OpenChess

Homayoun wrote: Tue Apr 02, 2024 11:52 am

Homayoun wrote: Tue Apr 02, 2024 11:43 am

Cr@ck wrote: Tue Apr 02, 2024 11:24 am Friends, please tell me at what time controls STC and VLTC testing of Stockfish is carried out? It says that VLTC is carried out with control 60+0.6. Is it 60 minutes or 60 seconds per game?

What do you personally think?VLTC means , very long time control. Then it should be second or minute?
Regards

Ok. I think Ibay is right.with excuse I wrongly read 60+ 6. But if it is 60+ 0.6 , Then Ibay is right.
Regards

I think it’s 60 minutes, but I’m absolutely not sure about it, so I want to find out how it really is. STC is carried out with control 10+0.1 and I also can’t understand, is it 10 minutes per game or 10 seconds?

VLTC is 60 seconds + 0.6 seconds increment per move, STC is 10 seconds + 0.1 seconds increment per move. That are the time controls, not minutes. It does not make sense to test at 60 minutes + 0.6 seconds, that would take too long if you want to run a significant amount of games. You can anyways check it yourself here https://tests.stockfishchess.org

IbaiBuR wrote: Tue Apr 02, 2024 2:44 pm VLTC is 60 seconds + 0.6 seconds increment per move, STC is 10 seconds + 0.1 seconds increment per move. That are the time controls, not minutes. It does not make sense to test at 60 minutes + 0.6 seconds, that would take too long if you want to run a significant amount of games. You can anyways check it yourself here https://tests.stockfishchess.org

IbaiBuR wrote: Tue Apr 02, 2024 2:44 pm VLTC is 60 seconds + 0.6 seconds increment per move, STC is 10 seconds + 0.1 seconds increment per move. That are the time controls, not minutes. It does not make sense to test at 60 minutes + 0.6 seconds, that would take too long if you want to run a significant amount of games. You can anyways check it yourself here https://tests.stockfishchess.org

Thank you very much for the clarification

Tuesday April2, Author Viren6
https://abrok.eu/stockfish/?

Author: Disservin
Date: Sat Apr 13 22:05:19 2024 +0200
Timestamp: 1713038719

Fix wrong sign for 200 TB score

Fix another case of 9032c6cbe74ccf7e8963755501e7e6cc473ae471

* TB values can have a distance of 0, mainly when we are in a tb position but haven't found mate.
* Add a missing whitespace to UCIEngine::on_update_no_moves()

Closes https://github.com/official-stockfish/S ... /pull/5172

No functional change

https://abrok.eu/stockfish/?

April 2024: SFS mpv 130424 by Solista, engine based on the latest updates from Stockfish dev.
The influence of the smaller network (dimension 256) has been reduced, so the engine is not a bullet monster.

A quick test on 4 cores against Stockfish 020424 (Random Op MUltiPV = 1) yielded the following results: Bullet 1m+1s -> https://up.picr.de/47414279ba.png and 2m+1s -> https://up.picr.de/47414281sg.png.

By the way, I always test under server conditions, so with Ponder ON.
As you can see, an increase of 1min is enough to get past Stockfish dev (Bullet-King).

For gimmicks I therefore recommend Random Op.MultiPV=1. MultiPV but set to 3.

In my ENET-2023 test, SFS manages 106 of 110 solutions at max 60s per train.

Download Win 64-Bit (avx2, bmi2, sse41, source code): https://pixeldrain.com/u/doXDbsJh

Thank you Eduard !

Danke Eduard
Ein Gutes Engine.
Gute Arbeit
Matjaz

Thank you, Eduard !

April 14, 2024: Normal version of SFS 130424 without Random Op MultiPV mode, some prefer to analyze in normal MPV mode.
Download: https://pixeldrain.com/u/zyGYH1Bp.

SFS stands for Stockfish Solista, I have not only changed the nets and the threshold for the small network. The search has also been modified and is slightly slower than in Stockfish dev, but slightly faster than Incognito. SFS 130424 is excellent for analyses with 2 variants (under Fritz GUI -> https://up.picr.de/47423219zw.png). For analyses with tactical motives, SFS mpv 130424 with Random Op.MultiPV = 3 (preset) seems a little better. An FS player told me that he uses Random Op.

Thank you Eduard !