Originally raised on chess, I had heard that Go was older and harder than chess. I had been curious about Go for many years, and when the opportunity presented itself, I bought a CD-ROM (Mind Games entertainment pack) that included a computer Go opponent. I played and played, soon spending nearly all my free time playing the program. After what I estimate to be somewhere between 50 to 100 games, I had taught myself to beat the computer opponent on it's top level. By that time I was hooked, and having tired of finding the way to crushing the computer by the most points, I logged on to the No Name Go Server (NNGS). Since even cheap chess programs could play close to master level, I figured go programs would play the game at a similar level, and I must be fairly good at go could beat the computer on it's highest setting.
I got a major ego adjustment that day.
It turns out that all computers are horrible at playing go. When confronted with real opponents, I quickly discovered how poor the computer opponent had been! I lost repeatedly even with very large handicaps. I am now able to give the 9 stone maximum handicap to players slightly stronger than I was when I first logged on. Sounds like I might be a top level player? NOT EVEN CLOSE!
To give you an idea where I was, my first on-line ranking was 17kyu, which is a very good start, but still in the beginner ranks. The following table will give you an idea what the ranks are and where I was. All ranks describe relate to AGA ranks, though ratings in the AGA go as high as 9d, players with an 8d or 9d rating, are still refered to as a strong 7d for traditional reasons. The distinction is that a rank, is more akin to a title, and a rating is a statistical means of generating hanidicaps. In Asia, ranks are awarded (complete with a Diploma), a rank is yours to keep for life, a rating may fall or rise according to your tournament results.
Also note that the American Go association only gives amateur ratings. There is no such thing as an American or European professional, though a few western players have been admitted into the asian professional associations. There simply are not enough professional level players in the west for an western pro association to have much meaning (yet).
| Rank Category | Rank Grade | Comment |
|---|---|---|
| Professional Dan | 9p | Reserved for the top professionals, who win multiple international tournaments, and many national tournaments |
| 8p | ||
| 7p | Major international tournaments are usually won by players 7p and above | |
| 6p | ||
| 5p | Consdiered a strong professional player | |
| 4p | ||
| 3p | Successful professional, usually must win at least one national tournament, Reliably wins against almost all amateurs | |
| 2p | ||
| 1p | Permitted to enter professional tournaments, some top Amateurs may be stronger, especially if the pro has not been competing in tournaments for several years | |
| Amateur Dan | 7d | Top amateur. Some rating systems (especially on line) may denote higher levels, but traditionally this is the higest amateur level, games decided by 1 or 2 points are not uncommon. A few of the strongest 7d players can play on even terms with weak professional players. A tiny handful have actually beat stronger pros in open tournament play (but generally, the strong pro would still win most of the time, particularly if pro tournament time limits were used) |
| 6d | Very Strong Amateur, can occasionally give a good game to a professional, but will almost always loose, games between 6d players are often decided by less than 5 points | |
| 5d | ||
| 4d | A Strong Dan player, games are often decided by less than 10 points | |
| 3d | ||
| 2d | ||
| 1d | Dan Level players are generally considered strong amateurs, Most Dan level games are decided by less than 20 points | |
| Kyu | 1k | |
| 2k | My current AGA rank | |
| 3k | Maybe 50% of games decided by under 20 points | |
| 4k | ||
| 5k | Strong intermediate player, komi points begin to matter. May be almost Dan strength in some areas of the game but lacking in others. | |
| 6k | ||
| 7k | ||
| 8k | ||
| 9k | Single digit kyu are generally considered intermediate players, and have a general grasp of all phases of the game. | |
| 10k-15k | Advanced Beginner, knows many tesuji/joseki, occasionally selects joseki correctly, plays moderately reasonable opening moves, can kill difficult shapes, knows many end-game tesuji and tries to sequence them (with variable success) Can estimate score at mid game within 20-30 points, Looses big groups over 30points in less than 50% of games | |
| 16k-20k | Experienced Beginner, has begun to think about strategic concepts like thickness and sabaki can kill all the basic dead shapes easily, can sometimes use ko to intentionally gain advantage, but more likely is still scared of ko fights, advancement begins to require study of books, or with stronger players. | |
| 25k-20k | Beginner, understands live/dead groups, ladders, nets, may know a couple joseki, knows what sente and gote are | |
| 30k-25k | Novice, Just learned the rules |
In amateur play games are usually handicaped by giving the weaker player extra stones at the start. The number of handicap stones is the difference in the player's ranks (with a maximum of 9 stones). This method of handicapping is one of the wonderful benefits of the game's simple design. Even players who are very different in ability can play with an approximately even chance of either player winning. This is very good because Go has more levels of play than any other game I know of. Go is one of those games where, only very inexperienced players ever claim to fully understand the game.
By nature of it's use of a square board and black and white pieces, Go is often compared with chess. The games aren't very similar, but one of the things that Go players like to remind chess players of is that the very best Go playing software run on the best available hardware can usually be beaten easily by a talented 12 year old. (I have actually seen just such a youngster play and win at a tournament, against adult opponents who are rated much stronger than the best Go playing programs). Truly, mean Go players follow up by asking if Kasparov would like to learn to play Go.
The lack of good Go software is NOT for lack of people trying. For many years the Ing Foundation offered a 1.6 million dollar prize for any program that could beat a strong amateur player. This inspired scores of attempts at writing programs that build Go, and regular computer vs computer Go tournaments. Eventually, after no one managed to come close enough to even try to claim the prize, the foundation decided that the money could be better spent elsewhere.