Forums

A chess problem that computers can't solve.

Sort:
Rasta_Jay

 Brendan you can remove both the bishops and still the chess engines(Not computer or artificial inteligence, like some people in the forum are calling it) will show that black has advantage, but it doesn't.. It has nothing to do with how computers think, rather how they are programmed to think.. We(humans) are not yet intelligent enough to create an autonomous inteligence, That can learn new things by itself and improve intelligently.. All we have are computers that humans have programmed, from might I say, not apex intelligence..

Brendan_UK

Deep learning & AI are developing at an exponential pace, let's hope it makes the World a better place.

DikyAV

version of the solution of Penrose's chess puzzle

 

null

 

igorbielopolskyi

Sir Roger Penrose and the chess puzzle

Anyone thought of that green cross on d7?

Can it be a hint?

Because if we treat it as "incomplete problem", which is possible for a human, but not for a computer, and this cross represents a missing piece, we say that solution is: Qb7#

AndrewSainsbury

http://www.thequietmove.com/man-against-the-machine/ This is our interpretation of the solution. You may disagree...

Xonatron

I do not understand. How does a modern chess computer not solve this position? Aside from the -30 score (which eventually turns to 0.00 score), does a modern chess engine ever make an incorrect move?

Xonatron
uri65 wrote:

"...the primary purpose of evaluation function is not to say who is going to win, not to predict an outcome of the game, but to compare relative values of the candidate moves" 

 

Exactly.

So, are chess engines view as misunderstanding this problem because the evaluation shows -30 rather than 0.00 (even though they eventually solve it to 0.00 as play continues)? This would only be wrong if there were candidate moves view with a higher score that are worse, and when you open up the engine analysis you can see this is not correct (for modern engines like Stockfish 8).

Kezzba

Hi Im Kezzba, stalemate is ezy simply let black kill the loose pawn at the top, then checkmate is impossible, there's only black bishops that black can use as all other pieces locked up if nothing is captured with the pawns essentially white squares are never defended there's always safe squares . White should win every time!!!! To win simply take your King on a stroll visiting all his black friends (undefended black squares) there's 4 coincidently(or perhaps not)  the same number as white pawns,  do this until the the bishops are configured in a way that allows safe promotion of the pawn to a bishop or queen or a hoarse(horse is harder). This will happen, as black  will move the bishops to check you when on a undefended black square. BTW I haven't actually tried this solution so if I sound like an idiot it's probably because I am lol, I don't be interested to know if I'm right wrong or close does anyone here know a link where I can try this puzzle? Cheers!! just found this site, so looking forward to a game or 2,  hard to find people that play see yaz on the board

PageBrix

just another clickbait bullshit...

Scott54935

Easy draw.  The Bishops can't mate, and the rest are trapped.  Just don't move the pawns.

Philloveschess
How do you mate. That seems iimmmpppooosssssiiiiiiiiiibbbbblllllllllllleeeeeeeeeeeee
Philloveschess
Can anyone just notate the answer to checkmate?????
Philloveschess
It’s impossible to checkmate sens u can’t move a pawn because it will open up a pease that is trapped
Winecorkscrew

From a purely mathematical standpoint, a computer would need to calculate extremely deeply as there are many many ways to draw here. On the other hand, a human can recognize the patterns involved in this position and realize that the only option is to draw with K moves, or else lose. A bit silly in my opinion. It's an exercise in computational depth and nothing else. The 50 move rule will eventually determine a draw. White's only option is to keep on truckin' with the K eventually leading to a draw. Besides, how is a position like this even possible in a real game? The question I've pondered since I learned how to play many moons ago is if there is a solution to chess. I believe there is since chess is a closed and finate system despite is extreme complexity - just my humble opinion.  

ConcernedFork

@Winecorkscrew Completely agree (funny that your name is also a tool). The engine can't see it's a draw because having several bishops makes the branching factor to high for each move, in short, it needs to calculate too many positions each turn before reaching the 50 move draw rule. If you delete the bishops and only leave one, it's still a draw and the computer sees it immediately:
You can make many different positions that are also a draw like this, but that the computer can't solve:
This is also a draw because the black king can't ever capture the white king, he can always escape using the light squares (you can see it's a draw more easily if you fill all black squares with black bishops). At most he could stalemate him, and not even that. The computer here also thinks black is better because the bishops generate too many possible moves. However, I'm not sure how the engine knows that is a draw when there is only one bishop in the first puzzle. I guess is because of pruning and some specific rule (like king and bishop vs king is a draw or something like that).