Hi all,
It's been a long time since I posted. I moved and, while I haven't unpacked everything, I'm afraid I may have lost my GoGoD disks (and the hard drive that had the SGFs, too!).
I couldn't find a repository only that had all of Dosaku's games. I found one with 67 or so of them, but I know GoGoD has ~150.
Do you guys/girls, by chance, know of a place where these can be downloaded?

[edit: of course, within minutes of posting this, I find -something- ... This site, http://sites.google.com/site/byheartgo/ , has 146 games. Can anyone do better? ]

We have 153, but that also includes all the recent new discoveries with different and/or extra moves.

The Dosaku book of the Chinese series of collected Honinbo games has 153 too, but those were published a couple of years ago, so between the book and GoGoD, there are probably a few more. Tchan001 has a picture with info up, so I didn't need to get my camera out.

http://tchan001.wordpress.com/2010/08/2 ... ollection/

Hmm, I just checked my database, and between GoGoD (Winter 2010), SmartGo and go4go, I seem to have 178 games (without obvious duplicates).

The first thing to look at would be dates. We use the correct Gregorian dates, giving the lunar dates only for reference. Other sources may give the lunar date misrepresented as a Gregorian date. This applies to Edo games in general. A Kombilo search would be useful.

SmartGo filters duplicates based on the final position of the board, I believe.

Yes, but we often find extra or variant moves for games. I think we are the only people to maintain our database in this way. This is a very frequent occurrence and T Mark has discussed it with Anders, I believe. From memory, I think it amounts to several hundred differences across databases.

There are also differences in final position where an internet version is used (i.e. with dame filled and sometimes extraneous passes added) and a paper version is used.

Kombilo uses the Dyer method of checking moves at several points early in the game and reporting that matches "possibly" represent duplicates. My experience is that this is much more reliable.