Fringe- Special Engine and Book Tests

[Fringe407]

Keep the book! I would give it to you for free too.

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

This line is funny, which was changed by the developers to VVLTC search tune:

SF dev "VVLTC search tune" changed:

Depth Search::Worker::reduction(bool i, Depth d, int mn, int delta) {
int reductionScale = reductions[d] * reductions[mn];
- return (reductionScale + 1091 - delta * 759 / rootDelta) / 1024 + (!i && reductionScale > 950);
+ return (reductionScale + 1107 - delta * 725 / rootDelta) / 1024 + (!i && reductionScale > 956);


And what does it say in HypnoS 240324 VVLTC search tune? So:

Depth reduction(bool i, Depth d, int mn, int delta, int rootDelta) {
int reductionScale = Reductions[d] * Reductions[mn];
return (reductionScale + 1346 - int(delta) * 896 / int(rootDelta)) / 1024
+ (!i && reductionScale > 880);

Absolute wrong (very old SF code, no VVLTC search tune).

My dear friend Eduard,

I have no doubt you'll give me your book. Beautiful people must always exist. Thank you again, my friend, for the good work you do on behalf of chess engines and books.

Best regards,
Fringe

[Solista]

INCOGNITO 8 is really good. I have carried out a test against HypnoS 240324. It should be noted that INCOGNITO 8 is a special engine for analysis and is clearly superior to HypnoS here.

Nevertheless, I did a bullet test, but not 1+1 but 2+1.

GUI Fritz 19
8 Threads/Engine (5000 kns)
Hash 256 MB/Engine
Ponder ON (important for online gaming)
Syzygy all 3456 men
Book EN-Select (with color swap)
Timecontrol 2m + 1s

Result:



All games:
https://pixeldrain.com/u/fFxC8qDC

[Fringe407]

My dear friend Eduard,

I'm doing two big tests and it's being played now. There are 3 engines belonging to you, in total 10 engines are competing.

The INC8 is very powerful and is the engine I use for my lichess bot. But I'm still a fan of Light 26.3. I wish you would make the NNUE embedded version in your next series.
It will take about 2 days for my test results. Played for 3 days. I will share the results.

Best regards,
Fringe

[Fringe407]

First test:
Test book: Balsa270423
Time:1+1
Threads:1
Hash:64

Code: Select all

Rank Name                          Elo     +/-   Games   Score    Draw 
   1 Light 26.3                     14      11     180   51.9%   95.0% 
   2 SF-PB-220324                   10      10     180   51.4%   96.1% 
   3 JigSaw v5.1                    10       8     180   51.4%   97.2% 
   4 stockfish_24032617              6       8     180   50.8%   97.2% 
   5 Big-SF 220324                   6      11     180   50.8%   95.0% 
   6 HypnoS 240324                   4       9     180   50.6%   96.7% 
   7 INCOGNITO 8                     2       7     180   50.3%   98.3% 
   8 TacticaL Gear 3.1               0       9     180   50.0%   96.7% 
   9 corchess220324                 -2      11     180   49.7%   95.0% 
  10 dragon-3.3                    -49      19     180   43.1%   85.0%
  

1. pgn: https://pixeldrain.com/u/9pvmYAEM


Second test:
Test book: UHO2024
Time:1+1
Threads:1
Hash:64

Code: Select all

Rank Name                          Elo     +/-   Games   Score    Draw 
   1 Big-SF 220324                  35      38     180   55.0%   45.6% 
   2 SF-PB-220324                   29      35     180   54.2%   52.8% 
   3 HypnoS 240324                  21      37     180   53.1%   47.2% 
   4 Light 26.3                     15      36     180   52.2%   51.1% 
   5 TacticaL Gear 3.1              15      35     180   52.2%   52.2% 
   6 stockfish_24032617             14      35     180   51.9%   51.7% 
   7 corchess220324                 14      35     180   51.9%   51.7% 
   8 INCOGNITO 8                    -2      37     180   49.7%   47.2% 
   9 JigSaw v5.1                    -6      38     180   49.2%   45.0% 
  10 dragon-3.3                   -143      34     180   30.6%   52.2%
  

2. pgn: https://pixeldrain.com/u/cE441CCJ


Dear friends

Here are two tournaments with the same engine but different books. With the UHO opening book, we ensured fewer draws, while the Balsa opening book was evaluated by offering higher draws.

Conclusion:

I think we need more games to evaluate the results of the tournament where no parameters changed. As my dear friend Sedat said, we cannot see a similarity in the results.

Best regards,
Fringe.

[Gannicus]

Your statements are incorrect, you lack the expertise to back them up
Otherwise the programmers in the stockfish community wouldn't waste time testing my code
https://tests.stockfishchess.org/tests
The HypnoS code is simply structured based on my needs for compatibility with the various vb.NET tools that I want to continue using
- so refactoring is missing
- all patches related to it are missing
- Patches intentionally omitted due to incompatibility are missing with the tools I use
so stop talking nonsense.

Like I said, it's a mix of old and new in the Hypnos 190324 version. Just two quick comparisons, the hypnos code here is 2 months older than the SF dev code.

SF dev:

for (size_t i = 1; i < reductions.size(); ++i)
reductions = int((19.80 + std::log(size_t(options["Threads"])) / 2) * std::log(i));

Hypnos dev

for (int i = 1; i < MAX_MOVES; ++i)
Reductions = int((20.37 + std::log(Threads.size()) / 2) * std::log(i));

---

SF dev.

Depth Search::Worker::reduction(bool i, Depth d, int mn, int delta) {
int reductionScale = reductions[d] * reductions[mn];
return (reductionScale + 1091 - delta * 759 / rootDelta) / 1024 + (!i && reductionScale > 950);

Hypnos dev.

Depth reduction(bool i, Depth d, int mn, int delta, int rootDelta) {
int reductionScale = Reductions[d] * Reductions[mn];
return (reductionScale + 1346 - int(delta) * 896 / int(rootDelta)) / 1024
+ (!i && reductionScale > 880);

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

My Solista Attack book is commercial, I sell it, it's not free. I have two versions, BIN and CTG. Which one do you use? The current version is 8.4.

[Solista]

It's simply not the same as SF dev "VVLTC search tune". If you take 33 of the 38 parameters and call that a "VVLTC search tune", then that is not 100% correct. Feel free to call it nonsense, but I can tell 33 from 38. It's about the name, not your engine. When I see "VVLTC search tune" then I believe that this is exactly 100% the case with SF dev. That's not the case, it's suggesting to me something that isn't the case. I would therefore like to point out the difference. You're welcome to go your own way, it's all good. But point out the difference to SF dev.

[Gannicus]

I was and am on my way
I don't care about your wrong opinions.
However, if you don't want to be bitten, don't pull the dog's tail.
You write in all the forums that everyone is against you... but it's you who instigates people first and believe me this is quite annoying in my case.
Your best choice would be to ignore me like I do you.
Go and make war on those who attack you, I am not among them.

[Solista]

All I did was point out that your engine is not the same as Stockfish. That's all. I'll take the liberty of pointing this out even if you don't like it. I'm sorry, but I don't see how I should believe what you would like. Everyone is welcome to be against me, I don't mind. It is what it is, and your version is not identical to Stockfish - everyone likes to believe that - I don't.

[Fringe407]

Your statements are incorrect, you lack the expertise to back them up
Otherwise the programmers in the stockfish community wouldn't waste time testing my code
https://tests.stockfishchess.org/tests
The HypnoS code is simply structured based on my needs for compatibility with the various vb.NET tools that I want to continue using
- so refactoring is missing
- all patches related to it are missing
- Patches intentionally omitted due to incompatibility are missing with the tools I use
so stop talking nonsense.

Like I said, it's a mix of old and new in the Hypnos 190324 version. Just two quick comparisons, the hypnos code here is 2 months older than the SF dev code.

SF dev:

for (size_t i = 1; i < reductions.size(); ++i)
reductions = int((19.80 + std::log(size_t(options["Threads"])) / 2) * std::log(i));

Hypnos dev

for (int i = 1; i < MAX_MOVES; ++i)
Reductions = int((20.37 + std::log(Threads.size()) / 2) * std::log(i));

---

SF dev.

Depth Search::Worker::reduction(bool i, Depth d, int mn, int delta) {
int reductionScale = reductions[d] * reductions[mn];
return (reductionScale + 1091 - delta * 759 / rootDelta) / 1024 + (!i && reductionScale > 950);

Hypnos dev.

Depth reduction(bool i, Depth d, int mn, int delta, int rootDelta) {
int reductionScale = Reductions[d] * Reductions[mn];
return (reductionScale + 1346 - int(delta) * 896 / int(rootDelta)) / 1024
+ (!i && reductionScale > 880);

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

My Solista Attack book is commercial, I sell it, it's not free. I have two versions, BIN and CTG. Which one do you use? The current version is 8.4.

This is not a language of discussion, not a language of criticism. On the contrary, you're the one who ruined a good conversation. If something is wrong or needs to be corrected, you should state it in proper language. You are attacking the subject like a provocateur. We would be happy if you change your spelling and express what you know, my word to you Gannicus.

[Gannicus]

I finally agree with what you say
My engine is absolutely not identical to the stockfish
I don't want it to be and it never will be.