Yotsu preference statistics reprise

[Rich Magahiz <MAGAHIZ@ernest.phys.cmu.edu>]

Since I mensioned this subject in a note a few weeks back,
I have corrected and extended the statistics to measure "yotsu" 
preference.  For my purposes, yotsu kimarite are those which
emphasize gripping and grappling the body, as opposed to
tsuki-oshi techniques (pushing with one or both hands), trips,
bends, etc.

See my other note for the way in which I have divided up the
70 kimarite into yotsu, tsukioshi, and "other".  For each of the
wrestlers in this note, the detailed bout-by-bout information is
available on the Web:
<URL://www.contrib.andrew.cmu.edu/usr/rm6s/ythat95.html> (Hatsu)
<URL://www.contrib.andrew.cmu.edu/usr/rm6s/ytnat95.html> (Natsu)
<URL://www.contrib.andrew.cmu.edu/usr/rm6s/ythar95.html> (Haru)
<URL://www.contrib.andrew.cmu.edu/usr/rm6s/ytnag95.html> (Nagoya)
and
<URL://www.contrib.andrew.cmu.edu/usr/rm6s/yt95.html> (summary tables)

(Contact me if you would like to obtain them via email instead).

I apologize in advance for the semi-random choice of sumotori,
if I have left out some of your favorites (or included your non-faves).
I tried to get most of the recent san'yaku rikishi plus a few farther
down the banzuke.  All these numbers were generated with a minimum
of technological aids (mainly spreadsheets and Emacs), so I wouldn't
guarrantee that they are free of computational errors.  I hope I can come
up with something better in time for Aki-basho.

Thanks to Ian Frank, I have gone away from a single "yotsu preference"
number, in favor of a three-part scheme.  Two issues are important:
(A) the fraction of the time a rikishi spends using yotsu techniques,
regardless of the success or failure of such, and (B) the relative
effectiveness of yotsu techniques vs. tsukioshi techniques.  

To measure (A), I compute the percentage of bouts
with a yotsu outcome, or (YotsuWin+YotsuLoss)/total*100.
Scores between 50 and 100 are high, between 0 and 50 are low.
 
To measure (B), I compute the percentage of Yotsu wins minus the percentage of
Tsukioshi wins, or ((YotsuWin/total)-(TsukioshiWin/total))*100.
Positive numbers are high, while negative numbers are low.

To allow comparison with my previous note, I also include another measure
of (B), the "yotsu effectiveness" (YotsuWin-YotsuLoss)-(TsukioshiWin-TsukioshiLoss).
Positive numbers are high, while negative numbers are low.

For the first four bashos of 1995 (only rikishi for whom I have
records for all these bashos):

Rikishi        Yotsu          Tsukioshi     %Yotsu  %Y-%T   Y-T
             Wins  Losses   Wins   Losses    bouts   wins   eff.
Akebono     13        4       37        6     28      -40   -22                    
Akinoshima  20        9       14        8     57       12     5
Hamanoshima 19       25        7        8     75       20    -5
Kaio        24       15       10       11     65       23    10
Kenko       24       16        7       12     68       29    13
Konishiki   17       17        9       15     59       14     6
Kotonowaka  26       22        3        7     83       40     8
Musashimaru 13        5       35        6     31      -37   -21
Takanohana  41        5       12        2     77       48    26
Takanonami  29       16        6        9     75       38    16
Takatoriki   1       10       25       22     19      -41   -12
Terao        8       14       16       21     37      -14    -1
Wakanohana  21        7       23        9     47       -3     0

Wakanohana looks like the most "even-handed" of
all, with nearly identical records in yotsu and tsukioshi overall. 

Here is one interpretation of the extreme cases.

(A)     (B)
HIGH   HIGH   Rikishi who prefer yotsu and are capable of imposing their
                preference on others (offensively by gaining morozashi &
                armlocks, nage; defensively by closing in on opponent)
                E.g., Takanohana, Takanonami
LOW    LOW    Rikishi who prefer tsukioshi and are capable of imposing their
                preference on others (offensively by tsuppari & nodowazeme;
                pressing for quick advantage, etc.).  E.g., Akebono, Musashimaru

Rikishi with middling numbers can either be interpreted as possessing equal
talent in both areas (Wakanohana, Akinoshima) or as having a non-specialist approach
(Kaio, Kenko).  There are a few who have a preference in one direction despite
only so-so records using that technique:  Takatoriki and Terao favoring tsukioshi,
Kotonowaka, Konishiki, and Takanonami favoring yotsu.  Can someone find a common
factor accounting for these?

Musoyama is an interesting case: he sat out most of Hatsu and Haru with his
shoulder injury, then for the last two bashos he has a line that looks like this:

Musoyama     7        7       14        2     47      -23   -12

He did a little better in yotsu-technique in Natsu than in Nagoya, but was
much better in tsukioshi for both.

So I hope this will stimulate discussion during the time we wait for the
next round of action!  What are your thoughts and suggestions?

Regards,

Rich Magahiz
-- 
Richard A. Magahiz                  Physics Department
                                    Carnegie Mellon University
magahiz@ernest.phys.cmu.edu         5000 Forbes Ave.
rm6s+@andrew.cmu.edu                Pittsburgh, PA 15213-3890
http://www.contrib.andrew.cmu.edu/usr/rm6s/homepage.html