![]() StartingPoints Paths Referenced by
|
Handicap
Path: HandicapGoPath · Prev: · Next: HandicapForSmallerBoardSizes
Keywords: Go term
If two players differ in strength (see rank), the weaker player gets a handicap to compensate for the difference. That way, both players have a chance to win. In general, the ideal handicap is equal to the difference in kyu or dan ranks. Traditionally, handicap stones are placed on the star points, but one can also play with free placement of the handicap stones.
The traditional placement of handicap stones, from Black's viewpoint, is (see diagram):
Bill Spight: The traditional scheme gives an advantage to White. In an even game Black gives komi. If Black is 1 stone weaker, the proper handicap is for Black to put 2 stones and give komi. Alternatively, Black could play first and White could give komi, since komi is worth 1/2 stone. Simply having Black play first is inadequate compensation for the difference in strength. Also, the equivalence between one stone and one rank is inexact. Is the fifth handicap stone worth the same as the fourth? However, through 9 stones the deviation appears to be quite small. This could be the topic of a good statistical study. :-) Jasonred Gotta agree with this one. Not only that, but the size of the board has gotta make a difference too. IMHO, the strength of the players also makes a difference, as even 3 stones on a 19x19 is not much help to a 30 kyu, but is a HUGE difference to a 9 Dan... Especially if you're following French handicap, or whatever you call the "free placement of handicap stones" thing. Instead of statistics though, I would ask a top Dan pro how much komi he'd be willing to give an equally strong opponent for a 1 stone handicap, for a 2 stone, and so forth. I DO know that the 26th stone isn't worth as much as, say, the 5th stone... and on a 9x9, anything over 5 is probably a liability. More than 10 stones is just silly on a 9x9, you're just wasting your own territory, so I can safely say that the 10th stone should have a NEGATIVE value. Dieter Verhofstadt: Let me get this straight: A and B are of equal strength. A takes Black and B White. B gets 5,5 points of komi. Second game: B takes Black and A White. A gets 5,5 points of komi. We expect the same result. This means that the difference between having the first move or not is equivalent to 11 points. Removing komi is only worth 5.5 points. The next step we add a stone to Black's handicap. If one stone is worth 11 points, then this is a bigger leap than the first one, and so White consistently receives 5.5 points more than he deserves. If one stone is worth 5.5 points, then this system is OK. So, Bill, your assumptions are:
? Bill: Each handicap stone (up to 9) has approximately the same value. Playing White (sans komi) is equivalent to having a handicap of 0 stones (i.e., of playing first but passing on the first play). Receiving (proper) komi makes the game even. So (proper) komi is worth approximately half a stone. About 25 years ago I did some statistics on pro-pro handicap games and concluded that each handicap stone was worth around 13.5 points (territory scoring). (Of course, the estimate was not very precise. I was surprised to find, however, that the relationship was nearly linear. My initial assumption had been that it was not.) Since that implied a proper komi of 6.75, I predicted that the Japanese komi would increase to 6.5 by the turn of the century. Wrong again! ;-) (Nearly: In October 2002, the Nihon Ki-in announced it would start using 6.5 komi). Somebody else published an article in the AGA journal (Summer, '77, I think) based on over 2,000 Japanese pro games, that indicated a proper komi of around 7. In the late '70s in New Mexico I ran 4 tournaments a year (around 20 players showed up for each tournament). We divided our ranks into two half ranks. If there was a 1/2 rank difference, the lower ranked player took Black (White winning jigo); if there was a 1 rank difference, he took Black and White gave komi; etc. Komi was 6.5. After a few years game statistics indicated that White had a slight advantage. However, most games were decided by such large margins that a small change in the komi would not have mattered. Besides, in our small pool of players, the better players (who were more often White) tended to advance more than the weaker players. They were probably more interested in learning the game. :-) Before the adoption of komi, a different handicap system was in use among top players. See: tagai-sen, sen-ai-sen and josen. In my experience various Go clubs in Tokyo use a similar point ranking scheme to track players' results and determine the handicap in games. Dave Sigaty Jasonred: The correct answer is, at a 25 stone difference, the weaker player should be allowed to do whatever he likes... or perhaps should be playing on a smaller board. To be honest, I don't see how a 25 stone handicap makes any sense. I'm probably 20 kyu or something, and I don't see a 5 Dan player giving me a 25 stone and winning... especially in the arrangement you just showed, I have no idea how ANY stone my opponent puts down isn't dead... Confused: I advise you to find a 5-Dan player willing to play such a game and give it a try. The last high handicap game I played was a very humbling experience for me (I'm also around 20 kyu strength). unkx80: It might be possible for a 5-Dan to win a 20-Kyu at a 25 stone handicap, although IMHO the game would be rather meaningless (either the Black player plays way too weakly or the Black player blunders horribly). However, a good exercise for single digit kyu players is also a 25 stone handicap, but free placement, and the objective is not to allow the stronger White player to have any living group. That is, if White manages to make any living group, White wins the game. Some years ago when I attended Go classes the instructor Yang Jinhua suggested such a game during one particular lesson. It was a very interesting experience. See also:
Path: HandicapGoPath · Prev: · Next: HandicapForSmallerBoardSizes This is a copy of the living page "Handicap" at Sensei's Library. ![]() |