Japanese Baseball: How Good Is It?

Japanese baseball performance should, in theory, be as translatable as
performance from any baseball league in the United States. The process has
had its challenges, though: the data is not as easy to find, and much of
what is available is in a language and a character set that I can’t read. (I
still don’t have complete data for 1996).

More serious a problem is the small number of players moving between Japan
and the United States. The Translations system depends on being able to set
a difficulty level for each league. To do that, I need to have a sizable
group of players who have played in both the leagues I am testing and in
leagues whose difficulty level I already know. Every player who played in
both leagues needs to be compared to the league average; if, as a group, one
set is league average, and the second set is 10% above average, you can
assume that the second league is 10% worse that the first league.

With the Japanese leagues, there really haven’t been enough comparisons to
get a firm grip on the appropriate difficulty level, especially since almost
all the comparisons were of players who went from the U.S. to Japan, and not
from Japan to the U.S..

Last year, for the first time, there were more common plate appearances from
the prior year moving from Japan to the U.S. than vice versa. (A common
plate appearance is the lesser of a player’s plate appearances in League 1
and in League 2; it is what I use to ensure that a given player is always
weighted equally.) Alex Ramirez was the only player who went to Japan
last year to log more than 100 CPA, while five players–Ichiro
Suzuki, Tsuyoshi Shinjo, Orlando Merced, Tony
Fernandez, and Lou Merloni–came to the U.S. and reached that
standard.

Using one-year differences, here’s how the difficulty ratings for Japan
shape up:

You can do the same analysis with Triple-A players. In fact, it is better to
use Triple-A players: the Japanese leagues have generally taken players who
were stuck in the minors, guys who tend to end up in Triple-A again even if
they come back. There have been far more common PA between Japan and
Triple-A than there have been between Japan and the majors.

In Baseball Prospectus 2001, I wrote that the difficulty level of
Japan was "about even with the Triple-A leagues." Looking at it
more comprehensively–I was basing my assessment on a scattering of players,
rather than off a full list of Japanese player data–that was a silly thing
to say, as the Japanese leagues have clearly and consistently rated as
tougher than the American Triple-A leagues. Every case from the 1990s shows
that players do worse as a CPA-weighted-average group in Japan than they do
in Triple-A, and by a considerable margin. The Triple-A/majors multiplier is
.860; if the transitive property holds, then Japanese EqA is worth about
.948 of a major-league EqA, which conveniently enough is almost identical to
what we got from major leaguers.

You can, of course, do the same thing with pitchers. The total ERA ratios
for those come out as:

Triple-A weighted average:       17991 IP     1.154
Major league weighted average     7178 IP      .903

Remember, though, this is in runs, and runs are proportional to EqA to the
2.5 power. These ratios, in EqA terms, are the same as 1.058 and .960. The
1.058 between Triple-A and Japan is equivalent to a .910 ratio between Japan
and the majors. So we have, all in EqA and major-league terms:

Hitting Triple-A to Japan         .948
Hitting Major Leagues to Japan    .946
Pitching Triple-A to Japan        .910
Pitching Major Leagues to Japan   .960

The mean of these values is .941.