Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 3:42 am
Peter Berger wrote:
Welcome Rebel 16.2 Rebel ( with the Noomen book) introduced itself with a great game on my computers, which was basically a book victory though, in a line that was initially worked out by an otherwise pretty unknown German correspondence player in like 2001 ( kudos to Mr Winckelmann). I wonder whether there is a bug in the book utility though. Rebel had to play moves 3 and 9 on its own, that I am pretty sure should have been in Jeroen‘s book. [Event "Lang 120min+10sek"] [Site "Berlin"] [Date "2023.02.27"] [Round "?"] [White "Crafty 25.6"] [Black "Rebel-16.2"] [Result "0-1"] [ECO "B33"] [PlyCount "56"] [TimeControl "7200+10"]
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 8:18 am
mwyoung wrote:
Flawless victory by Rebel 16.2! CP loss = 0.00.
What is CP loss and why is that important?
I assume CP stands for centi pawn.
Peter Berger
Posts : 126 Join date : 2020-11-20
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 10:52 am
mwyoung wrote:
I show it as 2010, unless another ICCF player fell into the same line almost a decade later. And that is possible.
The relevant games were Satici-Winckelmann and Petraitis-Winckelmann, 2001 (I know this because Mr. Winckelmann’s analysis made the ECO, quite an honour for an amateur player ). Another great very positional game today (see below). This also means that is pretty obvious that the book utility won’t play any captures from the book – no biggy, but it is conceivable to lead to problems.
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 11:02 am
Admin wrote:
mwyoung wrote:
Flawless victory by Rebel 16.2! CP loss = 0.00.
What is CP loss and why is that important?
I assume CP stands for centi pawn.
Centipawn loss is just another analysis tool to show accuracy in the game play. Or more important in a serries of games.
It means in this case no detectable errors were found in the game play using retrograde analysis, and measured in centipawn loss per move average.
For reference the best human players GM Carlsen, GM Fischer, GM Kasparov measures around .12 CP playing at peak form.
On average a typical or just average Grandmaster player scores below .40 CP for the average Grandmaster player.
Meaning playing Stockfish GM Carlsen, GM Fischer, GM Kasparov would be down 1.2 pawns ever 10 moves playing Stockfish on average.
Using this measurement it very easy to see why no human player has a chance playing the best chess engines.
Admin and Brendan like this post
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 11:42 am
mwyoung wrote:
Mclane wrote:
mwyoung wrote:
Chris Whittington wrote:
Admin wrote:
Thanks Thorsten, can you mail me the pgn ?
Those results don’t make a lot of sense
Why?
Maybe he dislikes that some engines have problems with the time controls. IMO it is not my job judge why some engines, e.g. Stockfish 15 or 15.1 have time control problems no matter if you test in 40/3‘ or 40/10 or even 40/20.
An engine that makes zillions of NPS and goes into the 40-50 ply should be capable to play 40/10‘ games. If not it’s obviously a bug.
Rebel16.2 is doing a very good job and plays very strong. Thats what counts.
I agree.
There were a lot of forfeits, also for Rebel, and the total favored Rebel and so the outcome is not fully trustworthy. Maybe Thorsten should consider to drop Arena.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 11:45 am
mwyoung wrote:
Admin wrote:
mwyoung wrote:
Flawless victory by Rebel 16.2! CP loss = 0.00.
What is CP loss and why is that important?
I assume CP stands for centi pawn.
Centipawn loss is just another analysis tool to show accuracy in the game play. Or more important in a serries of games.
It means in this case no detectable errors were found in the game play using retrograde analysis, and measured in centipawn loss per move average.
For reference the best human players GM Carlsen, GM Fischer, GM Kasparov measures around .12 CP playing at peak form.
On average a typical or just average Grandmaster player scores below .40 CP for the average Grandmaster player.
Meaning playing Stockfish GM Carlsen, GM Fischer, GM Kasparov would be down 1.2 pawns ever 10 moves playing Stockfish on average.
Using this measurement it very easy to see why no human player has a chance playing the best chess engines.
Interesting tool.... is it available for download or must I buy Fat-Fritz?
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Rebel 16.2 release Tue Feb 28, 2023 11:54 am
You are now on the enemies list of the Stockfish fanboys. They do not like it when another chess engine challenges Stockfish. As that is a hard and fast rule for testers. As the programmer of Rebel 16.2. You will need to hire body guards.
You are now on the enemies list of the Stockfish fanboys. They do not like it when another chess engine challenges Stockfish. As that is a hard and fast rule for testers. As the programmer of Rebel 16.2. You will need to hire body guards.
If that's the case Chris will need more body guards
Maybe he dislikes that some engines have problems with the time controls. IMO it is not my job judge why some engines, e.g. Stockfish 15 or 15.1 have time control problems no matter if you test in 40/3‘ or 40/10 or even 40/20.
An engine that makes zillions of NPS and goes into the 40-50 ply should be capable to play 40/10‘ games. If not it’s obviously a bug.
Rebel16.2 is doing a very good job and plays very strong. Thats what counts.
I agree.
There were a lot of forfeits, also for Rebel, and the total favored Rebel and so the outcome is not fully trustworthy. Maybe Thorsten should consider to drop Arena.
There is nothing wrong with Arena. Older Stockfish versions do not have these time forfeits.
Mclane
Posts : 3009 Join date : 2020-11-17 Age : 57 Location : United States of Europe, Germany, Ruhr area
Subject: The religious sects in computerchess Tue Feb 28, 2023 1:13 pm
Admin wrote:
mwyoung wrote:
You are now on the enemies list of the Stockfish fanboys. They do not like it when another chess engine challenges Stockfish. As that is a hard and fast rule for testers. As the programmer of Rebel 16.2. You will need to hire body guards.
If that's the case Chris will need more body guards
They treat Stockfish as god like. I have no idea why they do it. There is a tendency for religious behaviour for technical things. Apple products comes to my mind. The fans of this brand do also behave like a religious sect. We have a couple of apple fans in the company i work. And almost all of them have to reload their mobile phone during the day because the battery is nearly empty. All the android colleagues, we never reload our batteries at work. We do it at home before or after work. When you refer to this weakness they react very upset or even deny the fact that this is a weakness of this brand. They simply deny the technical fact that the apple batteries are smaller then the batteries of the competitors.
And similar it is wit the „best“ chess engine.
In the moment you critisize apple products or stockfish or in older days it was chessbase, the religious fans attack you.
If you play fast games the game is too fast (from their point of view). If you play slow games you do not have “enough games” to make a point (for them). If you present a PGN noone answers. Because these guys do not play chess or do not replay chess games as it seems. They are mainly interested in RESULTS not in chess content.
They are used to show and discuss elo list tables instead of pgns.
Last edited by Mclane on Tue Feb 28, 2023 1:22 pm; edited 1 time in total
You are now on the enemies list of the Stockfish fanboys. They do not like it when another chess engine challenges Stockfish. As that is a hard and fast rule for testers. As the programmer of Rebel 16.2. You will need to hire body guards.
If that's the case Chris will need more body guards
They treat Stockfish as god like. I have no idea why they do it. There is a tendency for religious behaviour for technical things. Apple products comes to my mind. The fans of this brand do also behave like a religious sect. And similar it is wit the „best“ chess engine.
In the moment you critisize apple products or stockfish or in older days it was chessbase, the religious fans attack you.
If you play fast games the game i too fast. If you play slow games you do not have “enough games” to make a point. If you present a PGN noone answers. Because these guys do not play chess as it seems. They are mainly interested in RESULTS not in chess content.
They are used to show elo list tables instead of pgns.
Correct. The die hard Stockfish fanboys take their devotion very seriously.
In other words they are nuts!
Mclane likes this post
Chris Whittington
Posts : 1254 Join date : 2020-11-17 Location : France
Centipawn loss is just another analysis tool to show accuracy in the game play. Or more important in a serries of games.
It means in this case no detectable errors were found in the game play using retrograde analysis, and measured in centipawn loss per move average.
For reference the best human players GM Carlsen, GM Fischer, GM Kasparov measures around .12 CP playing at peak form.
On average a typical or just average Grandmaster player scores below .40 CP for the average Grandmaster player.
Meaning playing Stockfish GM Carlsen, GM Fischer, GM Kasparov would be down 1.2 pawns ever 10 moves playing Stockfish on average.
Using this measurement it very easy to see why no human player has a chance playing the best chess engines.
Interesting tool.... is it available for download or must I buy Fat-Fritz?
You will need to buy Chessbase 17.
Can't we write one ourselves?
Algorithm: Take PGN game. For each move, ask SF for best move+eval, second best move+eval etc. If PGN move = SF move, score cp-loss=0 else score cp-loss=bestmoveEval minus PGNmoveEval.
calculate the average cp-loss. that's it.
If you tell me how to ask SF for Nth move eval, I'll Python it in a few minutes.
Centipawn loss is just another analysis tool to show accuracy in the game play. Or more important in a serries of games.
It means in this case no detectable errors were found in the game play using retrograde analysis, and measured in centipawn loss per move average.
For reference the best human players GM Carlsen, GM Fischer, GM Kasparov measures around .12 CP playing at peak form.
On average a typical or just average Grandmaster player scores below .40 CP for the average Grandmaster player.
Meaning playing Stockfish GM Carlsen, GM Fischer, GM Kasparov would be down 1.2 pawns ever 10 moves playing Stockfish on average.
Using this measurement it very easy to see why no human player has a chance playing the best chess engines.
Interesting tool.... is it available for download or must I buy Fat-Fritz?
You will need to buy Chessbase 17.
Can't we write one ourselves?
Algorithm: Take PGN game. For each move, ask SF for best move+eval, second best move+eval etc. If PGN move = SF move, score cp-loss=0 else score cp-loss=bestmoveEval minus PGNmoveEval.
calculate the average cp-loss. that's it.
If you tell me how to ask SF for Nth move eval, I'll Python it in a few minutes.
Yes except Chessbase 17 weights their score by difficulty. And Chessbase 17 does not count opening moves or captures and other forced moves.
Mclane
Posts : 3009 Join date : 2020-11-17 Age : 57 Location : United States of Europe, Germany, Ruhr area
You are now on the enemies list of the Stockfish fanboys. They do not like it when another chess engine challenges Stockfish. As that is a hard and fast rule for testers. As the programmer of Rebel 16.2. You will need to hire body guards.
If that's the case Chris will need more body guards
They treat Stockfish as god like. I have no idea why they do it. There is a tendency for religious behaviour for technical things. Apple products comes to my mind. The fans of this brand do also behave like a religious sect. And similar it is wit the „best“ chess engine.
In the moment you critisize apple products or stockfish or in older days it was chessbase, the religious fans attack you.
If you play fast games the game i too fast. If you play slow games you do not have “enough games” to make a point. If you present a PGN noone answers. Because these guys do not play chess as it seems. They are mainly interested in RESULTS not in chess content.
They are used to show elo list tables instead of pgns.
Correct. The die hard Stockfish fanboys take their devotion very seriously.
In other words they are nuts!
I have posted 2 sacrificial games of CSTAL in PGN format and there is NO comment. Nothing.
Maybe he dislikes that some engines have problems with the time controls. IMO it is not my job judge why some engines, e.g. Stockfish 15 or 15.1 have time control problems no matter if you test in 40/3‘ or 40/10 or even 40/20.
An engine that makes zillions of NPS and goes into the 40-50 ply should be capable to play 40/10‘ games. If not it’s obviously a bug.
Rebel16.2 is doing a very good job and plays very strong. Thats what counts.
I agree.
There were a lot of forfeits, also for Rebel, and the total favored Rebel and so the outcome is not fully trustworthy. Maybe Thorsten should consider to drop Arena.
There is nothing wrong with Arena. Older Stockfish versions do not have these time forfeits.
Rebel had 3 forfeits, impossible.
Admin Admin
Posts : 2536 Join date : 2020-11-17 Location : Netherlands
Subject: Re: Rebel 16.2 release Sat Mar 11, 2023 2:05 pm
A short update. I am still playing these slow Rebel 16.2-Crafty games , when I have the time to let them play and follow the games live (which doesn’t happen too often because of real life). Current score is 8-0. Not only is Rebel completely destroying its opponent, the games are also very nice and impressive – they are looking way better than Stockfish’s to my personal taste. I am getting the impression that Rebel profits from the long time control, but this is guesswork of course as I am testing nothing else . But nothing really has to be posted, nothing strange. I can also give a feedback on the Noomen book now: I really like – high quality work. Much as expected the book utility bug doesn’t matter much, other than stealing a little time from Rebel’s bank, that doesn’t matter at all. Only thing is that a few lines Jeroen might well have chosen will never get played (e.g. QGA).
Admin, Mclane and matejst like this post
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: I added Leptir to my contest Tue Mar 21, 2023 4:27 am
567 of 900 games played Name of the tournament: Tdog Site/ Country: DESKTOP-426Q3JG, United States Level: Blitz 5/2 Hardware: AMD Ryzen Threadripper 3970X 32-Core Processor with 63.9 GB Memory Operating system: Windows 10 Home Home Edition (Build 9200) 64 bit PGN-File: C:\Program Files (x86)\Arena\Tournaments\Tdog.pgn
Dann Corbit
Posts : 189 Join date : 2020-11-26
Subject: Threading test, with Leptir added for kicks Tue Mar 21, 2023 2:45 pm