Life In 19x19

I rather like the attached image, from the latest go manga hit (sold out already, I'm told) Hoshizora no Karasu, of the fellow old codger telling little Waka: "This is MY territory - and this is mine, too". (Trembling, she asks who decides it's his.)

Not long after reading that I was going through some old papers and saw a note that relates to territory in Go Seigen's games, which I thought you might find useful.

In all (or a sample - can't remember) of his games the average territory per game was about 116 vacant points plus 12 dead stones plus 20 captures, making the average final territory per player around 74 points.

I'm sure these figures will be similar for most players, and they provide a useful yardstick. You can suggest your own uses, but obviously they tell us that if you have the usual maximum of five groups and you have allowed one of them to be bullied into living with just two points, you had better make your other groups work much harder than usual. More specifically, it follows that mo matter how much of a clever general you may think you are if you invade and live with just two points, the lieutenants in charge of your other platoons are not likely to be impressed.

On the topic of manga, you may wish to support the Nihon Ki-in's own new venture, Hibi Goseki. It's for the older types, perhaps (maybe even the old codgers who say "washi" as in Hoshizora), but it does give, as the title implies, a passably accurate, if idealised, idea of the life and range of characters in an everyday Japanese go club. It's a cheap reprint of the monthly episodes from Gekkan Go World, and it also has a short but nice essay by the creator, Kasa Taro, in which he says he made 3-kyu at 30, 1-dan at 32 and 6-dan at 35 - so there's hope for us all.

John Fairbairn wrote:I was going through some old papers and saw a note that relates to territory in Go Seigen's games, which I thought you might find useful.

In all (or a sample - can't remember) of his games the average territory per game was about 116 vacant points plus 12 dead stones plus 20 captures, making the average final territory per player around 74 points.

Maybe you are getting the 74 pts. this way? (116 + 12 + 20)/2 = 74.

If so, there are a couple of other ways of looking at those numbers. One is how many points of territory are there on average when it is counted? The dead stones are exchanged without changing the number of vacant points. Subtract 20 for the captures and you get an average of (116 - 20)/2 = 48.

Another way is how many points are on the board after taking off the dead stones? The average per player is (116 + 12)/2 = 64.

I remember a lecture at a US Go Congress many years ago by the Japanese pro Izumitani Masanori in which he did some reasoning similar to John's above. He did it this way: An average game that is not resigned lasts around 220 moves. That leaves 141 points for territory and half of that is around 70 points. His conclusion was that if you see youself as having around 70 points of territory you don't have to make much more to win so you don't have to take any big risks or even push hard to build your territory.

Bill Spight wrote:

John Fairbairn wrote:I was going through some old papers and saw a note that relates to territory in Go Seigen's games, which I thought you might find useful.

In all (or a sample - can't remember) of his games the average territory per game was about 116 vacant points plus 12 dead stones plus 20 captures, making the average final territory per player around 74 points.

Maybe you are getting the 74 pts. this way? (116 + 12 + 20)/2 = 74.

If so, there are a couple of other ways of looking at those numbers. One is how many points of territory are there on average when it is counted? The dead stones are exchanged without changing the number of vacant points. Subtract 20 for the captures and you get an average of (116 - 20)/2 = 48.

Another way is how many points are on the board after taking off the dead stones? The average per player is (116 + 12)/2 = 64.

For over the board games, when I count in-game, I count my territory, with two points for dead stones, and adding on prisoners in my bowl. In a close game this is usually around 70-80 points. At the end of game, we count the opponents score, and use prisoners to fill in the opponent's territory. Here, in close games, the score is usually somewhere around 50.

I have often wondered and asked about the expected number of points, groups, or lengths of plays games have, I don't think I have ever seen any statistics in those regards, overall it seems people have the attitude that such statistics are irrelevant because there are so few games that are counted.

Overall, I think this is highly stylistic though, some players tend to play tight and compact with many invasions. Others have a large amount of amashi, both of which would lead to lower over all scores.