Some Thoughts On Handicapping

British Go Journal No. 12. August 1970. Page 6.

Francis Roads, Secretary and Founder of Enfield Go Club

There is no doubt that the handicapping system contributes much to the popularity and spreading of the knowledge of Go. A Chess Tournament with 60 players of such varied strengths as those of the Go players who attended the recent Congress would hold little interest for anybody. Yet the undoubted convenience and elegance of our handicapping system often blinds us to the fact that it contains several anomalies.

The most obvious one is that handicap games appear to have a built-in advantage for White. The addition of one handicap stone is supposed to represent an advantage to Black of around 10 points. However, the advantage of the opening move is about 5 points, so that, when the players strength differ by one grade, Blacks advantage is about 5 points; by two grades 15 points, and so on. More logically, one might have expected the advantage to be equivalent to 10,20 and so on. White appears to have a 5 point advantage, though perhaps this is deliberate, so that when Black wins regularly he knows he is due for promotion.

However, it is difficult to talk in terms of so many points advantage for one player. I believe British players tend to think too much about how many points they win by. After all, a 25 point victory may well mean that you are a weaker player than one who would have beaten the same opponent by 10 points, if you took an unnecessary risk to win 15 extra points.

Japanese professional players of equal strengths playing nine stone handicap games have shown that this handicap has a points value of about 140, rather than the 90 one would expect if the handicap stones were in fact worth 10 points each. This points to the rather obvious fact that the differences in value between successive handicaps are far from equal. Many players experienced in giving large handicaps would agree that the difference between four and five stones is less than that between five and six; likewise, that between seven and eight stones is less than that between eight and nine. Clearly the value of each additional handicap stone must be affected by its position and relationship to those stones already placed.

However, even if it could be established that the added advantage of each additional handicap stone were equal, there is still no particular reason to think that what we call "playing strength" is a quantity capable of addition and subtraction in a simple linear way. If player A is evenly matched with player B giving four stones, and player B is matched with player C giving four stones, what a priori reason is there for thinking that player A will be matched with player C giving eight stones? Clearly there is scope for some club to carry out carefully controlled experiments to see what extent our assumption of the additive nature of playing strengths is justified.

Another anomaly of the system is that a players strength may vary according to the size of the handicap. An obvious example is a player familiar with handicap joseki, but not even game joseki. Such a player may well have a weaker strength when playing low handicap or even games, whether with a weaker or stronger player, than when playing with a large handicap. It would be worth while experimenting with arranging the stones differently in high handicap games so as to give the Black player practice in even-game joseki; e.g. in a six stone game the three handicap points on one side might be left vacant, the stones being placed on the six remaining (handicap) points. It would also be worth experimenting with placing the handicap stones on points other than the nine conventional star points, although there are in fact good reasons why the conventional points were chosen.

The only conclusion to be drawn from these thoughts is that convenient though our handicapping system is, it is only a rough and ready one, and not so precise as we are apt to think. It is worth pointing out that the most reliable promotions in playing strength are those based on low handicap and even games with stronger players. It would be interesting to hear from anyone who has tried the experiments I have suggested or who has other ideas on the subject of handicapping.


This article is from the British Go Journal Issue 12
which is one of a series of back issues now available on the web.

Last updated Thu May 04 2017. If you have any comments, please email the webmaster on web-master AT britgo DOT org.