Format of rikishi data files
The file allrslt.dat (about 500K) contains one line per entry.
This file is also available in compressed form as
allrslt.zip or as allrslt.dat.gz (about 90K each).
The first line looks like:
Akebono:Akinoshima:30:1:67:2
Fields:
1. Rikishi shikona
2. Opponent shikona - this is sometimes "Unknown<n>" or
"Makushita<n>", where n is an integer to make the name unique
for a given basho. Unfortunately, for basho prior to 1995 Kyushu,
in some cases I only knew whether a juryo rikishi won or lost on a
given day. In a few cases, I did not know the name of a makushita
rikishi who fought in juryo.
3. Basho number. This starts at 00 for Hatsu 90, is 01 for Haru 90,
etc. There is a file allbasho.dat
for which you can use this number as an index. Results for some
basho before 1996 are absent or partial.
4. Result. 1 for win, -1 for loss.
5. Kimarite. Index into Kimarite.dat
This is occasionally 0 (Unknown) because of a typo in the
original input.
6. Day (1-15) of match. Ketteisen matches are not in the database.
Notes:
In addition to the Unknown and Makushita entries described above,
some juryo matches from 1995 Hatsu, Nagoya and Aki basho are totally
missing. If you have detailed results for these or earlier basho,
please email them to me and they will be put in.
Most matches appear twice - e.g. Akebono beats Takanohana and
Takanohana loses to Akebono. But a makushita rikishi, "Makushita<n>",
or "Unknown<n>" does not appear in field 1.
The first line of rankwins.dat is:
Akebono:30:NY:12:3
Fields:
1. Shikona
2. Basho number (as above)
3. Rank in this basho, e.g. NY = nishi yokozuna, HS2 = higashi
sekiwake (second row), NM12 = nishi maegashira 12.
4. Number of wins (except ketteisen).
5. Number of losses (except ketteisen).
A rikishi on the official banzuke who did not fight is listed
as 0-0.