New commercial engine Revenge 1.0

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Frank Quisinsky
Posts: 5643
Joined: Wed Nov 18, 2009 6:16 pm
Location: Trier, Germany
Full name: Frank Quisinsky
Contact:

Re: New commercial engine Revenge 1.0

Post by Frank Quisinsky » Sun Jul 11, 2021 12:28 pm

Made no different to set Stockfish to 2250 Elo or to set Revenge to 2250 Elo.
The question is more ... which can play the most interesting chess (human-view) with 1600, 1800, 2000 or 2200 Elo.

OK, Stockfish have a great playing style also but the style from Pedone is differently to Stockfish.

The comparsion with Stockfish for each new commercial product made no sense for me.
Can be make sense if humans are clearly stronger as best chess engines today (older chess-computer times).
Today it make no sense, not important for new commercial products which Elo Stockfish can give. 3500, 4000 or 5000 Elo ... no matter.
Nothing to do with all that what for humans can be intersting!

Elo alone is in reality "Nothing" if computer chess engines Elo is clearly higher as Elo from best humans.
Max. interesting for analyzes.

But the style of an engine is pure gold!

Best
Frank
I like computer chess!

User avatar
CMCanavessi
Posts: 1078
Joined: Thu Dec 28, 2017 3:06 pm
Location: Argentina

Re: New commercial engine Revenge 1.0

Post by CMCanavessi » Sun Jul 11, 2021 2:45 pm

Will Pedone still be updated, or is 3.1 the final version ever?
Follow my tournament and some Leela gauntlets live at http://twitch.tv/ccls

Fabio Gobbato
Posts: 184
Joined: Fri Apr 11, 2014 8:45 am
Full name: Fabio Gobbato
Contact:

Re: New commercial engine Revenge 1.0

Post by Fabio Gobbato » Sun Jul 11, 2021 5:17 pm

Pedone is no more under development, I will work only on Revenge.

carldaman
Posts: 2129
Joined: Sat Jun 02, 2012 12:13 am

Re: New commercial engine Revenge 1.0

Post by carldaman » Mon Jul 12, 2021 3:33 am

Frank Quisinsky wrote:
Sun Jul 11, 2021 12:28 pm
Made no different to set Stockfish to 2250 Elo or to set Revenge to 2250 Elo.
The question is more ... which can play the most interesting chess (human-view) with 1600, 1800, 2000 or 2200 Elo.

OK, Stockfish have a great playing style also but the style from Pedone is differently to Stockfish.

The comparsion with Stockfish for each new commercial product made no sense for me.
Can be make sense if humans are clearly stronger as best chess engines today (older chess-computer times).
Today it make no sense, not important for new commercial products which Elo Stockfish can give. 3500, 4000 or 5000 Elo ... no matter.
Nothing to do with all that what for humans can be intersting!

Elo alone is in reality "Nothing" if computer chess engines Elo is clearly higher as Elo from best humans.
Max. interesting for analyzes.

But the style of an engine is pure gold!

Best
Frank
Playing style is more important than some people realize. For example, in my tests the latest SF and NNUE derivatives are at times unable to defeat opponents rated 600 Elo less (!!), when having the Black pieces AND playing without a book. This lack of aggression and risk-taking is actually very serious and costly, resulting in the engines playing not to lose, rather than for a win, with Black. It happens too frequently when considering the Elo difference.

I know the rating lists don't like to test the strongest engines against much weaker opponents, or from the starting position, but perhaps the underlying assumptions should be rethought and some allowances made for such tests. The inability to defeat much weaker opposition should definitely cost Elo! This is currently not measured, unfortunately, because of the absence of such testing.

This is why engines that always play for a win and don't mind the risk of losing should deserve more recognition, and Elo should be secondary, as long as sufficient strength is present. I would rather analyze with and spar against a more interesting and intelligent engine, even if it's 200 Elo or more below SF NNUE! :|

User avatar
pohl4711
Posts: 1690
Joined: Sat Sep 03, 2011 5:25 am
Location: Berlin, Germany
Full name: Stefan Pohl
Contact:

Re: New commercial engine Revenge 1.0

Post by pohl4711 » Wed Jul 14, 2021 5:50 am

Fabio Gobbato wrote:
Fri Jul 09, 2021 12:43 pm
I'm glad to announce the birth of a new commercial engine: Revenge 1.0
It's based on Pedone 3.1 but with many improvements that make it stronger of about 50 elo.
7000 games testrun of Revenge 1.0 finished:

https://www.sp-cc.de/

User avatar
Sylwy
Posts: 3781
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI - the historical capital of MOLDOVA
Full name: SilvianR

Re: New commercial engine Revenge 1.0

Post by Sylwy » Thu Jul 15, 2021 5:03 pm

A nice game of Revenge_1.0_x64_bmi2:


User avatar
Sylwy
Posts: 3781
Joined: Fri Apr 21, 2006 2:19 pm
Location: IASI - the historical capital of MOLDOVA
Full name: SilvianR

Re: New commercial engine Revenge 1.0

Post by Sylwy » Thu Jul 15, 2021 5:51 pm

My test after 7 games (from 100):

Image

Cornfed
Posts: 313
Joined: Sun Apr 26, 2020 9:40 pm
Full name: Brian D. Smith

Re: New commercial engine Revenge 1.0

Post by Cornfed » Thu Jul 15, 2021 6:58 pm

carldaman wrote:
Mon Jul 12, 2021 3:33 am
This lack of aggression and risk-taking is actually very serious and costly, resulting in the engines playing not to lose, rather than for a win, with Black. It happens too frequently when considering the Elo difference.

I know the rating lists don't like to test the strongest engines against much weaker opponents, or from the starting position, but perhaps the underlying assumptions should be rethought and some allowances made for such tests. The inability to defeat much weaker opposition should definitely cost Elo! This is currently not measured, unfortunately, because of the absence of such testing.

This is why engines that always play for a win and don't mind the risk of losing should deserve more recognition, and Elo should be secondary, as long as sufficient strength is present. I would rather analyze with and spar against a more interesting and intelligent engine, even if it's 200 Elo or more below SF NNUE! :|
Actually, doesn't that all depend on the engine - I mean to play more 'to win' almost by definition = playing to potentially lose.

carldaman
Posts: 2129
Joined: Sat Jun 02, 2012 12:13 am

Re: New commercial engine Revenge 1.0

Post by carldaman » Thu Jul 15, 2021 10:47 pm

Yeah, except that SF-NNUE has so much strength that it could afford to take a few more risks without losing Elo.
Unfortunately, it doesn't 'know' or 'understand' that.

Remember, when they added contempt to classical SF, it played more aggressively AND gained strength.

Post Reply