[ 3 / a / adv / an / asp / cgl / ck / co / diy / fa / g / gd / int / jp / k / lit / m / mlp / mu / n / o / out / p / po / sci / sp / tg / toy / trv / tv / v / vg / vp / vr / w / wsg / x]

/sci/ - Science & Math - Chess !

<< back to board
[Delete this thread]

Chess ! Anonymous 06/26/14(Thu)22:38 UTC+1 No.6613001 Report

I read something that said that the number of possible Chess games up to move 40 exceeds the number of stars in the universe.
Is that right??
(and no, it wasn't neil degrasse tyson lol)

How many possible Chess games are there where someone wins?
Is that even possible to calculate?

Chess people in the know: Did IBM cheat in '97?

Let's talk about Chess and its cool / interesting mathematical associations and implications. :)
Anonymous 06/26/14(Thu)22:43 UTC+1 No.6613010 Report

>the number of possible Chess games up to move 40 exceeds the number of stars in the universe.

White starts with 18 possible moves (2 moves per Pawn, both Knights)
Black has 18 possible responses,
Then the pieces start interacting...
Each player has X possible moves, Y possible counter moves and Z possible counter moves to EACH SET of possible moves that took place before...

I don't know anything about math,
but I would think that the number of "possible games" goes off the charts very, very fast.
Anonymous 06/26/14(Thu)23:08 UTC+1 No.6613059 Report

from memory, so I stand to be corrected...
>number of possible Chess games up to move 40 exceeds the number of stars in the universe
number of possible chess games = 10^120, number of ATOMS in the universe = 10^68
this is also the argument for why chess will never be 'solved' (like checkers has been) - insufficient storage space.
>How many possible Chess games are there where someone wins?
either player can resign on any move for any reason; therefore twice as many as the number of possible games.
>Did IBM cheat in '97?
supposedly they withheld logs of deep blue's games, and deconstructed it after the match, whether or not that's cheating... there are youtubes about it if you're really interested.
Anonymous 06/26/14(Thu)23:16 UTC+1 No.6613077 Report

>number of possible chess games = 10^120, number of ATOMS in the universe = 10^68

>this is also the argument for why chess will never be 'solved'

It's a bad argument because when calculating chess moves you can prune moves that are known to give the other side a clear path to checkmate in x moves.

let's say white has 18 moves available to it. One of the moves is known to result in a 7 move checkmate. So we can remove it an all the moves that come after.

This seems like it won;t prune much, however after the game starts, the number of possible moves that do not result in checkmate in x moves falls very quickly where most of the time each side only has maybe 2-3 moves available that are delaying checkmate more than 7 moves. Computers can calculate 7 move checkmate very easily for all positions.

This way it would probably be possible to put into memory a data base of all outcomes.

Current algorithms are very good and the elo is rising by 100-200 a year for chess machines. If this rate continues chess will be "solved" in the sense you cannot beat a machine in a few years.
Anonymous 06/26/14(Thu)23:40 UTC+1 No.6613118 Report

Not really arguing with your logic which makes sense, but to know it's checkmate in 7, you (as a machine) have to store that data otherwise you reach that position and don't know you're won.
Humans can prune moves in the sense you're talking about, because they can intuit the outcome, machines cannot.
'Conventional Wisdom' considers the number of legal moves in any given position (late opening / middlegame) to be about 30.
Elo ratings are valid only within a pool of players - humans don't play machines in tournament chess - so a 2800 Elo machine is not the same as a 2800 Elo human. That said, between about 1990 and 1995 machines progressed quickly from the joke of being too weak to contend with humans, to being too strong to make a meaningful contest worthwhile, the algorithms are now that good (but this is not the same as solving chess). Like why human weightlifters don't compete with forklift trucks.
Anonymous 06/26/14(Thu)23:45 UTC+1 No.6613132 Report

So? the amount of possible frequencies you can fort exceeds infinity. So what?

Pathetic combinatoric example.
Anonymous 06/27/14(Fri)00:03 UTC+1 No.6613183 Report

you can save a lot with subroutines
Anonymous 06/27/14(Fri)00:04 UTC+1 No.6613187 Report

however this game...
Anonymous 06/27/14(Fri)00:09 UTC+1 No.6613202 Report

It would seem like Go is much simpler than chess, how can it be that it's actually harder for a computer to solve? Is it just because of the sheer size of the board?
Anonymous 06/27/14(Fri)00:15 UTC+1 No.6613218 Report

>just because of the sheer size of the board
Anonymous 06/27/14(Fri)00:18 UTC+1 No.6613226 Report

it's kind of ridiculous
it's kind of silly to think of go strategy in advance, it's more like learning a language and having good conversations
Anonymous 06/27/14(Fri)00:18 UTC+1 No.6613227 Report

Go and chess are diametrically opposed; the one starts with an empty board that must be filled, the other starts with a full board that must be emptied.
Anonymous 06/27/14(Fri)00:21 UTC+1 No.6613231 Report

That has nothing to do with his question.
Anonymous 06/27/14(Fri)00:24 UTC+1 No.6613237 Report

Go has nothing to do with OP's question... what are the mathematics of Chess?
Anonymous 06/27/14(Fri)00:35 UTC+1 No.6613257 Report

>either player can resign on any move for any reason; therefore twice as many as the number of possible games.
Well no, because each game where one side wins is necessarily a possible game. But there are some games where neither side wins (i.e. draws) so there are strictly fewer games where somebody wins than possible games.

Unless the number of "possible games" excludes things like resignation or mutually agreed draws and only counts games played completely out. But then Chess technically becomes an infinite game since there are hypothetical games which do not end in checkmate, stalemate, or where either side wants to claim a draw (things like three-fold repetition have to be claimed, they're not automatic).
Anonymous 06/27/14(Fri)00:41 UTC+1 No.6613272 Report

Let's refine that statement to "can choose to resign"; but true, a repetition could go on ad infinitum without either player declaring the draw.
Anonymous 06/27/14(Fri)00:59 UTC+1 No.6613301 Report

if the 50 move rule and 3 rep rule are enforced there is a finite number of possible games
Anonymous 06/27/14(Fri)01:02 UTC+1 No.6613304 Report

Did you even bother to read what you're commenting on?
Anonymous 06/27/14(Fri)01:06 UTC+1 No.6613306 Report

The possible move space of Go is vastly larger than in Chess. In chess, there are (starting out) 18 possible initial moves for each side, leading to a total of 324 possible openings (most of which are terrible); with Go, you can place a piece literally anywhere on the 19x19 board each move, so there are 129,960 possible opening board states.

Also, Go depends vastly more on long-term strategy, so you need to look many more moves ahead compared to chess; and because of the enormous possibility space compared to chess, looking N moves ahead is exponentially harder than it is in chess.
Anonymous 06/27/14(Fri)02:01 UTC+1 No.6613390 Report

Bumping an actually interesting thread.

If anyone wants to play on Lichess let me know btw.
Shitty chess-n00b here, so you'll probably beat me, but I like playing :)
Anonymous 06/27/14(Fri)02:08 UTC+1 No.6613395 Report

I wish people actually played this in the West. Obviously it exists somewhere but it has a weak presence.

Playing gokgs is very intimidating though, and playing it in real life looks so much more awesome.
Anonymous 06/27/14(Fri)02:09 UTC+1 No.6613402 Report

Those rules don't force a draw though. A draw has to be accepted.
Anonymous 06/27/14(Fri)02:40 UTC+1 No.6613442 Report

I think there are severe real estate problems, and a multitude of moves which will never happen, because any player with a desire to win won't sacrifice pieces or refuse to take pieces offered in sacrifice.
Anonymous 06/27/14(Fri)02:54 UTC+1 No.6613478 Report

>any player with a desire to win won't sacrifice pieces or refuse to take pieces offered in sacrifice.
Amateur detected.

Tell that to Bobby Fischer, whose Immortal Game is famous for its sacrifice of the Queen,
or to Deep Blue, who caused an extreme controversy when it refused to accept a sacrifice made by the 15-year world champion -- and in doing so secured the long-term win!
Anonymous 06/27/14(Fri)02:54 UTC+1 No.6613481 Report

It depnds. Sometimes it is really worth sacrificing pieces if your opponent will be forced from a position on the board. Or if sacrificing them means you get to bring your pieces out where your opponent does not effectively giving you the tempo lead and leaving the other player responding to your moves.
once the other player has to respond to your moves you are more likely to be able to force an advantage
Anonymous 06/27/14(Fri)02:58 UTC+1 No.6613485 Report

Jesus Christ, I knew this would pull you haughty fuckers out of the woodwork. There are many more random moves which will offer useless sacrifices than there are FAMOUS SACRIFICES DONE BY GENIUSES. Great job applying your understanding to the TOPIC AT HAND.
Anonymous 06/27/14(Fri)03:25 UTC+1 No.6613525 Report

Why are you so upset?

It's not our fault you can't write clearly
and like to opine on subjects about which you know very little lmao
Anonymous 06/27/14(Fri)04:23 UTC+1 No.6613601 Report

How do I get better at chess?

I've been reading about tactics / strategy every day and watching commentary of famous games,
but I still SUCK :(

Although, I managed to check-mate the computer at Lichess the other day for the first time, which was very encouraging :D
Anonymous 06/27/14(Fri)04:47 UTC+1 No.6613664 Report

There are an infinite amount of move sequences that result in a victory in chess. After all, there is no limit to the amount of turns that can be taken.

If you limited the number of moves, then you could calculate an actual number, but it would be incomprehensibly large.
Anonymous 06/27/14(Fri)05:14 UTC+1 No.6613712 Report

The stockfish ai is very good.

Sometimes you will get unlucky with it and there is absolutely nothing you can do. I have seen even on the 1 setting the stockfish play a perfect game. basically every move there are odds the ai will make a blunder or a mistake. You need to recognize when it makes a mistake and take advantage of it or you will loose.

A normal game it might only make 3 mistakes a;though it is possible it will make none if you get unlucky, so if you are not paying attention you will miss it.

You should probably just use the take back button every time you screw up and play it until you can at least draw it. That is probably the best way to train because it lets you undo mistakes and you won;t waste a huge amount of time going through multiple openings.
Anonymous 06/27/14(Fri)05:18 UTC+1 No.6613718 Report

your an idiot

making / refusing sacrifices are an extremely common part of chess...
Anonymous 06/27/14(Fri)05:30 UTC+1 No.6613729 Report

>Did IBM cheat in '97?

I don't know but they acted like dicks about it. Why be so secretive if you know your computer program is great?
All the content on this website comes from 4chan.org. All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster. 4chanArchive is not affiliated with 4chan.