First Move in the Corner
I would like to present some analysis of the first moves played in the corners based on the games present on the GoGoD CD. I used Kombilo to perform a search for empty quadrangles across all the games included on the January 2002 version of the CD. Games that start with 4+ handicap stones on the board will not turn up any hits for the search pattern since there are no empty corners at the start of the game.
Kombilo returned 51,747 hits from 13,362 games (unfortunately there is a slight problem with the number of hits since there is some double counting involved at the edges, but that is as good as it gets for now :-)
- 3-4 point: 24,996 = 48.3% of the total
- B at a: 15,146 = 55.6% of Black's plays
- W at a: 9,850 = 40.6% of White's plays
- 4-4 point: 21,103 = 40.8% of the total
- B at n: 10,096 = 37.0% of Black's plays
- W at n: 11,007 = 45.4% of White's plays
- 5-3 point: 2,223 = 4.3% of the total
- B at b: 1,127 = 4.1% of Black's plays
- W at b: 1,096 = 4.3% of White's plays
- 3-3 point: 1,926 = 3.7% of the total
- B at i: 345 = 1.3% of Black's plays
- W at i: 1,581 = 6.5% of White's plays
- 5-4 point: 1,054 = 2.0% of the total
- B at c: 370 = 1.4% of Black's plays
- W at c: 684 = 2.8% of White's plays
(SAS: Do you have the corresponding figures for the 5-5 point? DaveSigaty: I skipped all other choices for this analysis since there are far fewer examples, 0.9% of the total or less than 500 altogether. In the case of the 5-5 point there are only about 25 cases - good for study but meaningless for trends I think.)
- Total shown for five plays 51,302 = 99.1% of total hits reported
- Total shown by Black 27,084 = 52.8% of total shown
- Total shown by White 24,218 = 47.2% of total shown
I have the following thoughts about these figures (all of them off the top of my head):
- The preponderance of Black's plays over White's plays in the empty corner reflects the fact that in various fuseki White foregoes the opportunity to play in the empty corner in order to prevent Black from making a shimari. This was particularly true in classical Go before komi was used. The Shusaku fuseki is a good example of this.
- The preponderance of komoku as the first move by Black in the empty corner reflects the long history of the move plus the fact that it was not replaced by hoshi as Black's first move choice until the period 1975-1977 (see FusekiTrends - First Move Analysis).
- The preponderance of hoshi as the first move by White in the empty corner reflects its development already in the 19th century as a means of blunting Black's advantage of the first move.
- The 5 to 1 ratio between the use of san-san by White and by Black reflects White's attempts to slow down the pace of the game when receiving komi.
Kombilo reports the winning percentage for the results that it finds. The winning percentages from the empty corner analysis let us ask whether Move One Loses the Game