

[Homayoun]

I agree with you completely. Nowadays chess world suffers from lack of original and innovative thoughts in chess. Both in engine games and either in human ( GM ) games. Although the condition of GM games is a little better. Chess fans miss for pure chess. The recent tournament of yours was a good example of pure chess and original thoughts of engines as the low draw rates showed. Thank you very much. Another important advantage of tournament was the size of the engines. It’s is very important that an engine maker could create a small size but very strong engine at the same time. Even nowadays in NNUE era I always appreciate the super strong engines with small and sophisticated nets. Like we see in Obsidian project. I appreciated the author of engine. Now with the help of our dear friend OrgZ we have Klasman engine in our archives again. I really like and support this kind of tournaments about pure and innovative chess.
Hope we have the chance to follow these kinds of tournaments here. Again ,I want to thank for your recent tournament.
Best regards

[Sedat Canbaz]

I agree with you completely. Nowadays chess world suffers from lack of original and innovative thoughts in chess. Both in engine games and either in human ( GM ) games. Although the condition of GM games is a little better. Chess fans miss for pure chess. The recent tournament of yours was a good example of pure chess and original thoughts of engines as the low draw rates showed. Thank you very much. Another important advantage of tournament was the size of the engines. It’s is very important that an engine maker could create a small size but very strong engine at the same time. Even nowadays in NNUE era I always appreciate the super strong engines with small and sophisticated nets. Like we see in Obsidian project. I appreciated the author of engine. Now with the help of our dear friend OrgZ we have Klasman engine in our archives again. I really like and support this kind of tournaments about pure and innovative chess.
Hope we have the chance to follow these kinds of tournaments here. Again ,I want to thank for your recent tournament.
Best regards

Yes, I totally agree with you too...

And let Me say please some notes more )

As we know, all NNUE is started/inspired after 2017 year..
For example, after Google's Alpha Zero vs Stockfish 8 match,
Where Alpha Zero is managed to win with a BIG margin..

Meanwhile,
Alpha Zero team is introduced the great NN's power to community...
But later... they are switched to others Google's non-chess projects...
Sure it would be much better to stay in chess...at least we could see
More powerful games by Google's Deep Mind company...

On the other hand,
This is other question of course, but day by day it's clear that
The played matches (by Google) was not fair..e.g Alpha Zero was
Played with trained NN data, where Stockfish played as without...
So nothing strange here that Stockfish is lost...but I admit also that
NNUE technology and architecture is very effective, impressive!!
Sure I mean if the target is to win...but vs weaker opponents!
This is as fact..reality is reality... but just in case of test 3700+
I mean close in strength...then almost all games are ending 'draws'
And these stats are with FAST TC, with SLOW TC I can't imagine.. )

Btw, another interesting point is that, as we know too,
As 1st available, plus free and super strong NN engine is appeared:
LCZERO (Lc0), but for maximum strength, it requires a fast GPU card
Yes...Lc0 is free and strong.. but fast GPU cards are not so cheap ))
And as usual I say over, Leela is like a caprice lady ))

Where later, I think in summer of 2020 year:
Stockfish is became as NNUE engine too...

One thing more
Nowadays, there are so many strong NNUE engines,
Where I am not going to count all.. because
I may miss to count some...but the sad news is that
Many are NOT capable to play serious (Classic) chess..
Only via previous played data...anything more else ?))
Or some are capable, but their performances is poor..
I mean much below than expected....

Ok..that's all for now.. see you later )

Greetings

[Sedat Canbaz]

Hello Chess Friends,

I am very pleased to announce that,
Another new interesting CS is coming very soon!

Some Notes,
Current tourney will be organized in the name of Mr. Homayoun
Definitely he deserves more, but for right now that's all what I can do..

Well, as planning eng players,
I hope to test again the small Top engines in size (about up to 2 MB)
This time I pick (for more fair cond.) which are close in release dates!
E.g all engines are belonging to 2017, mainly Sept, Oct, Nov, Dec 2017
Exception, only Xiphos 0.6 is much newer (released on November 2019)
As reason...that's as small request/interest by Mr. Homayoun

On the other hand,
And you may know too, and from my experience I can say:
It's a soft touch to release a 2024 engine and in same time
To be stronger than older ones, such as released 7 years ago...

But this is also true that,
Not so easy as it looks to create World's strongest engine...
Sure, I mean about which are very close in release dates!
As we know, definitely newer, better ideas are required !!
And once more, no one can be on Top by the work of others!
By the way, as same words are going for opening books too!

Note also that (about the current planning participants),
Mostly current engines played each other in my older tour/machines:
https://sites.google.com/site/computers ... m1s-rating
Note: I used Bullet TC in 2017, now it's planning to be used Blitz TC
In short, I wonder what will be influences under better hardware cond.
Because 2x EPYC is much faster (in chess) than 2 E5-2686 / i7 980X
In other words, If nothing else, it's expecting better/stronger games..
Plus I wonder too, what will be the overall draw percentages..

Besides, who can predict about which engine will be Number One ?)

Code: Select all

Andscacs 0921
Aristides v2d
asmFish 191117
asmFish 041017 WD
BrainFish 230917
CFish 090917
CiChess 220917
CorChess 180917
Ethereal 8.61
Fire 6.1
Fizbo 2.0
Ginkgo 1.96
Houdini 6.03 
Klasman Father
Komodo 11.2.2 
McBrain v2.7
Raubfisch ME262
Stockfish 051217
Toltec Sambala
Xiphos 0.6

Be aware of that too (as in recent organized tour -555kb):
The new championship will consist two separate rankings:
Derivative Eng Ranking plus Original Eng Ranking as well!
Otherwise, it will be not so much fair... right ?)

And please stay tuned...more info very soon )

Greetings

[Homayoun]

Hi Mr. sedat
I can’t find appropriate words to say thank to you. Thanks for your kindness and great favor to me. You maid all my chess nostalgias being alive again. I will follow the tournament with pleasant, Enjoy and pride. It will be honour for me. All the engines are great competitors but let me have a small prediction this time. I never do this as usual but this time I want to have a prediction before the match. I think in original engine section , Houdini and ethereal have the most chance. If we consider Houdini an original engine. And in derivatives section Klasman and toltec Sambala have the more chance. Let’s see what would happen in reality and tournament condition.
Yours faithfully: Homayoun Parsanecou

[Sedat Canbaz]

Hi Mr. sedat
I can’t find appropriate words to say thank to you. Thanks for your kindness and great favor to me. You maid all my chess nostalgias being alive again. I will follow the tournament with pleasant, Enjoy and pride. It will be honour for me. All the engines are great competitors but let me have a small prediction this time. I never do this as usual but this time I want to have a prediction before the match. I think in original engine section , Houdini and ethereal have the most chance. If we consider Houdini an original engine. And in derivatives section Klasman and toltec Sambala have the more chance. Let’s see what would happen in reality and tournament condition.
Yours faithfully: Homayoun Parsanecou

Not at all dear Homayoun )

And if you like the current new tour idea,
Then how happy I am...

Btw, I hope you never mind that instead of 'Mr.' I say as 'dear' ?
Because in this way...I think is better for close chess friends )

Regarding the predictions..
You can be right, exc. Ethereal 8.61 is never tested before..
I mean, it can perform/ gain lower points than expected...
But.. after all, time will tell...right ?)

Btw, the derivative eng tour is already started....!!

Good luck to all and have fun

Greetings

[Sedat Canbaz]

1st UPDATE

Meanwhile, just to be more clear:
Exc. two Champion tittles, there will be GRAND Champion tittle as well
E.g in the Finals,
The Winner of Original Engines will be played Vs the Winner of Derivative...

Note also that,
All current matches are playing at Blitz (2m+1s), min 660 games (per player)
Where the GRAND final will be played at Rapid (10m+5s), min 100 games...

And here are latest results of Derivative Eng Standing:

Code: Select all

Rank Name                          Elo     +/-   Games   Score    Draw 
   1 asmFish 191117                 14      11     487   52.0%   86.2% 
   2 BrainFish 230917               11      10     492   51.5%   88.8% 
   3 SF-020917 IP                   10      11     487   51.4%   86.4% 
   4 CorChess 180917                 7      11     486   51.0%   88.1% 
   5 asmFish WD                      4      11     490   50.5%   88.0% 
   6 Klasman Father                  3      11     490   50.4%   86.1% 
   7 Toltec Sambala                 -1      10     490   49.9%   88.8% 
   8 Raubfisch ME262                -1      12     490   49.8%   84.9% 
   9 McBrain v2.7                   -4      12     489   49.4%   84.9% 
  10 CiChess 220917                 -9      12     487   48.8%   85.6% 
  11 CFish 090917                  -13      13     489   48.2%   82.0% 
  12 Aristides v2d                 -20      12     489   47.1%   84.5% 

2933 of 3960 games finished.

[Homayoun]

Many thanks, for first round ( match)blitz.