11 ms·

Questions I would ask God about the game of Go

111 points by ivee 7y ago

blowski 7y ago
I don't know how to play Go, and I'm not very good at maths. How many of these questions could be answered without going to an omniscient heavenly being?
- hyperpape 7y ago
  Go is a finite game, so almost all these questions can be answered by just writing an algorithm that enumerates all the possible games and gives the answer. However, the game space is so big that approach would never finish. Further, I doubt any of them could practically be proven, even with some smarter algorithm. It's akin to how even with all the advances made in computer chess, we still don't know what perfect play looks like. I do think we already have good evidence the best komi is 7 or 7.5, but it's not definitive. The question of what the best ko rule is value based (different rules have different consequences/aesthetics for gameplay).
  - IggleSniggle 7y ago
    From Wikipedia: Komidashi https://en.m.wikipedia.org/wiki/Komidashi https://en.m.wikipedia.org/wiki/Komidashi "Komi in the game of Go are points added to the score of the player with the white stones as compensation for playing second. Black's first move advantage is generally considered to equal somewhere between 5 and 7 points by the end of the game. Standard komi is 6.5 points under the Japanese and Korean rules; under Chinese, Ing and AGA rules standard komi is 7.5 points. Komi typically applies only to games where both players are evenly ranked. In the case of a one-rank difference, the stronger player will typically play with the white stones and players often agree on a simple 0.5 point komi to break a tie in favour of white."
- praptak 7y ago
  Most of them if we either 1. discover an amazing technology for blazing fast computing. or 2. advance our algorithms so that they can prune the search space of Go to something manageable by our CPUs while provably keeping correctness. Does not apply to questions about aliens.
- CydeWeys 7y ago
  Likely none of them; the search space is simply way too big. Heck, we can't even answer most of these questions for Chess, and Chess is a comparatively much easier to compute game.
phkahler 7y ago
The best placement for handicap stones seems most for perfect play. For tic-tac-toe I prefer to start in the middle because "not losing" takes more thought for the other player, but assuming perfect play it doesnt matter where you start - the game is a draw.
elcomet 7y ago
Do things like "best opening move", "worst handicap placement" or similar make sense if you assume perfect play ? I would think not, because when you have perfect play, you either win, lose or draw. So you just have "winning opening moves", "losing opening moves", and "draw opening moves". You cannot really rank them. Maybe the ranking would be the number of different games that lead you to a win, in a monte-carlo tree search fashion. Which could then be useful if you can't play perfectly (highest probability of winning).
- afthonos 7y ago
  Go has points. An opening move is better than another if, given perfect play thereafter, you end up with more points.
  - elcomet 7y ago
    But are those points counted in tournaments, or are they just counted to know who won the game ? So, if you have a very clear path of how to win by 1 point, or a risky way to win by 10 points, does it make sense to try the risky path ?
    - afthonos 7y ago
      They are not officially counted, but they are often used as an informal indication of strength difference among players.
CydeWeys 7y ago
> Is go played anywhere else in the universe, and was it discovered by chance there? > What openings do aliens play? I love these questions about if anyone else in the universe is playing Go. My gut response would be yes -- it's a simple enough game (regular game grid of any size roughly 5x5 on up, one type of piece, two players take turns alternating placing pieces, and any that are surrounded are removed). I suspect other simple enough games that likely exist elsewhere in the universe would be tic-tac-toe, connect-N, dots and boxes, and Othello/Reversi -- any others? Contrast with, say, Chess, which has too many different types of pieces, all with specific movement rules, and with a specific initial layout. I wouldn't be surprised if Chess only existed here, though I also would expect plenty of games similar to Chess to exist out there in the universe, i.e. a game on a regular board with multiple types of pieces and different rules for moving the different pieces. If Shogi, Checkers, or Stratego weren't human games I wouldn't be surprised to learn them from aliens. It's a great question -- what is the most universal board game? What board game has been independently invented the most times throughout the universe? What if it's a game that we haven't even come up with?
- SamPatt 7y ago
  I imagine dice, or some other physical object that introduces randomness, is a sure thing if intelligent life exists elsewhere.
  - GrumpyNl 7y ago
    That implies that they think and behave like us.
    - CydeWeys 7y ago
      Does it? A lot of concepts in board games are pretty natural results of universal mathematics. There's only three regular polygons that will tesselate a board (we have plenty of board games that use all three of them), and there's only three dimensions that games can work in (and we have plenty of games of all three types; Candyland, for instance, is played along a 1D path, and the first player to reach the end wins). Even if some aliens think so differently from us that the concept of a square is unknown to them, surely there are plenty of aliens that would think similarly enough?
mark_l_watson 7y ago
That was fun. My favorite two questions are: Is go played anywhere else in the universe, and was it discovered by chance there? What openings do aliens play? Our Universe is a huge place. I think it was Nature Magazine that had an article estimating 20,000 worlds with earth like conditions and life. My formal education is in physics and personally I subscribe to the multiverse hypothesis so with a potential infinite number of universes these questions have meaning. My older brother taught me to play Go when I was eight and I played actively in the 1970s. Two years ago, on a whim, I decided to go to the US Go Open. I drastically overestimated my own rating when registering, played too strong of players and lost every game. A good lesson in humility and I made some friends. To lick my battle wounds, I paid a South Korean Go professional for lessons for three months and I feel like that really helped my game.
- treve 7y ago
  20,000 seems a little low. I imagine this needs at least 10 more zeroes
  - CydeWeys 7y ago
    Yup. As our telescopes get better we keep finding Earth-like planets just orbiting nearby stars. And that's only what's nearby, here in this one galaxy.
  - mark_l_watson 7y ago
    I agree. My Dad is a physicist and in his late 90s still a member of the National Academy of Science. We riff a lot on science and technology and this is a favorite topic :-)
  - marcoperaza 7y ago
    Any number is a bad guess because we don’t have any sense of the likelihood of life emerging, even in the “optimal conditions”. It could be so unlikely that it’s probable there is no other life, it could be so likely that almost every “viable” planet has life, anywhere in between, or it could be that God created exactly as much life as he wanted, which could be, again, any number. We have a sample size of one and what must be, in the grand scheme, a primitive understanding of physics and chemistry and biology. Throwing around numbers and probabilities for the existence of other life is frankly ridiculous. The best we can intelligently talk about is what the number is given some guess about other numbers. But remember, we have no good numbers for the emergence of life, the survival of life, the likelihood of evolution into intelligent life, etc. Any number multiplied by a totally unknown factor can be literally anything. And before someone throws the infinite universe argument at me, of course. Any probability above zero times infinity is infinity. (Though it’s also possible that there is still no extraterrestrial life even in that case, that outcome always has a non-zero probability!) If we accept a probabilistic emergence of life, then we need to talk about useful things like “in our solar system”, “in our galaxy”, “in the visible universe”, and such. Though it is theoretically possible that the conditions for the emergence of intelligent life are so narrow that only planets exactly like ours, so exactly like it that someone else on that planet has or will typed out this exact comment, have it. In a truly infinite, random (hmm, another bad assumption perhaps?) universe, there would be infinitely many such exact replicas of Earth and its entire visible universe. This isn’t a very useful concept.
mcbuilder 7y ago
Would not Alpha Go Zero and related algorithms give us a good insight into the question about what openings alien's might play? Isn't it really tackling the game from first (or zeroth if you will) principles, just simulated games without the historical baggage coming from human study. It has also considerably changed our own knowledge of the opening.
- Sharlin 7y ago
  I don’t think so. The go that AlphaGo plays is one very specific type of ”alien go” that exploits the strengths of the computing substrate it thinks on. Alien entities that are still evolved via natural selection would probably play go differently from humans based on their own biological and cultural idiosyncracies, but I doubt it would on average be closer to how a Monte Carlo tree search running on transistors plays rather than in another direction entirely.
  - tarvaina 7y ago
    This is a good point. While Alpha Zero etc. are much stronger than humans, they may have weak areas. Actually we know one: ladders. Ladder is a fundamental technique that is easy for humans but hard for both monte carlo and neural network static analysis. It is plausible that an equally strong but more ladder-aware player would have very different tactics from the current best computer players. The original AlphaGo actually had a hand-crafted ladder feature input layer. It was removed in Alpha Zero. We know from Leela Zero and ELF Go game records that they don’t understand ladders very well, so likely Alpha Zero didn’t understand them either.
    - tarvaina 7y ago
      Actually if anyone here follows current go bot development, I would like to know if they currently have some specific ladder features to address this weakness?
- daseiner1 7y ago
  Are you implying alien societies wouldn’t (don’t?!) have their own historical baggage?
kentiko 7y ago
> For that matter, how did humans discover go? I like how he used the word "discovered" and not "invented".
- foobar_ 7y ago
  I always likened it to nature - https://www.youtube.com/watch?v=8wl8ZxAaB2E https://www.youtube.com/watch?v=8wl8ZxAaB2E Some of the rules neatly map onto it.
- sunstone 7y ago
  Some time ago I had this discussion with my math prof. Are new math things discovered or invented? We did not agree :)
H8crilA 7y ago
Funny that God is just a PSPACE oracle. At least in the sense that the oracle can answer essentially all of those questions.
- jofer 7y ago
  I think you didn't read far enough... (See aliens questions later on) That having been said, I think it's fascinating that most of these questions are just that, but are impossible to answer at present. The probability space is just too large. It's one of the fascinating aspects of go. It's a very simple set of mathematical rules that yields incredibly complex behavior and is practical impossible to brute force "perfect" gameplay as a result. I.e. you can't search the full space, you have to learn to predict optional strategies and always wonder if there isn't a better solution somewhere.
- somesortofsystm 7y ago
  Well, the premise of the exercise is that One can even Communicate to God, so this is two or three incongruencies away from being infinitely unsolvable. Atari!
- jerf 7y ago
  I don't think that's a characteristic of "God" here (which I mean in the sense that you are using it), I think it's the characteristic of asking about a game that "fits" into PSPACE.
Trufa 7y ago
Vaguely related, I still hold the hope that a perfect game of chess ends in black winning with the longest ever zugzwang! That would definitely would be among my questions!
pressurefree 7y ago
go is where asian nurses think th3y can turn out any american boy where theres no back row we are all equal pawns. Goh is like the humanities majors of board games. fuck u asians. asians aint shit
kburman 7y ago
Now i wonder will God know everything that universe has to offer. For example i wrote a 3 page article. Even would be surprised be surprised if a person came and tell me if we took all the characters at prime number it would be a valid sentence about an event going to happen in future.
Iv 7y ago
Unfortunately Lemy is dead
tzs 7y ago
I once tried to write a math/CS nerd's description of the rules of Go. There's an XHTML file of the results at https://pastebin.com/Cn7tbv98 https://pastebin.com/Cn7tbv98 if anyone is curious. Below is a text version: Here is an attempt to describe the rules of Go precisely. This is actually for a generalization of Go that I call MANGO, which stands for MAth Nerd Go Equipment To play Mango, you need the following: • A countable set, I. • A set, C, whose members are subsets of I, each of which contains exactly two members of I. • Three subsets of I, called E, B, and W, such that their union is I, and the intersections of any pair of them is empty. • A real-valued function F, whose domain is I. • A function, T, whose range is {0,1}, that is defined on the real numbers. Definitions • If P is a finite subset of C, and n is a member of I, then the INDEX of n in P is the number of elements of P that contain n. • A finite subset, P, of C is called a PATH if the following conditions are met: 1. Each element of I has an index in P of 0, 1, or 2. 2. There are exactly two elements of I whose index in P is 1. • An element of I whose index in a path is 1 is called an ENDPOINT of the path. • An element of I whose index in a path is 2 is called an INTERIOR point of the path. • Let S be one of the sets B or W. Let s be a member of S. Let L be the set of all members, e, of E, such that there is a path whose endpoints are s and e, and whose interior points are all in S or E. Let z be the sum over L of F. Then s is ALIVE if T(z) = 1. • The ordered triple (E,B,W) is called the CONFIGURATION. Playing The players must first obtain a Mango set. This consists of agreeing to the sets I, C, E, B, and W, and the functions F and T. The players must agree to an initial score for each player. The players than decide who shall have the first turn. Players alternate turns. On a players turn, that player may do one of two things: • The player may PASS. It then becomes the other players turn. • The player may make a LEGAL MOVE. Note that a player MUST either pass or play a legal move. If there is no legal move, the player is forced to pass. A MOVE consists of performing several actions. In the following, if it is Black's turn we will use the symbol M to refer to the set B and the symbol H to refer to the set set W. If it is White's turn, M will be W and H will be B. Here are the actions that are taken by a player on that players turn: • A member, n, of E is selected. • n is removed from E and added to M. • All members of H that are not alive at the end of the above step are removed from H and placed in E. • All members of M that are not alive at the end of the previous step are removed from M and placed in E. A move is a LEGAL MOVE if the configuration, (E,B,W), produced by the move is new. The game ends when two consecutive turns are passes. Scoring Each player uses the following procedure to compute his score. We will use the symbol M to refer to B if the player is Black, and to refer to W if the player is White. We use the symbol H to refer I-(E union M). The player starts with the initial score agreed upon at the start of the game. For each m in M, the player receives F(m) points. A player receives F(n) points for each member, n, of E for which the following conditions both hold: • There exists a path with n as one endpoint and the other endpoint in M, and which contains no members of H as interior points. • All paths that contain n as one endpoint and a member of H as the other endpoint contain a member of M as an interior point. The player with the most points wins. Example To play ordinary 19x19 Go, with a 5.5 point Komi, the players might agree to the following: • I = { (x,y) | x and y are integers in [1,19] } • C = { {(x,y),(u,v)} | (x,y) and (u,v) are in I, (x-u)^2+(y-v)^2 = 1 } • B = W = {} • E = I • The initial scores are 0 for Black, 5.5 for White. • It is Blacks turn. • F((x,y)) = 1 • T(z) = int((z+360)/361) To play a Go-like game on an infinite board, the players might agree to this: • I = { (x,y) | x and y are integers } • C = { {(x,y),(u,v)} | (x,y) and (u,v) are in I, (x-u)^2+(y-v)^2 = 1 } • B = W = {} • E = I • Initial scores are Black:0, White:0. • It is Black to move. • F((x,y)) = exp(-x^2-y^2) • T(z) = 1 if z > 1/1000, otherwise T(z) = 0
Izkata 7y ago
Can someone explain this Q/A to me? > Is there a safe way to prevent Mirror Go? > A really sad result of talking to God about go would be that there is a way to just play mirror go and win. If that was all god did, it would be really annoying. If mirror go doesn't actually work, how does God defeat it? Do the variants where W is just not allowed to mirror fix the issue? The first way I was aware of was that the primary player could surround the center. They'll capture the mirroring player's stones one turn before the mirroring player can surround theirs, and the mirroring player loses the stones they need to do a mirrored capture. Sensei's Library [0] has a few other ways as well. [0] https://senseis.xmp.net/?CounteringMirrorGo https://senseis.xmp.net/?CounteringMirrorGo
- bluecalm 7y ago
  Mirror go isn't considered a strong strategy in a game with komi. You don't need to do anything extraordinary either. Just play natural moves and at the end you win by komi.
tunesmith 7y ago
One question I never found an answer for was what AlphaGo thought komi should be. Since Black goes first, it has an advantage of x stones. But if Black is given a handicap of 2 stones (placed on the board before play), white then goes first. Since AlphaGo plays itself, it could have been set to play itself with a handicap, such that the win rate was 50%. I want to know what that handicap would have been, because wouldn't that indicate a pretty accurate measurement of what komi should be?
- bluecalm 7y ago
  I agree. There are many interesting questions Deep Mind could have answered about both go and chess. I am still bitter they didn't take a few days more to run experiments of great interest for fans of both games. For chess the simplest one would be to take most popular positions from top level human play and let A0 play Stockfish from that (say 10 times from every of 100 positions). Then do the same for self-play and then most importantly publish all the games not selected sample. They kinda did something similar but the positions they selected were chosen for entertainment value. We still don't know how Alpha Zero plays from positions that are actually interesting for human players.
  - monkeypizza 7y ago
    Open source research is continuing. There are a few open source implementations of the alphago algorithm and some new work being done to improve it for go. Check out lizzie, which is a nice front end to leela zero. https://lifein19x19.com/viewtopic.php?f=10&t=17095 https://lifein19x19.com/viewtopic.php?f=10&t=17095

111 points by ivee 7y ago

I don't know how to play Go, and I'm not very good at maths. How many of these questions could be answered without going to an omniscient heavenly being?

hyperpape 7y ago
Go is a finite game, so almost all these questions can be answered by just writing an algorithm that enumerates all the possible games and gives the answer. However, the game space is so big that approach would never finish. Further, I doubt any of them could practically be proven, even with some smarter algorithm. It's akin to how even with all the advances made in computer chess, we still don't know what perfect play looks like. I do think we already have good evidence the best komi is 7 or 7.5, but it's not definitive. The question of what the best ko rule is value based (different rules have different consequences/aesthetics for gameplay).
- IggleSniggle 7y ago
  From Wikipedia: Komidashi https://en.m.wikipedia.org/wiki/Komidashi https://en.m.wikipedia.org/wiki/Komidashi "Komi in the game of Go are points added to the score of the player with the white stones as compensation for playing second. Black's first move advantage is generally considered to equal somewhere between 5 and 7 points by the end of the game. Standard komi is 6.5 points under the Japanese and Korean rules; under Chinese, Ing and AGA rules standard komi is 7.5 points. Komi typically applies only to games where both players are evenly ranked. In the case of a one-rank difference, the stronger player will typically play with the white stones and players often agree on a simple 0.5 point komi to break a tie in favour of white."
praptak 7y ago
Most of them if we either 1. discover an amazing technology for blazing fast computing. or 2. advance our algorithms so that they can prune the search space of Go to something manageable by our CPUs while provably keeping correctness. Does not apply to questions about aliens.
CydeWeys 7y ago
Likely none of them; the search space is simply way too big. Heck, we can't even answer most of these questions for Chess, and Chess is a comparatively much easier to compute game.

phkahler 7y ago

The best placement for handicap stones seems most for perfect play. For tic-tac-toe I prefer to start in the middle because "not losing" takes more thought for the other player, but assuming perfect play it doesnt matter where you start - the game is a draw.

elcomet 7y ago

Do things like "best opening move", "worst handicap placement" or similar make sense if you assume perfect play ? I would think not, because when you have perfect play, you either win, lose or draw. So you just have "winning opening moves", "losing opening moves", and "draw opening moves". You cannot really rank them. Maybe the ranking would be the number of different games that lead you to a win, in a monte-carlo tree search fashion. Which could then be useful if you can't play perfectly (highest probability of winning).

afthonos 7y ago
Go has points. An opening move is better than another if, given perfect play thereafter, you end up with more points.
- elcomet 7y ago
  But are those points counted in tournaments, or are they just counted to know who won the game ? So, if you have a very clear path of how to win by 1 point, or a risky way to win by 10 points, does it make sense to try the risky path ?
  - afthonos 7y ago
    They are not officially counted, but they are often used as an informal indication of strength difference among players.

CydeWeys 7y ago

> Is go played anywhere else in the universe, and was it discovered by chance there? > What openings do aliens play? I love these questions about if anyone else in the universe is playing Go. My gut response would be yes -- it's a simple enough game (regular game grid of any size roughly 5x5 on up, one type of piece, two players take turns alternating placing pieces, and any that are surrounded are removed). I suspect other simple enough games that likely exist elsewhere in the universe would be tic-tac-toe, connect-N, dots and boxes, and Othello/Reversi -- any others? Contrast with, say, Chess, which has too many different types of pieces, all with specific movement rules, and with a specific initial layout. I wouldn't be surprised if Chess only existed here, though I also would expect plenty of games similar to Chess to exist out there in the universe, i.e. a game on a regular board with multiple types of pieces and different rules for moving the different pieces. If Shogi, Checkers, or Stratego weren't human games I wouldn't be surprised to learn them from aliens. It's a great question -- what is the most universal board game? What board game has been independently invented the most times throughout the universe? What if it's a game that we haven't even come up with?

SamPatt 7y ago
I imagine dice, or some other physical object that introduces randomness, is a sure thing if intelligent life exists elsewhere.
- GrumpyNl 7y ago
  That implies that they think and behave like us.
  - CydeWeys 7y ago
    Does it? A lot of concepts in board games are pretty natural results of universal mathematics. There's only three regular polygons that will tesselate a board (we have plenty of board games that use all three of them), and there's only three dimensions that games can work in (and we have plenty of games of all three types; Candyland, for instance, is played along a 1D path, and the first player to reach the end wins). Even if some aliens think so differently from us that the concept of a square is unknown to them, surely there are plenty of aliens that would think similarly enough?

mark_l_watson 7y ago

That was fun. My favorite two questions are: Is go played anywhere else in the universe, and was it discovered by chance there? What openings do aliens play? Our Universe is a huge place. I think it was Nature Magazine that had an article estimating 20,000 worlds with earth like conditions and life. My formal education is in physics and personally I subscribe to the multiverse hypothesis so with a potential infinite number of universes these questions have meaning. My older brother taught me to play Go when I was eight and I played actively in the 1970s. Two years ago, on a whim, I decided to go to the US Go Open. I drastically overestimated my own rating when registering, played too strong of players and lost every game. A good lesson in humility and I made some friends. To lick my battle wounds, I paid a South Korean Go professional for lessons for three months and I feel like that really helped my game.

treve 7y ago
20,000 seems a little low. I imagine this needs at least 10 more zeroes
- CydeWeys 7y ago
  Yup. As our telescopes get better we keep finding Earth-like planets just orbiting nearby stars. And that's only what's nearby, here in this one galaxy.
- mark_l_watson 7y ago
  I agree. My Dad is a physicist and in his late 90s still a member of the National Academy of Science. We riff a lot on science and technology and this is a favorite topic :-)
- marcoperaza 7y ago
  Any number is a bad guess because we don’t have any sense of the likelihood of life emerging, even in the “optimal conditions”. It could be so unlikely that it’s probable there is no other life, it could be so likely that almost every “viable” planet has life, anywhere in between, or it could be that God created exactly as much life as he wanted, which could be, again, any number. We have a sample size of one and what must be, in the grand scheme, a primitive understanding of physics and chemistry and biology. Throwing around numbers and probabilities for the existence of other life is frankly ridiculous. The best we can intelligently talk about is what the number is given some guess about other numbers. But remember, we have no good numbers for the emergence of life, the survival of life, the likelihood of evolution into intelligent life, etc. Any number multiplied by a totally unknown factor can be literally anything. And before someone throws the infinite universe argument at me, of course. Any probability above zero times infinity is infinity. (Though it’s also possible that there is still no extraterrestrial life even in that case, that outcome always has a non-zero probability!) If we accept a probabilistic emergence of life, then we need to talk about useful things like “in our solar system”, “in our galaxy”, “in the visible universe”, and such. Though it is theoretically possible that the conditions for the emergence of intelligent life are so narrow that only planets exactly like ours, so exactly like it that someone else on that planet has or will typed out this exact comment, have it. In a truly infinite, random (hmm, another bad assumption perhaps?) universe, there would be infinitely many such exact replicas of Earth and its entire visible universe. This isn’t a very useful concept.

mcbuilder 7y ago

Would not Alpha Go Zero and related algorithms give us a good insight into the question about what openings alien's might play? Isn't it really tackling the game from first (or zeroth if you will) principles, just simulated games without the historical baggage coming from human study. It has also considerably changed our own knowledge of the opening.

Sharlin 7y ago
I don’t think so. The go that AlphaGo plays is one very specific type of ”alien go” that exploits the strengths of the computing substrate it thinks on. Alien entities that are still evolved via natural selection would probably play go differently from humans based on their own biological and cultural idiosyncracies, but I doubt it would on average be closer to how a Monte Carlo tree search running on transistors plays rather than in another direction entirely.
- tarvaina 7y ago
  This is a good point. While Alpha Zero etc. are much stronger than humans, they may have weak areas. Actually we know one: ladders. Ladder is a fundamental technique that is easy for humans but hard for both monte carlo and neural network static analysis. It is plausible that an equally strong but more ladder-aware player would have very different tactics from the current best computer players. The original AlphaGo actually had a hand-crafted ladder feature input layer. It was removed in Alpha Zero. We know from Leela Zero and ELF Go game records that they don’t understand ladders very well, so likely Alpha Zero didn’t understand them either.
  - tarvaina 7y ago
    Actually if anyone here follows current go bot development, I would like to know if they currently have some specific ladder features to address this weakness?
daseiner1 7y ago
Are you implying alien societies wouldn’t (don’t?!) have their own historical baggage?

kentiko 7y ago

> For that matter, how did humans discover go? I like how he used the word "discovered" and not "invented".

foobar_ 7y ago
I always likened it to nature - https://www.youtube.com/watch?v=8wl8ZxAaB2E https://www.youtube.com/watch?v=8wl8ZxAaB2E Some of the rules neatly map onto it.
sunstone 7y ago
Some time ago I had this discussion with my math prof. Are new math things discovered or invented? We did not agree :)

H8crilA 7y ago

Funny that God is just a PSPACE oracle. At least in the sense that the oracle can answer essentially all of those questions.

jofer 7y ago
I think you didn't read far enough... (See aliens questions later on) That having been said, I think it's fascinating that most of these questions are just that, but are impossible to answer at present. The probability space is just too large. It's one of the fascinating aspects of go. It's a very simple set of mathematical rules that yields incredibly complex behavior and is practical impossible to brute force "perfect" gameplay as a result. I.e. you can't search the full space, you have to learn to predict optional strategies and always wonder if there isn't a better solution somewhere.
somesortofsystm 7y ago
Well, the premise of the exercise is that One can even Communicate to God, so this is two or three incongruencies away from being infinitely unsolvable. Atari!
jerf 7y ago
I don't think that's a characteristic of "God" here (which I mean in the sense that you are using it), I think it's the characteristic of asking about a game that "fits" into PSPACE.

Trufa 7y ago

Vaguely related, I still hold the hope that a perfect game of chess ends in black winning with the longest ever zugzwang! That would definitely would be among my questions!

pressurefree 7y ago

go is where asian nurses think th3y can turn out any american boy where theres no back row we are all equal pawns. Goh is like the humanities majors of board games. fuck u asians. asians aint shit

kburman 7y ago

Now i wonder will God know everything that universe has to offer. For example i wrote a 3 page article. Even would be surprised be surprised if a person came and tell me if we took all the characters at prime number it would be a valid sentence about an event going to happen in future.

Iv 7y ago

Unfortunately Lemy is dead

tzs 7y ago

I once tried to write a math/CS nerd's description of the rules of Go. There's an XHTML file of the results at https://pastebin.com/Cn7tbv98 https://pastebin.com/Cn7tbv98 if anyone is curious. Below is a text version: Here is an attempt to describe the rules of Go precisely. This is actually for a generalization of Go that I call MANGO, which stands for MAth Nerd Go Equipment To play Mango, you need the following: • A countable set, I. • A set, C, whose members are subsets of I, each of which contains exactly two members of I. • Three subsets of I, called E, B, and W, such that their union is I, and the intersections of any pair of them is empty. • A real-valued function F, whose domain is I. • A function, T, whose range is {0,1}, that is defined on the real numbers. Definitions • If P is a finite subset of C, and n is a member of I, then the INDEX of n in P is the number of elements of P that contain n. • A finite subset, P, of C is called a PATH if the following conditions are met: 1. Each element of I has an index in P of 0, 1, or 2. 2. There are exactly two elements of I whose index in P is 1. • An element of I whose index in a path is 1 is called an ENDPOINT of the path. • An element of I whose index in a path is 2 is called an INTERIOR point of the path. • Let S be one of the sets B or W. Let s be a member of S. Let L be the set of all members, e, of E, such that there is a path whose endpoints are s and e, and whose interior points are all in S or E. Let z be the sum over L of F. Then s is ALIVE if T(z) = 1. • The ordered triple (E,B,W) is called the CONFIGURATION. Playing The players must first obtain a Mango set. This consists of agreeing to the sets I, C, E, B, and W, and the functions F and T. The players must agree to an initial score for each player. The players than decide who shall have the first turn. Players alternate turns. On a players turn, that player may do one of two things: • The player may PASS. It then becomes the other players turn. • The player may make a LEGAL MOVE. Note that a player MUST either pass or play a legal move. If there is no legal move, the player is forced to pass. A MOVE consists of performing several actions. In the following, if it is Black's turn we will use the symbol M to refer to the set B and the symbol H to refer to the set set W. If it is White's turn, M will be W and H will be B. Here are the actions that are taken by a player on that players turn: • A member, n, of E is selected. • n is removed from E and added to M. • All members of H that are not alive at the end of the above step are removed from H and placed in E. • All members of M that are not alive at the end of the previous step are removed from M and placed in E. A move is a LEGAL MOVE if the configuration, (E,B,W), produced by the move is new. The game ends when two consecutive turns are passes. Scoring Each player uses the following procedure to compute his score. We will use the symbol M to refer to B if the player is Black, and to refer to W if the player is White. We use the symbol H to refer I-(E union M). The player starts with the initial score agreed upon at the start of the game. For each m in M, the player receives F(m) points. A player receives F(n) points for each member, n, of E for which the following conditions both hold: • There exists a path with n as one endpoint and the other endpoint in M, and which contains no members of H as interior points. • All paths that contain n as one endpoint and a member of H as the other endpoint contain a member of M as an interior point. The player with the most points wins. Example To play ordinary 19x19 Go, with a 5.5 point Komi, the players might agree to the following: • I = { (x,y) | x and y are integers in [1,19] } • C = { {(x,y),(u,v)} | (x,y) and (u,v) are in I, (x-u)^2+(y-v)^2 = 1 } • B = W = {} • E = I • The initial scores are 0 for Black, 5.5 for White. • It is Blacks turn. • F((x,y)) = 1 • T(z) = int((z+360)/361) To play a Go-like game on an infinite board, the players might agree to this: • I = { (x,y) | x and y are integers } • C = { {(x,y),(u,v)} | (x,y) and (u,v) are in I, (x-u)^2+(y-v)^2 = 1 } • B = W = {} • E = I • Initial scores are Black:0, White:0. • It is Black to move. • F((x,y)) = exp(-x^2-y^2) • T(z) = 1 if z > 1/1000, otherwise T(z) = 0

Izkata 7y ago

Can someone explain this Q/A to me? > Is there a safe way to prevent Mirror Go? > A really sad result of talking to God about go would be that there is a way to just play mirror go and win. If that was all god did, it would be really annoying. If mirror go doesn't actually work, how does God defeat it? Do the variants where W is just not allowed to mirror fix the issue? The first way I was aware of was that the primary player could surround the center. They'll capture the mirroring player's stones one turn before the mirroring player can surround theirs, and the mirroring player loses the stones they need to do a mirrored capture. Sensei's Library [0] has a few other ways as well. [0] https://senseis.xmp.net/?CounteringMirrorGo https://senseis.xmp.net/?CounteringMirrorGo

bluecalm 7y ago
Mirror go isn't considered a strong strategy in a game with komi. You don't need to do anything extraordinary either. Just play natural moves and at the end you win by komi.

tunesmith 7y ago

One question I never found an answer for was what AlphaGo thought komi should be. Since Black goes first, it has an advantage of x stones. But if Black is given a handicap of 2 stones (placed on the board before play), white then goes first. Since AlphaGo plays itself, it could have been set to play itself with a handicap, such that the win rate was 50%. I want to know what that handicap would have been, because wouldn't that indicate a pretty accurate measurement of what komi should be?

bluecalm 7y ago
I agree. There are many interesting questions Deep Mind could have answered about both go and chess. I am still bitter they didn't take a few days more to run experiments of great interest for fans of both games. For chess the simplest one would be to take most popular positions from top level human play and let A0 play Stockfish from that (say 10 times from every of 100 positions). Then do the same for self-play and then most importantly publish all the games not selected sample. They kinda did something similar but the positions they selected were chosen for entertainment value. We still don't know how Alpha Zero plays from positions that are actually interesting for human players.
- monkeypizza 7y ago
  Open source research is continuing. There are a few open source implementations of the alphago algorithm and some new work being done to improve it for go. Check out lizzie, which is a nice front end to leela zero. https://lifein19x19.com/viewtopic.php?f=10&t=17095 https://lifein19x19.com/viewtopic.php?f=10&t=17095