SedatChess

[Sedat Canbaz]

Hello Chess Friends,

Meanwhile, I'd like to inform all of you also that
SCCT NN CS conditions are going to be changed!

In short, next future tours are planing to be played
Into 'SIZE'' CS Leagues, since 'NN size' matters a lot..!
Sure not always... but if I'm going to check in overall,
I admit that the well-tuned and larger Evalfiles play as
Significant role, with clear advantage over smaller ones!

Actually this is nothing new in SCCT..depending on tours,
Size rules used since a long time... besides for these
Reasons, the well-known sports Boxing, Wrestling etc.
Are organizing under 'Weight' categories.. right ?

In other words,
It's time for more frequent Fair Fight Battles !!
Btw, as other good example is recent organized LIV,
Where all Top engines used up to 25 MB Evalfiles!
But now, via the new size rules, more entries will
Participate too..otherwise there will be BIG miss!

And to be more clear, the planning new/future
SCCT NN tours will consist 3 (three) Leagues:

Small = Up to 25 MB
Medium = 26 - 60 MB
Large = 61 MB +

Note also that, soon as possible:
Medium and Large championships will be organized as well!

Small CS is already organized, so it's not required... at
Least for nowadays but later..in case of newer versions
Why not again...?!

And as final words, let Me say also this saying over, when
FAST, FAIR FIGHT, FUN, STABILITY, STRENGTH, TIME on scene,
The rest factors, influences etc. do not matter so much!
Of course I am talking for my side...not sure about you!?

FAST = Powerful, modern fast hardware
FAIR FIGHT = Playing in SIZE categories
FUN = Fast Bullet + Unique opening suite
STABILITY = No time loses, crashes etc.
STRENGTH = Mainly 3500+ Elo NN engines
TIME = Organizing frequent/many tours...


Best,
Sedat

[Sedat Canbaz]

UPDATE

Finely, I've managed to start the planning Medium CS (26 MB - 60 MB)
But I admit too SIZE CS are one of hardest tours that I ever organized...
Sure I'm referring for my side... at least I can say: a lot of details...
But just for Computer Chess's progress and it's ok...no any regret !)
If nothing else...now you have some info...that it's not so easy...
Actually if these tours were easy... everyone would run it, right ?

Current Leading Engine: SF Polyglot 15.1 - Congrats to Massimiliano Goi!

Code: Select all

Rank Name                          Elo     +/-   Games   Score    Draw 
   1 SF Polyglot 15.1               96      42      78   63.5%   67.9% 
   2 Stockfish 220923               86      39      82   62.2%   70.7% 
   3 Genko 1.1                      77      39      83   60.8%   71.1% 
   4 XTD 010723                     72      36      83   60.2%   74.7% 
   5 LittleBeast 2                  70      39      80   60.0%   72.5% 
   6 BlueMarlin 15.7                68      35      83   59.6%   75.9% 
   7 Yuliirma 2.1                   64      37      82   59.1%   74.4% 
   8 Polyfish 070823                61      37      80   58.8%   75.0% 
   9 D. Blue 280623                 56      36      81   58.0%   76.5% 
  10 Raubfisch X48e                 52      35      81   57.4%   77.8% 
  11 Zeus 26.5                      51      36      82   57.3%   75.6% 
  12 Brainlearn 25.2                51      34      83   57.2%   78.3% 
  13 Fisherov M 1.2                 39      37      80   55.6%   76.3% 
  14 Honey v14.1.01                 34      41      81   54.9%   70.4% 
  15 CFishNN 260623                 17      45      81   52.5%   65.4% 
  16 K. Dragon 3.3                   0      39      79   50.0%   74.7% 
  17 Fat Fritz 2                   -60      45      82   41.5%   63.4% 
  18 Revenge 3.0                  -130      50      81   32.1%   54.3% 
  19 Igel 3.5.0                   -146      56      78   30.1%   47.4% 
  20 Arasan 24.2.2                -197      59      82   24.4%   41.5% 
  21 Booot 7.3                    -217      60      83   22.3%   39.8% 
  22 Black Marlin 9.0             -218      68      81   22.2%   32.1% 

893 of 11088 games finished.

--------------------------------------------------

Note: Same Evalfiles are not allowed in the official contest..
So another newer Candidate tour is organized, the target to
Determine the best/strongest via nn-c38c3d8d3920.nnue !

And Polyfish 070823 wins 'nn-c38c3d8d3920.nnue' tour,
Congratulations to Polyfish's author (Khalid Omar) !

Code: Select all

1   Polyfish 070823  273.5/528
2   SF-PB 030923     271.5/528
3   RapTora 3.0      270.5/528
4   SF-POLY 220723   268.5/528
5   Corchess 240823  265.0/528
6   ProteusSF Aureo  259.5/528
7   CatroP 310723    239.5/528

GAMES:
https://mega.nz/file/CpITTDJC#qqtlrX-Lc ... x4gVMpsxCo

Note also that
I ran another test where CFishNN 260623 performed
Better/stronger than 2 (two) other Cfish eng versions..

And more updates soon )

Greetings

[Sedat Canbaz]

UPDATE 2

Top standings changed..it seems hard to guess what will be next ranks..
Current Leading Engine: Stockfish 220923 - Congratulations to SF team!

Code: Select all

Rank Name                          Elo     +/-   Games   Score    Draw 
   1 Stockfish 220923               83      17     420   61.8%   72.1% 
   2 Yuliirma 2.1                   81      17     422   61.5%   70.9% 
   3 Brainlearn 25.2                81      16     419   61.5%   72.8% 
   4 Polyfish 070823                77      16     419   60.9%   73.5% 
   5 XTD 010723                     74      17     420   60.5%   72.4% 
   6 SF Polyglot 15.1               70      17     418   59.9%   71.5% 
   7 Zeus 26.5                      69      17     418   59.8%   72.2% 
   8 Genko 1.1                      68      17     419   59.7%   73.0% 
   9 BlueMarlin 15.7                65      17     421   59.3%   71.5% 
  10 D. Blue 280623                 65      17     419   59.2%   71.6% 
  11 LittleBeast 2                  59      15     417   58.4%   77.5% 
  12 Fisherov M 1.2                 45      16     418   56.5%   74.6% 
  13 Raubfisch X48e                 42      16     420   56.1%   75.0% 
  14 Honey v14.1.01                 41      19     418   55.9%   67.7% 
  15 K. Dragon 3.3                  14      18     419   52.0%   71.1% 
  16 CFishNN 260623                  4      20     419   50.6%   64.9% 
  17 Fat Fritz 2                   -63      21     419   41.1%   60.6% 
  18 Igel 3.5.0                   -150      24     418   29.7%   48.8% 
  19 Revenge 3.0                  -168      25     417   27.6%   44.1% 
  20 Black Marlin 9.0             -199      26     421   24.1%   40.1% 
  21 Arasan 24.2.2                -203      26     419   23.7%   40.3% 
  22 Booot 7.3                    -234      27     422   20.6%   37.0% 


4611 of 11088 games finished.

[Sedat Canbaz]

Breaking News (06.08.2024):
NNUE Size CS are still on... but this time, as players are 26 MB - 60 MB
Unfortunately in this league: not all of the Top engines are so stable..sad
For example, there are 31 games lost on time, mainly by Revenge 3 and
Black Marlin 9, sure for accurate standings: I replayed those lost games..
But anyhow, in case of interest..I've included them in download package..
On other hand, almost all rest engines are stable under these conditions!
You know.. no one's work is perfect, but however I suggest more serious
And various trainings, tests etc (before final releases) thanks in advance
While I was mentioning about the stability) I should not forget to mention
The strong performance by Stockfish, Polyfish, Brainlearn..nice! Sure
Many of current entries are old-dated, but quite new eng versions will be
In the next planing Large Size CS.. e.g will consist manly 2024 releases,
Which will be participated soon as possible as well..so please stay tuned )

NN Medium Size I's Champion: Stockfish - Congratulations to SF Team!

For FULL Standings:
https://sites.google.com/site/computers ... edium-cs-1

GAMES:
https://mega.nz/file/u4pDgARZ#HjV1x8jph ... HybWfcHs2A

Best,
Sedat

[Homayoun]

Thank and congratulations to the winner author team too.
Best regards

[Sedat Canbaz]

Thank and congratulations to the winner author team too.
Best regards

You are welcome and + 1

[Sedat Canbaz]

Helle there )

As usually, I'm very pleased to announce that,
NNUE Candidate (Large Size) CS I is organized!
Actually even official contest is almost over..

Btw, in case of interest to submit, please read carefully:
- Per tour: 1 (one) public engine by same author is allowed
- Engines with same kind of evalfiles are not allowed, exc.
In case, via Candidate CS: the qualified player will be in..
- Unknown SF-based evalfiles are not allowed..for reference,
Simply will be used: https://tests.stockfishchess.org/nns

And here are Candidate (Large Size) CS I's results:
Congratulations to all, who managed to qualify...!

nn-31337bea577c.nnue: Corchess 040824 is the qualified player!

And as we see,
The latest Corchess ver. fully deserves to be in...right ?

Code: Select all

1   Corchess 040824   126.0 - 124.0127.5 - 122.5**   253.5/500
2   Stockfish 030824  124.0 - 126.0126.0 - 124.0 **  250.0/500
3   SF-PB 230724      122.5 - 127.5124.0 - 126.0  ** 246.5/500

nn-ddcfb9224cdb.nnue: Kookaburra 3 is the qualified player!

Here I admit that Kookaburra's performance is unbelievable!!
Sure I wonder too what kind of 'secrets' include Kokoburra?

Note also that
I checked several times regarding 1 Core etc. used or not...!
And be sure all of entries are played under same conditions..

Code: Select all

1   Kookaburra 3     +39/-17/=444 52.20%  261.0/500
2   Killfish 280624  +17/-39/=444 47.80%  239.0/500

nn-baff1edbea57.nnue: ShashChess 35.3 is the qualified player!

Code: Select all

1   ShashChess 35.3  +8/-6/=186 50.50%  101.0/200
2   Brainlearn 28.1  +6/-8/=186 49.50%   99.0/200

CSTal vs Rebel: CSTal 2.0.0 is the qualified player!

Code: Select all

1   CSTal 2.0.0  +28/-10/=162 54.50%  109.0/200
2   Rebel 16.3   +10/-28/=162 45.50%   91.0/200

GAMES:
https://mega.nz/file/n0JFXQxa#3lVsOvMZ0 ... x2mfCBob7U


Best,
Sedat

[Sedat Canbaz]

Breaking News (07.08.2024):
Well, NN Large Size (61 MB+) CS organized too...as players are picked
As far as possible 3700+ Elo e.g Minic 3.41 is tested too, but out due to
It's much weaker than current players..yes, Kookaburra is performed as
Real 'BEAST'' under these conditions! Unstoppable at least for nowadays!
And exception that managed as Number One again..besides, eliminated
Other very strong opponent (Killfish) during the Candidates CS, moreover
Kookaburra 3 is not so so new release.and it seems powerful moves may
Appear by older engines too, 2nd place: Stockfish dev..nice, where 3rd:
Marauders.. not bad as well..On other hand, after checking current stats,
It's clear all stable.. no any game to be lost on time in short, just great..
Btw, you may notice or not Small Size CS's Draws are 55% Medium: 65%
Large Size CS: 77%...in same that means Unique suite is more Drawish
Via Larger + Stronger NN data.. but I'm satisfied..and I assume you too?
Ok friends that's all for now, many thanks and congrats to all Eng Authors!

NN Large Size I's Champion: Kookaburra 3 - My Congratulations to Dorzs!
Note: exactly same Kookaburra vers managed to be Champ in older tours!

For FINAL / FULL Standings:
https://sites.google.com/site/computers ... -size-cs-1

GAMES:
https://mega.nz/file/3pBDwBRB#wEdCaz__m ... hMDfBlIDRg

Best,
Sedat

[Sedat Canbaz]

Breaking News (15.08.2024):
Current CS is ended...many thanks and my congratulations to all book authors!
Via latest new 6 (six) entries, the standings changed..especially Tunafish 3.1 is
Managed to gain higher points than previous Tunafish 3.0 so it seems stronger!
Plus AdrianOlivia, Oran etc. performed not bad.. but to be honest I've expected
More from the rest entries, anyhow I'm grateful to all, for their efforts, interest..
And a few more words about the Winner ( OPTIMUS ), yes...under these cond.
No doubt that Mr. Roger's book seems to be one of best public ones !! sure as
Same words are going for Mr. Hurnavich (Saphire), Mr. Hamdullah (HCanFree)!
Yes the cond. were hard, for anyone missed: all entries played vs private books
Created by Me...I mean simply that no much chances to train, gain, block.!) and
I assume also that experienced book makers understand perfectly what I mean)
Ok friends that's all for now... I hope to see you in next planning tours as well )

The Public Book Champion: OPTIMUS - Congrats to Roger Constantinesco!

FOR FINAL STANDINGS:
https://sites.google.com/site/computers ... t-bn-cs-33

GAMES:
https://mega.nz/file/ClwmyYoS#ZRx_X01rU ... 4VFVtncf7c

Meanwhile, SF-CTG 240715 engine (by Mr. Tanick Ramz) is released:
https://pixeldrain.com/u/pXEugc24

Best,
Sedat

[Sedat Canbaz]

Hello Chess Friends,


As usually I am very pleased to announce that,
Here is another new Book / NN Championship!

1st of all, the current contest's target is simply that:
This time TREND books out because there is also other opening theory:
Short neutral openings, tuned + used for determining eng strength..and
Via this CS: to check well-known sources which prefer various openings!
And as usual, Perfect book's openings are as one of the strongest out..!
Why I say like that..? because Oliprog P23 book is tuned to prefer only
Perfect 2023 openings (based on recent data played in Oliprog tours),
Sure not only as best in this tour.. I refer from the past to the present..
In the end, a 'Perfect' name and this name should be deserved..right ?)
In same time once more we noticed Oliprog's produced data is not so
Bad in quality + strength, moreover it seems Mr. Adrian's machine is
Stable as a rock...otherwise I hardly doubt Oliprog P23 (Perfect 2023)
Would get these points but this is also true without strong openings:
No any way to be Number One, at least under these tour conditions,
For example, as clear proof is Oliprog others book,
Which is based on many others books (Perfect openings not included)
I mean simply that it's tuned to prefer other various books openings too..
Picture is perfectly clear from my side..I assume is clear for side too?!

And a few words about my latest new created SCCT opening book,
It's a micro one, almost 450 KB, based on Balsa + Unique openings,
I have older SCCT book, which seems to be stronger..but not played
As reason, my target to be not Chamoion anymore..once is enough..
In SCCT (since 2007) I never title my opening books as Champ! sure
Just out of curiosity I may test my Perfect 2024 (up to 8 moves) too,
Simply for fun... but anyhow, I have to create a newer book, which
Will be based and optimized to prefer others Top authors openings,
And I hope that I'll manage to create a stronger book, otherwise the
Ongoing XXXIV CS will remain as active and it will never end... !)

Note also that,
All my created books are very small, minis plus 1-2 MB and for
Each book creation, only wins are preferred.. (for less draw ratio)
And as databases, mainly their high-rated 2024 NN games used..
Where CCC, FCP, FRRL... not played, since their games are old,
But even if they were as players, I hardly doubt would be Champ,
Especially FCP + FGRL books are fighting for the last places..sad..

And for anyone missed too, once more I'd like to point out that,
The strong openings play as 'golden key' for best performance!
Where weak openings are 'catastrophe' for worst performance!
Sure fun is important but much 'fun' may down the quality, right?

Btw, what does it means a 'balanced' word?
Actually I use 'balanced/unbalanced' term too, but for car vehicles!
Yes, since NNUE: balanced word became popular in opening theory )
But it pronounced mainly by programmers, testers...but not by Me..
And who knows? maybe I'm missing something here? On other hand,
In reality, the situation is not so complicated as it looks, simply
The opening theory is divided into two groups: Weak and Strong!
If still not clear, Champion books prefer to play balanced openings?
No..no...they prefer to play the 'very' balanced openings... hehe ))

And meanwhile, if we place one's cards on the table:
My 6th sense says that, many guys have complexes to write the
Word 'critical' or 'weak' or 'bad' or 'strong' etc. openings.. really!
Besides, many are not aware, which line is strong or weak.. e.g
Some check openings strength via Eng outputs, e.g if close '0' then
They say balanced..strange ) forget everything, checking openings
With outputs is madness! In other words, I'd summarize all like that:
All that glitters is not gold! yes...proved, and guaranteed !)

Sure I can talk more...to give useful tips..but what will be changed ?
Again some may say: ridiculous for who use 'strong' openings...and
It seems in theirs eyes, all tours have to be with bad (weak) lines ))
Actually it's ridiculous who generate unusual mismatch stats numbers!
Yes...where theirs data is ending with wins almost all for Whites side!
And all these weak (bad, critical..) openings are useful? not of course..
At least for my side, they are simply waste of time.. no more no less..!
But anyhow, I don't feel any regret..just for computer chess progress!

And without to not mention this I can not as well,
Many books represent the latest strongest eng opening theory, e.g:
Oliprog P23, plus CEDR, SSDF, SCCT etc. are one of them for sure!
Plus theirs produced stats (draws + wins) are logical plus acceptable!
But unfortunately I can not say as same words for many others...e.g
Those weak ones prefer such lines like just born not to win, but to lose!
I mean in the weak books: mostly of the strong openings are missing...
In fact, chess is a ancient strategic game, playing by two color players,
And no any one is born to play great strategic moves...but it can be
Learned by hard work, practicing, trainings etc.. but definitely not only
Via Eng outputs...Believe Me..I'm in book making plus in serious tours
More than 25+ years...and I think that it's a long time and experience!
This is also true.. if not learning from the past, it's something like that:
You are a pawn that do not know it's a piece of chess game!

Be aware of that too (sure for more fair tour conditions again..),
Since current CS, plus future ones.. all Book Tourneys will be
Organized in 'SIZE' categories (as recent played Eng NN CS)

And I have good news as well:
Book (BIN + CTG) registrations will be allowed too...but up to 5 MB
Sure in uncompressed size, besides, another important note more,
Only those small books will be played out which prefer various lines...
In short, as players: at least 4 (four) main opening lines are required..

The Current Leader: Oliprog P23 - My Congratulations to Myself..!)
And urgently I wish to see other strong One but to be Number One...

FOR LATEST STANDINGS:
https://sites.google.com/site/computers ... k-nn-cs-34

Thanks for reading, interest.. and have a nice weekend )


Best,
Sedat