Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

gogameguru.com

Post reply on HN
Print view View original

41–50 of 134 posts

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#41

Post by gcp »

CocaKoala wrote: post #17
T-hawk wrote: post #12

Earlier quoted context omitted.

Would it be accurate to say those 4 stones are about equivalent to one pawn in chess? As a back-of-the-envelope calculation, one pawn is 1/43 of your starting material in chess if the king is included as a four-point fighting piece. Four stones in Go comes to 1/45 of your expected endgame material as half of a 19x19 board.

It's hard to make a comparison like that because Go and Chess operate on fundamentally different models. Chess is about capturing a specific piece and putting your opponent in a position where they have no options but to lose; to that end, it's a war of attrition where players slowly lose resources until the game is over (note that I do not play chess at any significant level). In contrast, Go is about capturing terr…

It's hard to make a comparison like that because Go and Chess operate on fundamentally different models

It is very easy. The nature of the game is completely irrelevant. The question is what expectation of winning rate a 4 stone advantage gives to a Go player. For middle ranks, one stone is about 100 ELO-equivalent, or 64% winning rate. (I think it's closer at top dan ranks, would have to look up the statistics). 4 Stones is about 92% chance of winning between equal players.

The equivalent in chess is slightly more than 4 pawns advantage, or about a minor piece.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#42

Post by baddox »

thinkpad20 wrote: post #35

Combinatorically, Go is considered pretty much the most complex game in existence, and computer go programs have been historically weak, so this seems to be quite an achievement (despite that it was against 4-stone-handicapped master passed his prime). One thing I wonder, though, if computers ever catch up to humans at Go, could we simply expand the board size a few spaces, thus dramatically increasing the problem sp…

> Combinatorically, Go is considered pretty much the most complex game in existence

Is that considering "game" to include only traditional turn-based board games? Presumably it would be trivial to make a video game that's vastly more complex (at least in terms of tree or state space complexity) than Go. Any real-time strategy game should quality.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#43

Post by awwducks »

valarauca1 wrote: post #22
cruise02 wrote: post #21

Earlier quoted context omitted.

A four stone handicap in Go is way less than a four piece handicap in chess. Also, Crazy Stone is 6d on KGS.

True I edited it down to just bishops. And KGS is very inflated, a 6d on KGS vs. a 6d international professional wouldn't be contest. The normal conversion I hear is a 3d on KGS is around a 1k on Eastern Servers. Or that's my experience for people who play on KGS and eastern servers. I'm a 4k on KGS, but I know my knowledge is a bit amateur or a South Korean 4k.

There is a difference between amateur levels and professional dan levels.

For reference, you probably have to be at least 7-8 dan on KGS to be close to being professional.

Here is a rank comparison of various amateur systems:

http://senseis.xmp.net/?RankWorldwideComparison

All of those systems will be weaker than any professional system.

It's inaccurate to say that KGS is "very inflated" since it's measuring amateurs instead of professionals.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#44

Post by dragontamer »

visarga wrote: post #14
ChuckMcM wrote: post #13

Wow, I was not aware that Go programs had gotten this good. That is excellent news. No a bunch of new papers to add to my list to read.

Crazy Stone got 4 stones handicap, I think. It wasn't a level playing field yet.

A few years ago, it was impossible for a computer to beat a pro with a 9-stone handicap.

This is a major advancement for Go AI, no matter how you slice it.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#45

Post by billforsternz »

89vision wrote: post #39

confusing title, I was expecting golang

Not confusing at all. Go laid claim to the name hundreds of years ago, and the title made it clear it was the game. If someone invented a computer language called chess and there was an article title "Houdini computer chess program defeats former world champion Garry Kasparov with knight handicap" would you expect it to be about the computer language ?

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#46

Post by awwducks »

baddox wrote: post #40
valarauca1 wrote: post #20

To put in more westernized chess terms, this would be like beating a 2100Elo master without bishops. Yoshio is a professional and very skilled, but he hasn't won a major tourniquet since 1989. Really not an achievement, its nice to see Computer Go advancing. A 4 stone handicap between master's is massive. The difference between a 9d (dan) and 1d isn't actually 8 stones. With an 8 stone handicap a 1d will play a 9d of…

What is a professional amateur circle?

He probably means "insei" or "yeon'gusaeng". [1]

That is, students striving to become professional.

[1] https://en.wikipedia.org/wiki/Insei

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#47

Post by awwducks »

dragontamer wrote: post #44
visarga wrote: post #14

Earlier quoted context omitted.

Crazy Stone got 4 stones handicap, I think. It wasn't a level playing field yet.

A few years ago, it was impossible for a computer to beat a pro with a 9-stone handicap. This is a major advancement for Go AI, no matter how you slice it.

Absolutely. A few years ago, even I, a low amateur dan, was able to beat the strongest AIs with the computer taking 9 stones.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#48

Post by thinkpad20 »

gcp wrote: post #38
rememberlenny wrote: post #37

Earlier quoted context omitted.

Interesting question. As a go lover I believe it would hinder the person more than the computer. You wouldn't be able to expand the board, because then its no longer Go. There are three commonly used board sizes (9x9, 13x13, and 19x19). Other than that, it wouldn't be Go. The computer would have an easier time playing with a larger board because its algorithmically making moves. The experience go player wouldn't have…

Some go programs (e.g. Leela) support playing on large boards, up to 37x37. Algorithmically, there are two issues: Go programs use Monte Carlo estimation by doing semi-random playouts to determine the game theoretic value of a position. The longer the game is from its end, the more noise in the estimate, which worsens the playing strength. Secondly, the longer the game is from its end, the longer the computation of a…

Yeah, this is my thought. It seems, for example, that precisely calculating the optimal strategy would be quite tenable on a 9x9 board, because the problem space is so small. On a 19x19 board, the space is absolutely immense. It doesn't seem to make a huge difference with humans, though, because we are so adept at pattern-matching. So it seems like while a larger board would certainly make things more difficult to a human, they wouldn't really have to make many adjustments in their fundamental strategy. On the other hand, it could dramatically weaken a computer player.

As to the other point raised, there doesn't seem to be any reason why it would suddenly cease to become Go if you increase the board past 19x19. The rules of Go are independent of board size, and I'm sure it took quite a long time before the "standard sizes" of 9, 13, and 19 were fixed.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#49

Post by wtallis »

baddox wrote: post #42
thinkpad20 wrote: post #35

Combinatorically, Go is considered pretty much the most complex game in existence, and computer go programs have been historically weak, so this seems to be quite an achievement (despite that it was against 4-stone-handicapped master passed his prime). One thing I wonder, though, if computers ever catch up to humans at Go, could we simply expand the board size a few spaces, thus dramatically increasing the problem sp…

> Combinatorically, Go is considered pretty much the most complex game in existence Is that considering "game" to include only traditional turn-based board games? Presumably it would be trivial to make a video game that's vastly more complex (at least in terms of tree or state space complexity) than Go. Any real-time strategy game should quality.

There are plenty of deterministic games more complex than Go, even turn-based ones. But Go is still a clear outlier due to its age and the simplicity of its rules - it's completely obvious how to enumerate the game tree.

Things like a real-time strategy game aren't really a fair comparison, since they are designed to approximate a continuous game space-time as opposed to the discrete board and moves of Go. You can inflate the game tree complexity just by switching from single-precision to double-precision, but that doesn't really make the game any harder.

Re: Crazy Stone computer Go program defeats Ishida Yoshio 9 dan with 4 stones

#50

Post by baddox »

wtallis wrote: post #49
baddox wrote: post #42

Earlier quoted context omitted.

> Combinatorically, Go is considered pretty much the most complex game in existence Is that considering "game" to include only traditional turn-based board games? Presumably it would be trivial to make a video game that's vastly more complex (at least in terms of tree or state space complexity) than Go. Any real-time strategy game should quality.

There are plenty of deterministic games more complex than Go, even turn-based ones. But Go is still a clear outlier due to its age and the simplicity of its rules - it's completely obvious how to enumerate the game tree. Things like a real-time strategy game aren't really a fair comparison, since they are designed to approximate a continuous game space-time as opposed to the discrete board and moves of Go. You can in…

> You can inflate the game tree complexity just by switching from single-precision to double-precision, but that doesn't really make the game any harder.

I think that absolutely makes the game significantly harder, just like playing Go on a smaller board is easier than Go on a larger board.

Post reply on HN