Page 1 of 1

Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 5:07 am UTC
by Six Fingers
It seems like a lot of math movies have the obligitory "Go scene", usually held in the park or at someone's house. Is there a reason for this? Why are mathematicians so attracted this game?

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 6:59 am UTC
by t0rajir0u
As far as I know, Beautiful Mind is the only one, although I wouldn't be surprised if there are others. Either way I don't think Go is better represented in math movies than chess, so this could all be chalked up to intellectual games in general.

But a big aesthetic reason to be interested in the game is that it's a great example, like the Game of Life, of how complex behavior can arise from a set of simple rules. Any given Go game involves the formation of continuously varying patterns that seem random to a casual observer but are full of meaning to the trained eye. And Go strategy is far more complex than the simple rules of the game would suggest.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 9:21 am UTC
by JayDee
I would imagine t0rajir0u is right. Chess is traditionally the board game people are protrayed as playing in film if they're supposed to be intelligent. Go has more complexity, so I guess it makes the statement more emphatically. Personally, I think it help that the binary nature of the game makes it seem more 'mathematical' compared to the 'strategical' intelligence of chess.

Oh, and the first place I'd've thought to check:
TvTropes wrote:Go: the supreme Eastern test of intellect. Several orders of magnitude more complex than chess (not quite the same as "more complicated than chess"). Knowing how to play well typically signals a character has likewise intellectually surpassed "mere chess". The aura of inscrutable Asian wisdom doesn't hurt either, though in reality playing either game at world championship level is equally difficult. And then, in the other direction, there's...

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 11:55 am UTC
by Quenouille
As far as I know, Beautiful Mind is the only one, although I wouldn't be surprised if there are others.


There's also pi.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 12:01 pm UTC
by skeptical scientist
Six Fingers wrote:It seems like a lot of math movies have the obligitory "Go scene", usually held in the park or at someone's house. Is there a reason for this? Why are mathematicians so attracted this game?

I think it's a myth. During my first year of grad school, I brought a board to my office, and tried to get some other people interested, or maybe play a few. Five years of leaving the board in my office later, I have yet to play anyone else in the department...

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 12:11 pm UTC
by notzeb
I was disappointed to find that out of my entire college (including grad students and staff), a mere six people show up to go club... and not one of us is dan level.

Stupid stereotype... getting my hopes up... But if it were true I'd say the reason is that mathematicians like challenges, and go is challenging.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 12:40 pm UTC
by stephentyrone
Go endgames have a remarkably interesting mathematical structure (much more so than chess). Some mathematicians become interested in Go that way.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 1:53 pm UTC
by lu6cifer
I think computer scientists are more inclined to play Go than mathematicians. I'll post a link about this that I once stumble upon'ed if I can find it.
Fun fact: Computers can already beat humans at chess, but not at Go.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 3:38 pm UTC
by notzeb
lu6cifer wrote:Fun fact: Computers can already beat humans at chess, but not at Go.
I've lost to computers at Go before.

Are you trying to imply that I am not human?

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 3:47 pm UTC
by Yakk
Chess:Go
==
C++:Group Theory

But yes -- Chess in a movie is a "this is a smart character" trope. Go is an alternative that is more exotic.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 3:55 pm UTC
by achan1058
skeptical scientist wrote:I think it's a myth. During my first year of grad school, I brought a board to my office, and tried to get some other people interested, or maybe play a few. Five years of leaving the board in my office later, I have yet to play anyone else in the department...
There's a few people who play it in our department, among with Magic the Gathering and the like. Strangely, nobody plays chess, though.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 3:56 pm UTC
by achan1058
notzeb wrote:
lu6cifer wrote:Fun fact: Computers can already beat humans at chess, but not at Go.
I've lost to computers at Go before.
I think he means a good player, not an amateur.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 4:34 pm UTC
by t0rajir0u
Actually, the status of that statement is changing; recently Go AI using Monte-Carlo methods have been doing quite well against human opponents.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 4:58 pm UTC
by Ended
Go possesses large amounts of symmetry, which mathematicians tend to be fond of. (In particular there is no distinguished orientation for the board at the start of the game, which is unusual for a board game.) I think the comparisons with the Game of Life and group theory are apt.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 5:22 pm UTC
by skeptical scientist
t0rajir0u wrote:Actually, the status of that statement is changing; recently Go AI using Monte-Carlo methods have been doing quite well against human opponents.

Changing, maybe, but it hasn't changed yet. According to the Wikipedia link, Many Faces of Go is the best computer program, and according to the M.F. Go website, at its strongest level it's 2 Kyu on a 19x19 board, or 3 Dan on a 9x9 board. That means on a 19x19 board it's equivalent to a top-tier amateur player, and on a 9x9 it's as good as a professional player, but not a top-tier pro. (This in contrast to Chess, where top computer programs are almost certainly stronger than the top pros.)

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 10:30 pm UTC
by Sir_Elderberry
Well, you could show them playing chess. But lots of people actually know how to play chess. So watching someone play chess isn't all that impressive, because I, Joe. Q. Viewershmuck, can play it too. On the other hand, far few people play Go.

Re: Why Do Mathematicians Play Go?

Posted: Fri Sep 04, 2009 10:56 pm UTC
by Yakk
True. And when you look at a board, and say "well, duh", or you see supposedly smart people play to checkmate...

Re: Why Do Mathematicians Play Go?

Posted: Sat Sep 05, 2009 12:49 am UTC
by Captain_Thunder
I started playing Go before I got interested in math, so maybe someone who enjoys one activity is likely to enjoy the other? Then again, I had been programming for years before I started playing Go, thus reinforcing the idea that computer scientists are the ones most likely to be drawn to it.

I really wish more people would learn about Go like they learn about Chess; it's a much deeper and more interesting game.

Re: Why Do Mathematicians Play Go?

Posted: Sat Sep 05, 2009 2:47 am UTC
by t0rajir0u
Captain_Thunder wrote:I started playing Go before I got interested in math, so maybe someone who enjoys one activity is likely to enjoy the other?

This statement doesn't really seem to be supported by the history of Go in, say, Japan. I'd say this is a side effect of the channels through which Americans in particular are likely to get interested in Go, which maybe preferentially attract a certain kind of person.

Re: Why Do Mathematicians Play Go?

Posted: Sat Sep 05, 2009 3:24 am UTC
by The EGE
JayDee wrote:Oh, and the first place I'd've thought to check:
TvTropes wrote:Go: the supreme Eastern test of intellect. Several orders of magnitude more complex than chess (not quite the same as "more complicated than chess"). Knowing how to play well typically signals a character has likewise intellectually surpassed "mere chess". The aura of inscrutable Asian wisdom doesn't hurt either, though in reality playing either game at world championship level is equally difficult. And then, in the other direction, there's...


...and I just lost half an hour. Dammit JayDee, do not link to the magical world of constant novelty...

Re: Why Do Mathematicians Play Go?

Posted: Sat Sep 05, 2009 9:39 am UTC
by Incompetent
I think mathematicians are just generally more inclined to engage in complex board/card games and puzzles than the general population. The same probably goes for scientists of all kinds. I think the buzz people get out of doing puzzles, solving mysteries and figuring out how to win at games is essentially the same as that of doing research, so it's not surprising that the same people are drawn to both.

As for Go specifically, it gets major mathmo points because it's so intricate and difficult to analyse, despite its simple rules.

Re: Why Do Mathematicians Play Go?

Posted: Sat Sep 05, 2009 2:36 pm UTC
by Captain_Thunder
t0rajir0u wrote:
Captain_Thunder wrote:I started playing Go before I got interested in math, so maybe someone who enjoys one activity is likely to enjoy the other?

This statement doesn't really seem to be supported by the history of Go in, say, Japan. I'd say this is a side effect of the channels through which Americans in particular are likely to get interested in Go, which maybe preferentially attract a certain kind of person.


So, the interwebs, basically? :mrgreen:

I like the Hikaru avatar, by the way.

Re: Why Do Mathematicians Play Go?

Posted: Sun Sep 06, 2009 9:12 pm UTC
by HenryS
skeptical scientist wrote: According to the Wikipedia link, Many Faces of Go is the best computer program, and according to the M.F. Go website, at its strongest level it's 2 Kyu on a 19x19 board, or 3 Dan on a 9x9 board. That means on a 19x19 board it's equivalent to a top-tier amateur player, and on a 9x9 it's as good as a professional player, but not a top-tier pro. (This in contrast to Chess, where top computer programs are almost certainly stronger than the top pros.)

I don't think this is accurate. Amateur ranks go from beginner at around 30 kyu up to 1 kyu, and then there are amateur dan ranks from 1 dan up to around 7 dan. Then there are professional dan ranks from 1 dan to 9 dan, although these are on a denser scale of around 1/3 dan rank for each amateur rank. I'm a British amateur 2 dan, so I should be able to handily beat Many Faces (in an even game without handicap stones), and I'm certainly not near professional level.

Re: Why Do Mathematicians Play Go?

Posted: Sun Sep 06, 2009 9:52 pm UTC
by lightvector
HenryS wrote:
skeptical scientist wrote: According to the Wikipedia link, Many Faces of Go is the best computer program, and according to the M.F. Go website, at its strongest level it's 2 Kyu on a 19x19 board, or 3 Dan on a 9x9 board. That means on a 19x19 board it's equivalent to a top-tier amateur player, and on a 9x9 it's as good as a professional player, but not a top-tier pro. (This in contrast to Chess, where top computer programs are almost certainly stronger than the top pros.)

I don't think this is accurate. Amateur ranks go from beginner at around 30 kyu up to 1 kyu, and then there are amateur dan ranks from 1 dan up to around 7 dan. Then there are professional dan ranks from 1 dan to 9 dan, although these are on a denser scale of around 1/3 dan rank for each amateur rank. I'm a British amateur 2 dan, so I should be able to handily beat Many Faces (in an even game without handicap stones), and I'm certainly not near professional level.


Indeed. I'd put 2k as a middling-level club player (this will vary a lot depending on the club however), and 3d roughly as an expert, but still far from the top amateurs and definitely not a pro.

These figures, I believe, are the strengths for top-line but common hardware. When parallelized over giant clusters with hundreds of cores, the current best Go programs reach maybe around 1d-2d on 19x19, and perhaps 5d to almost pro on 9x9.