There has been go reseatrch all the way to sixties. Progress has judt been very slow. And success really requred amount of computing power that would have been impractical even 1997.
Let alone the very basic idea of how to do chess program really did not change during that 40 years. While in most avenues had to be abandoned since they were dead ends
Go is far easier for algorithmical compression than chess and prior to 1997, nobody gave a shit about computer Go.
It took 40 years (1957 - 1997) and best IT scientist's effort to beat human champion in chess.
It took only 19 years (1997 - 2016) and amateurish research to beat human Go champion.
For a minimum truth, let's say: It took only 19 years (1997 - 2016) and amateurish research on top of 40 years (1957 - 1997) and best IT scientist's effort to beat human Go champion
OK, let's be more precise.
First complete chess program appeared in 1941, and human champion lost in 1997.*
First complete Go program appeared in 1968, and human champion lost in 2016.**
*Kasparov was declining already, underestimating artificial opponent, and there may have been a few GMs more competent (Anand, Shirov, Karpov).
**Very few of elite programmers were ever focused on Go.
It's more like - money for nothing vs nothing for money -.
Now, that's the right information.
Many avenues had to be abandoned since they weren't interested in actual AI
Yeah, chess programs weren't much stronger than the best humans until years later. There were some matches drawn in the mid 2000s.
and to put into context. Hardware needed to beat Lee Sedol
The version of AlphaGo that played against Lee used a similar amount of computing power as in the match against Fan Hui, where it used 1,202 CPUs and 176 GPUs.
Long long time anyone can get pro-level opponent onthe home computer
Go and chess cannot be described as "harder" whichever you think it might be. I play both games. I am only average as a casual chess player, and I am even worse at Go. But I do understand how computers calculate moves in both games and the mathematical difference between both of them. Go is more complicated for computers but recently the creators of alpha go have bridged that gap. Despite that computational difference and the fact that Go may have many more "possible positions" than chess it doesn't actually make chess easier and go harder for people to play.
The difference in all of the possibilities of piece arrangement on the board may be more in Go, but for human skill that difference is insignificant when it comes to mastering one game or the other. If anyone says Go is "better because it has more possible positions! So it must mean or requires greater brain power than chess!" Simply doesn't understand either of the two games, or how the mind works. If you take the skill of a bad Go player and skill of bad chess player and assign something like 1 to both, and a skill for some future super computer which HAS fully solved both games as one million then the difference between the skill of computers that solved both games will be like a 1 million (perfect knowledge in which every move is absolutely correct and perfect) and perhaps 1 million + 100 000. The skill of Go and Chess players will still be infinitely small compared to perfect knowledge of either game like maybe value of 100 for each, human champion of Go and human champion of chess.
i realize that above may be poorly worded. I couldn't describe it better without using complicated math terms. I hope if anyone reads that more or less understands what I'm mean. Basically the point is that for humans there is no significant difference for complexity of these games and skill at one game cannot really be compared to another. But we know that on average all human brains are of about the same potential. Go is more complex for computers but it's irrelevant for people.
I heard that the AlphaGo technology will next try to play starcraft at the pro level. That will be interesting. I don't know that many would describe starcraft as harder, but certainly the brute force computation approach would be even less effective than it in go.
I am probably wrong. Machine that would be perhaps little above average would match a pro on short time limits (about 30 sec/move) which is rate most people play casual games. On longer thinking time some special HW is needed.
I play both Chess and Go. For several years I did not play chess at all. I returned to chess because there are many more opportunitues to play OTB chess than OTB Go.
Nature of chess and Go compared
It is not true to say that they are completely different games. Both are two player games of strategy. Both are played on a finite board, and the players make discrete moves alternately. One difference I found is that a modest difference in playing strength produces a win by the stronger player more often and more certainly than in chess.
The thought processes during play are also very similar, but with different emphasis. Tactics are less important than in chess and in most cases easier to calculate, because the stones don't move around, but intuitive feel for position matters more.
The objectives of the games are different, but the things to study to become strong fall into much the same categories in each game.
To begin with there are standard tactical patterns. In Go these are called tesuji and they deal with things like capturing stones, making a group survive, making or breaking a connection between stones and so on.
Go also has its own endgame theory. The main difference here is that in chess the endgame is when the board is realtively empty, whereas in go it is about sealing up the boundaries between opposing groups of stones in the best way, so as to gain as many points as possible.
Strategy, Planning, Positional concepts
Just like chess Go has positional concepts such as "thickness", "high" or "low" positions, efficiency of the stones, good and bad shape, "Aji" (the potential of stones to cause trouble in future) and many more. There are also standard ways (in both the strategic sense and the tactical) of attacking and defending. A major difference is that in Go there can be several largely independent battles going on in different parts of the board. As the board fills up they often have more and more influence on each other.
Go has masses of opening theory. Against a strong player you have little hope of surviving unless you know what you are doing at the start of the game. Unlike Chess, where the centre is important the early stones are usually played close to the corners. THe idea is that to enclose territory in a corner you need enclose only two sides, whereas onthe edge it takes three and in the centre four.
Because Go takes place on a larger scale than Chess it has two different types of opening theory. Firstly there is large scale strategy. How are the early stones played with regard to the overall game. There are standard patterns such as playing on the 4,4, points of two adjacent corners adn on the point midway between with the first four stones (this is san-ren-sei - adopting a high position on the 4th line. It does not guarantee a lot of definite territory, but often leads to powerful stones that can be useful in attsacks on the opponent's groups).
The really detailed theory is about play in the corners. There are many standard patterns called joseki. Like chess openings they are precise. Most have subvariations. Some (e.g. the Taisha joskei) have many and complex sub-variations, just like the Najdorf or Botvinnik Semi-Slav in chess, and are just as difficult to memorise. It is not good Go to play your favourite joseki regardless of the position. The real skill is to choose joseki that work with the whole board position.
Similarities and Differences
Go is bigger in every way. Bigger board, more pieces, game last longer (typically 200 to 300 moves) and there exist more possible positions. But it makes little difference to human players as both games are beyond absolute mastery and offer enough interest and challenge to last a lifetime.
C. H. O'Donnel Alexander made the interesting point that if some alien civilisation has a strategy board game it is unlikely to resemble chess, but it might very well be similar to Go.
Move compulsion (zugzwang) is an integral and important part of chess. Without it there would be many more draws and chess would be a lot less interesting. In Go you are not compelled to move. If you see no advantage in playing a stone you may pass. If your opponent then plays you may answer or pass again. The game ends when both players pass.
One of the good things about chess is that when checkmate is delivered there can be no argument but that the game is over. It is a slight problem in Go that the territory has to be counted. To add to the difficulty there are different ways of counting territory, the Chinese and the Japanese. This is not a problem between professional players that know how to count, and rarely matters in amateur games where the margin of victory is often large. It is only a problem when an amatuer game is cose to within a handful of points. However you can be relatively strong, and still make mistakes in the counting.
Complexity of Rules
The basic rules of Go are very simple. It does not have such ad-hoc features as the double first move of a pawn, the ensuing possibility of en passant and castling. They make chess more interesting and more playable, but they have no connection to the basic logic of the game.
However Go is not entirely free of ad-hoc fixes. There is a situation called Ko where teh same two positions could be created ad infinitum by alternate captures. That would be something like a perpetual check in chess. The problem is that it is very easy to create a Ko so most games would end in a non-conclusive to-ing and fro-ing. To prevent this there is an artificial rule - when your opponent captures in a Ko you are obliged to play elsewhere before you can capture. This partly fixes the problem and it creates interesting play. It is often the case that if you cannot recapture in a ko you will take a heavy loss, so you make a threat elsewhere, the opponent answers (we hope!) and then the Ko can be retaken. Skilful players always seem to have an inexhaustible supply of such "ko threats". The rule of Ko does not completely fix the problem of a game getting stuck in repeated positions. If there are several Kos they can be played in a repeating sequence.
Advantage of moving first
There is widespread (but not universal) agreement in Chess that playing first is an advantage. In Go there is little argument that playing the first stone is advantageous, so long as it is played half-sensibly (e.g. not on a 1,1 point). But in Go this is neutralised by giving the second player, (and in Go that is white) a few points start. The start he gets is called "komi". Komi is often fixed at 4.5 or 5.5 points. The trouble with that is that there is no agreement about just how much it is worth to go first. To get around this some tournaments have an "Auction" for komi. The players state how many points deficit they'd be prepared to accept to have komi. The player that bids highest gets Black, and the other player gets that many points start in their score.
The problem of Draws
A criticism of chess is that the margin of draw is too wide and that too many games are drawn. In a straight game of Go a draw is possible, but unlikely. You win or lose by a definite margin. However, when komi is used a draw is impossible because komi always includes an odd half point.
In chess it is close to impossible to get a result against a very much stronger player, and playing at odds doers not help. Playing at odds of material in Chess distorts the nature of the game, and playing at time odds does not greatly affect the stronger player as he sees good moves instantly, as well as being able to play in your time. In Go it is possible to place up to 9 stones on pre-defined "handicap points" and this gives the weaker player genuine chances against a much stronger opponent. It does distort the game, but not so ruinously as in chess.
Ratings and Rankings
In chess most national federations use the Elo system of rating, and it is a pretty good way of comparing the relative strengths of contemporary players. The Elo system is also used in Go, but there is an older system of Dan and Kyu rankings similar to that used in Martial Arts. Amagteur Dan rankings go from 1 Dan to 7 Dan. 1 Dan is like a black belt in Go! difference of 1 corresponds to being one stone stronger or weaker. There also kyu rankings that start at 1 kyu, which is one stone weaker than 1 Dan and goes down as far as you like, though ranking lists generally do not go further than 25 kyu. There is a different system for professionals that, confusingly, goes from 1 Dan to 9 Dan. But in professional grades the difference between ranks is 1/3 of a stone and a 1 Dan professional is at least as strong as a 6 Dan amateur. Only the very strongest amateurs are stronger than the lowest of professionals.
Whereas Elo grades are calculated from competitive results professional Dan grades are awarded by the game's official regulating organisations, whioch looks not ony at results but also at quality of play. They are honorary titles for life, rather than an indication of current playing strength. Amatuer rankings are less controlled. You are expected to honestly figure out your own standing and state what you truly beieve your strength to be. In Europe, where Elo ratings are used, there is strong social pressure to be honest. You'd look a fool if you claimed to be a shodan (1-Dan) when your Elo rating suggested something closer to 10 kyu! [When I started playing Go I was immediately about 12-13 kyu simply because I played chess and could calculate sequences of moves. With about 2 years of study and practice I reached 3 or 4 kyu. I do not know if I could have improved further as that is when I returned to chess.]
ComputersIt was once widely thought that Go is superior to Chess because when computers could already play master strength chess GO programs were still beatable by weak amateurs. That was always sloppy thinking. Chess programming soon produced strong play because chess can be played well by calculating tactics deeply and making quite crude positional asessments. It lent itself well to a relatively straightforward approach and the architecture of most computers. This has changed. Go programs now play well enough to destroy most Go players as surely as strong chess programs beat everyone but strong masters. I am sorry to say that I know nothing about how computers have affected the way Go players study and prepare, but I suspect it will be a lot less than it has affected chess.
Go is played professionally in Japan, China and Korea. Each has a distinct national style of play. The best established professional system is in Japan , but for many years it is Korea that has been the strongest nation. Go in Japan is organised much more professionally than chess in any Western country. Top players are famous and can become very rich. Tournaments and big games are followed by large numbers and reported in most newspapers and on television. There are Go schools where promising children are trained intensively from a young age ... as young as five. This is more like the way that Soccer clubs develop young players than the way that chess skills are nurtured. It is rare for an amateur player to join the ranks of professionals. It is because of the intense training from childhood that most professionals are in a different class from most amateurs.
The expected behaviour of players is more formal than anything in Western Chess. This is in keeping twith the greater formality of social interactions in countries of the Far East. Naturally the players are required to shoiw mutual respect, but expectations extend as far as the way the stones are placed! There is a technique of playing the stones that makes a satisfying "thunk" that isd almost obligatory. Furthermore it is considered reprehensible to win a game by making ugly moves. You are meant to play beautifully and elegantly. [As a chess player this strikes me as barmy. If a move is strong then I reckon that thinking it to be ugly is a consequence of a faulty system of aesthetics]
FIDE and others have been trying hard to attract more spectators (and hence more money) to chess, but it is difficult when you have to be a tremendously strong player to understand high level chess. I believe that Go fares rather better. With very little knowledge you can appreciate the patterns that develop on the board and get an appreciation of how much territory the stones control. You can also see and follow raging battles for survival. A weak player or non-player might not be able to figure out the tactical details, but they will still have a better and truer impression of the state of a Go game than a similarly equipped spectator would make of a GM chess game.
I hope you've found this brief comparison helpful and/or interesting, and that it will encourage at least a few chess players to learn the rules of Go and give the game a try.
Sorry for any remaining typos
One of the many reasons why Go is less accessible and less popular with people is that it doesn't actually have a universal algebraic notation to record games! Yes, there are diagrams. Those can work on paper, but they are inferior to having a normal notation. Another problem is that no one has trasnlated and insisted upon using English terms for the Japanese terms used to study and play the game. Things like Joseki, Tesuji, kifu, moyo, etc are awkward and often difficult to remember for someone who isn't a speaker of Japanese. It is unfortunate because for every term in Japanese there is an appropriate word in English. For example Chess has its own terms for pieces, for game stages, and all terminology in the respective language of the players. Some terms from German have remained in other languages because they're very specific. But its easy for everyone because European and Middle Eastern languages all have unique, or similar terms that are specific to that language. I do not understand why this isn't the same way with Go. The South Koreans actually spend some government money to fund and popularize Go (which Koreans call Baduk) in other countries of the world. But for some reason they're not very successful so far.
I don't think there is a big correlation between popularity and whether the terms are in English. In chess there are 2 terms I had problem with as a beginner and that is "j'adoube" and "en passant". Those are terms you need to use when learning the rules. In go there is less rules you need to learn in order to play the game. "ko" and "komi" is probably the only terms you really need. I probably would not even bother with those terms at first when I teach go and start with capture go.
It's only when you get into studying the game other terms become more relevant, such as zugzwang in chess. There are a lot of terms in go, but since they're not essential to play the game I don't think they should matter a lot to the new player experience and therefore not have a big impact on popularity. I could be wrong on this though, since I haven't really studied the subject about what makes a game popular.
Also there are some English terms in go such as "eye", "knights move", "large knights move", "jump", "monkey-jump", "ladder" and "ladder breaker".
Ah, somebody knows Go exists.
Another difference I was thinking about today is that in chess, you can be completely winning for 99% of the game... and then late in the endgame, with 1 move, instantly lose. It seems there is much less room for error in chess (also the games are shorter of course). Or at least, the ability to lose a win game is much higher.
One of the many reasons why Go is less accessible and less popular with people is that it doesn't actually have a universal algebraic notation to record games! Yes, there are diagrams. Those can work on paper, but they are inferior to having a normal notation.
There is another side to this. Quality games in diagram form are a great learning tool. Why? Well for those thjat don't know the diagram will show the final position, or a position from an intersting stage of the game, with the stones numbered from 1 (or tha last move of the previous diagram) to the most recent move.
In diagrams with many stones you can hunt forages for the next move! In the process you begin to develop an intuition for roughly where to look, and this feeds into your playing skill. Just by playing over professional games from diagrams you can develop a good feeling for the right region to ;play to keep your stones efficient, and for good and bad "shape". By itself that is not enough to play good Go, but it is one of the things you need.
0110 ... ??? makes the excellent point that when you have a big lead in Go you can make quite big blunders and still win, whereas in chess there is often a fragile balance. In chess you can play 50 or 60 excellent moves and have an objectively won position, yet throw it all away with one slip. And to make chess even more tense the bad move does not have to be a gross blunder. It can be a move that superficially strong and could only be considered at all by someone that is strong!
It is as if a single game of Go is equivalent to a short chess match of several games. The stronger player might lose a game or two in a chess match but will still prevail over the match as a whole. In Go the stronger player might miss a few thikngs her and there but will prevail over the board as awhole. Whether this makes Go more attractive than chess or less (little chance of scoring a surprise win over a higher rated opponent) is a moot point.
Kudos to the posters from #348 to #355. It's nice to see a thoughtful and reality based discussion for once!
And #356 too!
One of the many reasons why Go is less accessible and less popular with people is that it doesn't actually have a universal algebraic notation to record games!
The South Koreans actually spend some government money to fund and popularize Go (which Koreans call Baduk) in other countries of the world. But for some reason they're not very successful so far.
Diagrams work better for Go because board is large. I have no problem following them. And if you follow games from database it is still easier.
Yes korean gov has spent some money. Buts lets face it: makin board games popular is verry difficult. And since Go is complex it is not most populat game even in those countries. Takes who lotta training not to Noob. Well true for about any game, but go is very knowledge burdened.
Most popular board game Japan is Shogi i.e. Japanese Chess. World most popular board game is Xian-qi aka chinese Chess. Dunno which is more popular in Korea chinese or korean chess. But has the one or another. They are very similar.
But Go has gained a lot in past two decades. And considering size addressable market and existence of competition, I do not think it will much more popular in west than it is
I'd say the difference of feeling between the two is chess is more like coordinating an assassination team and go is more like fighting a large scale war because it's not about any target but about dominance.
What I find interesting in go or even shogi, is their flexibility : you're not bound to one board size and you can use handicaps. Having handicaps in chess would be too hardcore, a piece lost is quite crippling and you won't get it back.
As for the depth of the difficulty, it's all about who is the opponent. I don't see an inherent difficulty in those games.