/chess/

Is there a way to create a perfect computer to win every game or is chess one of those games where one counter move can make the game unwinnable?

Kasparov was able to beat deep blue, and Magnus Carlson was able to beat the app that was created to beat him

I don't play chess but am wondering why a program couldn't account for all possible moves with regard to all possible counter moves

>>8657380
The number of possible moves is too great to possibly account for all of them. They're actually still trying to calculate the number of possible positions after 6 moves from each side (12 ply).

Computers use subjective valuations of positions to approximate the best moves. They have gotten much, much better than humans at this to the point where they beat top grandmasters consistently. Mind you, Deep Blue was almost 20 years ago. I don't know about the app put put Carlsen up against stockfish or Komodo with enough time and he'll get his ass handed to him 100 out of 100 times. Hikaru Nakamura recently completed a series of odds games (e.g. start a knight up) against Komodo and lost.

Every once in a while the impossible will happen though and a human will beat a computer, because the valuations are still subjective. Think of it as a normal distribution falling outside its 5 sigma zone once every few million times.

>>8657375
Is chess your mother? If so I banged chess last night.

File: 1485255863125.jpg (103KB, 960x724px) Image search: [Google]

103KB, 960x724px

chess is pretty dank. It's a shame I'm not particularly good at it. The homeless guys at Central Park could probably make a decent living off of my ineptitude.

Anyways, knight to g5 looks juicy and I would immediately resign as black were it to be played.

File: IMG_2429.png (13KB, 409x360px) Image search: [Google]

13KB, 409x360px

Chess is just a game for assburgers to think they're smart when in reality they arent. It requires absolutely no high level intelligence hence why computers are the best at it. It is the equivalent of thinking you're smart because you can multiply large numbers in your head. Everyone just thinks you're a cute autist and claps because you're a good little circus freak that no one cares about. I realised this after playing it for a couple weeks, if you actually play this game unironically as a hobby you should really reevaluate your life.

>>8657375

A difficult concept to learn but what is really important in an otherwise equal endgame position with king pawn endings is the idea of gaining the opposition in king pawn vs. king pawn endgames.

Since at that juncture the endgame results in a battle of tempo, the correct moves are often non-inuitive (moving your king back so they move their king forward so you can move your pawn up and they lose the opposition).

This is known as distant opposition.

>>8657375
Qxh7!!

I like how two continuations end with OO# and OOO# -- castling is almost never the checkmating move.

>>8657375
Mindless useless games for immature kids.

>>8657375
Is this a Josh Waitzkin game?

>>8657512
No, Lasker v Thomas
>>8657461
Correct! It's a shame Lasker played Kd2# instead, still a beautiful sequence though.

Chess is a waste of time brainlets play

>>8657380
>counter move
>not playing a counter move to counter the counter move
>chess brainlet

>>8657380

I'll answer this but bringing up something completely different. Recently there was a Go program that was able to beat a 3-dan player (or 5-dan?) and if you know anything about game engines, Go has been notoriously to get right from the computational aspect of it.

Deep Blue is a dated example -- and I don't know anything about the app thing -- but given our advances in convolutional neural networks and other ML/AI advances, we could almost assuredly create a system to beat Kasparov now (assuming he plays at the same strength as his prime years).

An app doesn't sound nearly powerful enough to contain any sort of meaninful amount of information in either terms of computational prowress or memory capacity to hold the ability to beat Magnus. That's if the app didn't contact a server.

>>8657375
>Mate in 8, slightly faster (assumed less moves) is possible.
>Wot/10

The most important endgames to study depend on you.

>>8657639
I said mate in 8 in the game, faster is possible.

The game continuation led to checkmate in 8 moves. It's possible in 7 however.

>>8657639
I'll specialize this statement. The most important endgames depend on your opening and the popular responses to those lines that create similar patterns of endings. Assuming you aren't under 2100 FIDE/USDF and are playing a level where these things consistently come up.

>>8657380
>Is there a way to create a perfect computer to win every game or is chess one of those games where one counter move can make the game unwinnable?
Perfect play has not been demonstrated for Chess or Go so the answer is unknown. That said computers are already as good or better than the best humans for both Chess and Go.

In the end Chess is a solvable problem like all perfect information games. Perfect play does exist, it just is too computationally expensive to be brute force proven with current computers.

File: CxOc7NuVIAAeAt.jpg (155KB, 1000x751px) Image search: [Google]

155KB, 1000x751px

>wants to git gud at chess to gain social cred points
>be a brainlet who can't reason on things outside my narrow field

Any brainlet friendly introduction to chess someone would recommend?

>>8657663
route memorization of openings and finishes

>>8657663
Anything by Yasswer Seirwan or Jeremy Silman's "The Amateur's Mind." This is assuming you now the basic moves and basic structure like opening/middle game/endgame.

>>8657663
Get a premium on chess.com and watch videos. It's brainless friendly because you don't actually have to do anything.

>>8657380
Magnus Carlsen isn't able to beat top engines or any other person for that matter. He beats a chess iPhone app. Engines in competitions are all rated over 3000 elo.

>playing chess
>not just using genetic programming to get AI to play on your behalf
It's like you don't want to be smart

>>8657666
>>8657694
>>8657720
Thanks guys.

>>8657722
>t. Ivanov

go away, you are still banned worldwide

(so much salt in this thread! is this how humanities majors whine when they flunk out of intro math and science classes?)

>>8657720
This is exactly why chess is a stupid game

>>8657437
> assburgers

kek

File: PawnEndgame.png (150KB, 523x505px) Image search: [Google]

150KB, 523x505px

>>8657445
here is a good example: there is only one move that wins this endgame. Eventually, lots of triangulation is involved.

>>8658239
White to move?

>>8657375
>go for brainlets

All other tabletop games are the brainlet version of catan

>>8657375
OP, show the solution to puzzle

>>8659815
Not OP, but a youtube search on "Lasker v Thomas Qxh7!!" (Reading the thread, it's the match and the winning move) raises : https://youtu.be/S6X6g_-W-rw

>>8657403
>>8657468
>>8657528
>>8657967
found the lamer-gamers

>>8659459
yes

>>8659823
and here is the faster mate ending in castling:

1. Qxh7+ Kxh7 2. Nxf6+ Kh6 3. Neg4+ Kg5 4. f4+ Kxf4 5. g3+ Kf3 6. OO#

>>8657437
b-but burgers are not good at chess, chess == europe, few usa chess players with high rating are just immigrants from south and east europe

>>8657375
>mate in 8
Nxf6
black forced to take
Qxh7 mate

>>8661072
Nevermind I'm retarded, this is why I'm only 1600