Latest Updates:
Hot Topic (More than 10 Replies) Theoretical lines where you beat the computer (Read 2360 times)
Confused_by_Theory
God Member
*****
Offline


I Love ChessPublishing!

Posts: 643
Location: Europe
Joined: 05/13/15
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #14 - 03/15/22 at 16:10:12
Post Tools
Hi.

TCEC has predetermined openings asfaik.

Regards.
/ CbT
  
Back to top
 
IP Logged
 
ErictheRed
God Member
*****
Offline


USCF National Master

Posts: 2533
Location: USA
Joined: 10/02/05
Re: Theoretical lines where you beat the computer
Reply #13 - 03/15/22 at 15:55:07
Post Tools
I beat Droidfish once on my phone about 5 years ago playing a quick blitz game.  It grabbed a bunch of material and the horizon effect set in, I suppose.  It's never happened again and I don't remember the score of that game, unfortunately.
  
Back to top
 
IP Logged
 
FMCharless
Full Member
***
Offline


I Love ChessPublishing!

Posts: 199
Location: Pompano Beach Florida
Joined: 10/29/21
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #12 - 03/14/22 at 19:10:16
Post Tools
no, because chess hasnt been solved, engine matches has what is called opening books which are customized for them. so their matches are all based on the opening books they have.
  
Back to top
 
IP Logged
 
Dink Heckler
God Member
*****
Offline


Love-Forty

Posts: 876
Joined: 02/01/07
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #11 - 03/13/22 at 10:17:40
Post Tools
Openings are pre-assigned for engine matches to avoid draw death, aren't they?
  

'Am I any good at tactics?'
'Computer says No!'
Back to top
 
IP Logged
 
cathexis
God Member
*****
Offline


No matter where you go,
there you are.

Posts: 591
Joined: 03/03/20
Re: Theoretical lines where you beat the computer
Reply #10 - 03/13/22 at 01:15:55
Post Tools
If 2 chess engines play each other it is not impossible for one side to score a win. But if we accept the Theory of Steinitz (which I know is debatable) then at least engine has made an error. Why would any engine ever make an error in the first place? Why isn't every game a draw? Perhaps there lies a strategy in the answer. Just saying.
  
Back to top
 
IP Logged
 
Konstriktor
Full Member
***
Offline


I Love ChessPublishing!

Posts: 171
Joined: 02/21/10
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #9 - 03/12/22 at 21:54:00
Post Tools
If chess engines are everything, as a wise man once said, how can you beat 'em?
  
Back to top
 
IP Logged
 
FreeRepublic
God Member
*****
Offline


I Love ChessPublishing!

Posts: 699
Location: Georgia
Joined: 06/08/17
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #8 - 03/12/22 at 16:28:50
Post Tools
ArKheiN wrote on 03/12/22 at 14:52:55:
It's already nearly impossible to beat a strong engine in corr play in human+machine vs machine and you think human alone can win by playing from start to end?


From "start to end," yes that is a great challenge.

I remember a poster someplace who said he was looking for opening lines mis-evaluated by computers. That seems to be a more realistic goal. It seems to me that Stockfish 14 is pretty negative on the French defense and the modern Benoni opening (MBO). Is that really warranted? I don't know.

I seem to recall a conversation between GMs de Firmian and  Yermolinsky. Some junior player dismissed the MBO as inadequate, and they had a good laugh. I wonder if the MBO was bread and butter for de Firmian.

How is one to judge? Well, you could trust the computer. There are lines where computers come around eventually to agreeing with humans. I take that as an indication that human judgement is later confirmed by computer calculations. Or, if you will, that the human was right and the computer wrong at an early stage.

The French and Benoni are different to be sure. But in each case white has more space (usually a pawn on the 5th rank) and black has counter-attacking potential. Does the computer dismiss play and activity prematurely perhaps?

I remember an introduction in an old opening book. Perhaps it was from Fine in an early edition of Modern Chess Openings. It was to the effect that:

with an opportunity to move a pawn forward two squares, and choosing to move it forward only one square, the French Defense and the Caro-Kan fail to equalize.

It is so simple that it seems naive. Yet simple is not necessarily simplistic.
  
Back to top
 
IP Logged
 
ArKheiN
God Member
*****
Offline


I love ChessPublishing.com!

Posts: 728
Location: Belgium
Joined: 03/30/05
Re: Theoretical lines where you beat the computer
Reply #7 - 03/12/22 at 15:09:07
Post Tools
An interesting match in slow game could be Carlsen vs Stockfish playing the ending line of the Riga Ruy Lopez where White is better but not winning. I am sure Stockfish would hold easily and why not win after a long fight but I believe Carlsen should not lose that in slow game as White against anybody. Carlsen as Black is another challenge!
  
Back to top
 
IP Logged
 
ArKheiN
God Member
*****
Offline


I love ChessPublishing.com!

Posts: 728
Location: Belgium
Joined: 03/30/05
Re: Theoretical lines where you beat the computer
Reply #6 - 03/12/22 at 14:52:55
Post Tools
It's already nearly impossible to beat a strong engine in corr play in human+machine vs machine and you think human alone can win by playing from start to end? You can if the theoretical line is very bad like "Jerome Gambit" or less bad like "Halloween gambit" and even there the human has to be strong not to slip the advantage. Even Nakamura lost with material odd vs stockfish, maybe he can with with a knight advantage...
  
Back to top
 
IP Logged
 
cathexis
God Member
*****
Offline


No matter where you go,
there you are.

Posts: 591
Joined: 03/03/20
Re: Theoretical lines where you beat the computer
Reply #5 - 03/12/22 at 13:32:50
Post Tools
It looks like Roman got pawn odds and white at all rounds, which I wouldn't call "full strength." Also, it seems from the write-up that, given he was always a pawn up, Roman's basic strategy was frequently to exchange down to a hopefully better endgame or draw. However, the write-up also says that Rybka made some errors. - at least that's what it sounds like to me. The hyper-links to the individual games appear broken, but in at least one case he played the Engiish. But I wasn't aware of this game so thx.

However, in fairness (and in criticism of my earlier remarks) it might be more accurate to ask if human and computer ELOs are even comparable and what about SF mounted on 64-core Ryzen with an SSD for tablebase look-up vs. on some old Pentium with an old HHD? So perhaps comparing SF with human play is like daring Usain Bolt to race your dad's Ford F-150, irrelevant comparison.
  
Back to top
 
IP Logged
 
FMCharless
Full Member
***
Offline


I Love ChessPublishing!

Posts: 199
Location: Pompano Beach Florida
Joined: 10/29/21
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #4 - 03/12/22 at 01:28:32
Post Tools
didnt roman dzindzichashvili play a match versus rybka?
https://chessok.com/?p=468
  
Back to top
 
IP Logged
 
cathexis
God Member
*****
Offline


No matter where you go,
there you are.

Posts: 591
Joined: 03/03/20
Re: Theoretical lines where you beat the computer
Reply #3 - 03/12/22 at 00:31:07
Post Tools
I think that Ukrainian GM (with a "P"?) was the last GM to beat a chess engine at full strength. That was Fritz 5 in 2005, IIRC.  I was just using Fritz 17 myself. Does anyone know how high a level Magnus can beat SF? Or any other related kind of rankings?
Might be an interesting idea for a simul; a solitary SF engine versus 4 or 5 Super-GMs. Who could beat it at what highest level compared to the others?

Just Sayin'
  
Back to top
 
IP Logged
 
FMCharless
Full Member
***
Offline


I Love ChessPublishing!

Posts: 199
Location: Pompano Beach Florida
Joined: 10/29/21
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #2 - 03/11/22 at 23:52:02
Post Tools
the reality is that engines keep the chess alive by working on their opening books. in other words the positions which stem out of the openings in their process of solving chess. different engines can give different assessments to different openings, and different solutions can be found for different systems. that is how engines keep the game alive.

the only opening book from computers which i have found is the rybka 4 opening book. I have it since i purchased it and its been wonderful. it goes into depth into many openings but leaves me wanting for more in certain lines.

having said that it is really deep and i find it fascinating. its creating books like those that could keep people alive for ages.

finding openings where the computer could lose would be fascinating. for instance, in the line 1.d4 d5 2.c4 e6 3.Nc3 c6 4.e4 dxe4 5.Nxe4 Bb4+ 7.Bd2 Qxd4 if black can hold on to the pawn in that line it would be a real deal. In other words a line where you could beat the computer. Anyone know of lines like this?
  
Back to top
 
IP Logged
 
TonyRo
God Member
*****
Offline


I'm gonna crack your skull!

Posts: 1826
Location: Cleveland, OH
Joined: 11/26/07
Gender: Male
Re: Theoretical lines where you beat the computer
Reply #1 - 03/11/22 at 23:41:15
Post Tools
It is fascinating to me that at this point you feel like you have a chance anywhere. Michael Adams couldn't beat Hydra in 2005. Is this real life?  Cheesy
  
Back to top
 
IP Logged
 
FMCharless
Full Member
***
Offline


I Love ChessPublishing!

Posts: 199
Location: Pompano Beach Florida
Joined: 10/29/21
Gender: Male
Theoretical lines where you beat the computer
03/11/22 at 22:46:08
Post Tools
does anyone know lines where you beat the computer?
like is the French really the opening to beat the computer?
I know in the Maccutcheon the engine likes black better in the main lines.

Do know how useful that could be but I am looking for someone that has discovered this.

beating the engine would be grandiose
  
Back to top
 
IP Logged
 
Bookmarks: del.icio.us Digg Facebook Google Google+ Linked in reddit StumbleUpon Twitter Yahoo