SF has now +33 ELO patch

[Jouni]

Code: Select all

LLR: 2.94 (-2.94,2.94) <0.50,3.50>
Total: 2048 W: 723 L: 525 D: 800
Ptnml(0-2): 10, 158, 511, 314, 31 

(Time-management fix in MultiPV mode.)

Sadly I don't see ANY point to play with multipv mode .

[Uri Blass]

I was surprised to see the small number of draws and found that the elo change is wrong even in multi-pv rating that I do not care about.

It seems that at some point stockfish changed the book to get less draws for some reason.

They used to have more than 90% draws with noob_3moves.epd and here is a typical result
https://tests.stockfishchess.org/tests/ ... 1525c9cb9b

suddenly in 21.8.2021 they started to test with a new book and had only something that is close to 50% draws

https://tests.stockfishchess.org/tests/ ... 1525c9cb19

[Cornfed]

Code: Select all

LLR: 2.94 (-2.94,2.94) <0.50,3.50>
Total: 2048 W: 723 L: 525 D: 800
Ptnml(0-2): 10, 158, 511, 314, 31 

(Time-management fix in MultiPV mode.)

Sadly I don't see ANY point to play with multipv mode .

Then...it is likely you just prefer to spend your time pushing engines thru their paces (playing against other engines or hiding behind them as others do while the engine mostly plays a game) as opposed to using them for analysis.

[Uri Blass]

Code: Select all

LLR: 2.94 (-2.94,2.94) <0.50,3.50>
Total: 2048 W: 723 L: 525 D: 800
Ptnml(0-2): 10, 158, 511, 314, 31 

(Time-management fix in MultiPV mode.)

Sadly I don't see ANY point to play with multipv mode .

Then...it is likely you just prefer to spend your time pushing engines thru their paces (playing against other engines or hiding behind them as others do while the engine mostly plays a game) as opposed to using them for analysis.

games do not test quality of multi-pv otherwise you can do normal search with one line in multi-pv and display random moves and random scores for the other moves.

[daniel71]

I thought this was a typo at first +33 patch, I'm very surprised because I thought all future gains would be small, very outstanding job by the Stockfish team. Thank you for all your fantastic work!

[notyetagm]

I thought this was a typo at first +33 patch, I'm very surprised because I thought all future gains would be small, very outstanding job by the Stockfish team. Thank you for all your fantastic work!

Damn, this is really a +33 increase in elo from this one patch?

Can someone please explain how does it do that? In simple English, since I am not a computer chess expert.

[Cornfed]

I thought this was a typo at first +33 patch, I'm very surprised because I thought all future gains would be small, very outstanding job by the Stockfish team. Thank you for all your fantastic work!

Damn, this is really a +33 increase in elo from this one patch?

Can someone please explain how does it do that? In simple English, since I am not a computer chess expert.

Read the whole thing ONLY revolves around Multi-PV - NOT regular play do don't get too excited.
From the development site:


Author: J. Oster
Date: Wed Oct 6 11:53:33 2021 +0200
Timestamp: 1633514013

Time-management fix in MultiPV mode.

When playing games in MultiPV mode we must take care to only track the
best move changing for the first PV line. Otherwise, SF will spend most
of its time for the initial moves after the book exit.

This has been observed and reported on Discord, but can also be seen in
games played in Stefan Pohl's MultiPV experiment.

Tested with MultiPV=4.

STC:
https://tests.stockfishchess.org/tests/ ... 38a969b990
LLR: 2.95 (-2.94,2.94) <-0.50,2.50>
Total: 1744 W: 694 L: 447 D: 603 Elo +49.54
Ptnml(0-2): 32, 125, 358, 278, 79

LTC:
https://tests.stockfishchess.org/tests/ ... 38a969b993
LLR: 2.94 (-2.94,2.94) <0.50,3.50>
Total: 2048 W: 723 L: 525 D: 800 Elo +33.70
Ptnml(0-2): 10, 158, 511, 314, 31

closes https://github.com/official-stockfish/S ... h/pull/372