Re: Finding the peak

["G. Jay Walker" <walker@nhrc.navy.mil>]

(more geek stuff)

>Mr. Walker's analysis seems pretty solid, but I'd like to see a break down by highest achieved rank.  That is:  What was the peak age for all those achieving Yokozuna status?  How about ozekis, the rest of sanyaku, upper and lower maegashira?  There might be some interesting differences there.

There were only 73 rikishi in the study (all those born between 1960 and 1969 who reached makunouchi), so I hesitate to divide them into too many groups as the numbers in some groups could get so small that the results could be misleading.  For example, there were only 5 yokozunas and 2 ozekis among the 73 rikishi.

However, I did try dividing them into two broad groups.  The "upper rankers" were defined as all those who spent at least two bashos in sanyaku (sorry, those guys who have their "basho of a lifetime" and then go 3-12 at komusubi don't make the grade).  The upper rankers numbered 27 of the 73 rikishi.  The other 46 were defined as the "lower rankers".

The original post describing the methodology can be found at:

http://www.banzuke.com/03-1/msg00058.html

Basically, a rank/result system is used that awards points to the top 40 rikishi every year.  A yearly rank in the top 5 is worth 8 points, a rank between 36th and 40th place is worth 1 point.  The chart below shows the age, points earned by the upper rankers at that age and points earned by the lower rankers.

20   16   6
21   47   1
22   76   8
23  100  23
24  110  43
25  136  55
26  158  64
27  150  75
28  139  50
29  118  41
30   79  28
31   60  15
32   43   5
33   35   6
34   18   5
35   11   6
36    8   3
37    1   2 

It's no surprise that the upper rankers record more points, but what we're interested in here is comparing the peaks.  The upper rankers show a peak age of 26, with age 27 ranked second.  The  lower rankers show a peak age of 27, with age 26 ranked second.  Perhaps you could conclude that the lower rankers peak a little later, but there's certainly nothing that would indicate any substantial difference.

If you graph and plot the two set of numbers (curve plot), the lower rankers show a significantly steeper rise and sharper decline than the upper rankers.  However, I tend to think much of that is an artifact of the methodology.  This could involve a fairly detailed discussion, but basically put, the method only awards points to those in the top 40 in any given year, which means that those fighting at the Juryo-equivalent level don't receive any points.  If the method ranked the top 65 or top 70 instead of just the top 40, I think the shape of these curve plots for the two groups would look more similar.

So in short, there doesn't seem to be major differences in the career peaks of upper rankers versus lower rankers, at least as broadly defined groups.  In both cases, we could expect a typical rikishi to peak around ages 26 or 27.

-George W.