Flowing Fuseki - Introduction
Inspired by some of the things on Flowing Data, I have been exploring the use of the information in GoGoD to visually represent the ebb and flow of different fuseki/joseki ideas over the last three hundred and fifty years, give or take. Starting with this post, I expect to present a few of the trends that can be seen.
First a couple of caveats. The Christmas 2011 update of GoGoD has 65,000+ even games, which form the base population that I used. This is not very many spread across such a long period of time and, indeed, we have very few games from various earlier periods. So the pictures presented are really impressions at best. Nevertheless, I hope everyone can enjoy speculating on what the players of the past were interested in and why their choice of fuseki changed over time.
(Larger version)
The older even games in GoGoD are almost all Japanese. The database contains quite a lot of historical Chinese games played with the traditional cross-hoshi initial position, but I have filtered those out. Only the last twenty years really contain a broad mixture of games from China, Korea, and Japan. This is natural given the development of professional Go, but it means that some themes are potentially reflections of the change in the source of games.
That First Basic Choice by Black
(Larger Image) Here we can easily see the early dominance of 3-4 and the rise of 4-4 in the New Fuseki era. More interesting, because less understood (at least by me) was the resurgence of 3-4 in the 60's and that it was really the 70's that made B1 on 4-4 the dominant theme in Go. BTW, the later rise and fall of 3-4 around 2006 is driven by the alternate form of the mini-Chinese with the marked stone below on 3-4 rather than 'a'.
White Responds:
To B1 on 3-4
(Larger Image) Note that naming all the variations on this graph required me to invent some new nomenclature. However, I haven't written down definitions of them all yet so that has to come in a later post. The problem is that I can't even remember them all myself so I have to re-research my data! RSN...trust me.
To B1 on 4-4
(Larger Image)
When we try to go beyond this point we are confronted with the fact that Go is played by two opponents. Therefore if we wish to look at the trends of say Black's fuseki choices, we have to be aware that as often as not White is deliberately playing to frustrate Black's plans. It becomes somewhat of a challenge to decide which games fall within a single theme, for example, "mini-Chinese fuseki". There are the various flavors chosen by Black: mini, micro, upper right 4-4, upper right 3-4, etc. There are also various counter-plays by White. Which fit into the theme and which do not? Naturally I will present my own ideas.
In future posts I will try to illustrate or at least explain what goes into each and then everyone can disagree as they like.As a last picture for this post, here is the graph of nirensei and sanrensei by Black. The most interesting thing to me is that despite the famous role of these in the history of modern Go and the great popularity among amateurs, sanrensei especially has never been a major fuseki choice by the pros!
(Larger Image)
Worst of all is that Excel does not properly (IMHO) protect the colors and transparency settings in the graphs. Thus, if I have to adjust the colors, I often end up going through them one by one manually changing each of them back to what they were the last time I did it. Since some of the graphs show up to 15 queries, this can get extremely frustrating! 
to counter Black's attempt to initiate the mini-Chinese, Black's earliest reaction was to enter the 3-3 point at "a". However, results were not so good and Black quickly shifted to making a shimari in the lower right. Although the small shimari at "b" has been most popular, the large shimari at 
has been second. I speculate that this caused a renewed interest in the large shimari in other lines. This variation has all but disappeared recently as the pincer at 
