Does Changing Leagues Affect Hitters?
by Kirk Allen
October 10, 2000
updated November 18, 2002
Everyone knows that Junior Griffey was traded from the Mariners to the Reds before the
2000 season. He proceeded to have a slightly down year by his standards. Some will argue that
he suffered from changing leagues, as in "not knowing the pitchers." Does such an effect really
exist?
To measure this, I compared a player's OPS (on-base% + slug%) the year before and after he
changed leagues. I used several qualifications for the player. First, the
player must be at least a semi-regular in both seasons under consideration. I used a
minimum of 300 at bats. Second, the player must be in his first change of leagues. This means if
a player went from AL to NL and then back to AL, only the first change would count. Third, I
tried to avoid players who were close to the beginning or end of their careers, in order to
avoid the effects of aging or maturing. There wasn't a specific age range, but I looked at
the seasons surrounding the change to decide if the player was noticably maturing or aging.
However, I found that the at-bat limitation usually made age considerations obsolete.
Finally, the player must have been with one team for each of the 2 seasons, which means
mid-season trades do not count. I made a couple exceptions for players who were traded very
early in a season, usually at around 20 games. I avoided players from the 1800's
and early 1900's because the leagues may not have been equal and statistics are less reliable.
Through the fine data available at Baseball
Reference.com, I was able to find 121 players who fit my qualifications. I also included data for OPS+,
which accounts for park effects. The list is not
intended to be exhaustive, as I am sure I missed some. That said, 121 players is a fairly
hefty sample size. To view the complete data file,
Click Here!
My results confirmed my suspicion.. Changing leagues does not make a player have a down year.
In fact, of the 121 players I investigated, there was an average increase in OPS of 6 points,
and a 1.8% increase in OPS+. Now, I'm not naive enough to say that this proves
that players actually benefit from changing leagues. Assuming 500 at-bats, an increase of 6 in OPS
means that the player had, for example, one extra double after changing leagues.
ONE HIT! Does this mean anything? Not really.
I definately feel I proved that changing leagues does not make a player have a down year.
My next hunch is that, perhaps, recent players benefit from advanced scouting more than players
of yesteryear. Maybe this trend will be noticed if you break the data down by decade or era.
I found that the appropriate breakdown was 1900-1959 (12 players), 1960's (17), 1970's (24),
1980's (32), and 1990's-2000 (36). The results..
Decade/Era cOPS cOPS+
1900-1959 -5.75 -0.25
1960's -19.59 -2.47
1970's +70.44 +15.60
1980's -40.47 -7.59
1990-2000 +19.42 +3.51
cOPS is average change in OPS, and cOPS+ is average change in OPS+
Is there a trend? I sure don't see one. My only conclusion is that changing leagues does not
have an effect in either direction, and it never has. Any advantage from scouting and increased media
coverage should be balanced because hitters and pitchers all have access to information.
I also broke it down by players who had a large change (over 50 points in either direction) and
those who had a small change (0-50 points) in OPS. This turned out again to be
slightly on the positive
side, with 26 players having small increases and 38 with large increases. On the negative,
24 were small and 33 were large.
My only possible explanation for the slight positive tilt
is that players were allowed to leave their team or were traded after they had down years. With
this in mind, perhaps it would be more meaningful to use 3-year OPS averages before the trade
instead of just one year. However, this would further limit the number of players due to age considerations.
Like I said, the increase is basically meaningless, and I
just think I somehow managed to miss some players who had decreases that would draw the average
down to a natural point of zero. There is a cumulative increase in OPS of 745, which would
take only 4 or 5 players having bad years to bring back to zero.
During follow-up research, I was able to determine a possible cause: Age. The peak age
for a player usually falls somewhere around 26 to 30. You expect younger players to improve,
peaking players to stay about the same or improve, and older players to decline.
Here are the numbers:
Age cOPS cOPS+
24 to 26 12.9 2.6
27 17.6 7.0
28 70.9 17.6
29 27.3 4.3
30 30.2 9.1
31 -5.9 0.8
32 -52.8 -11.3
33 to 34 -13.8 -4
35 to 37 -68.2 -18.3
Age refers to age in year after change.
The three age groupings are due
to a lack of players at those ages.
The correlations on the graph clearly show that age is a major factor in the change in OPS.
For young players, maturing accounts for 64% of the change in OPS (that's what the R2 means),
while aging accounts for
80% of the decline in OPS for the older group. If you look at OPS+, the relationship is even
stronger, with young and old both at 83%.
I realize that some people may considered this study flawed because it does not take a
whole season for a hitter to learn the pitchers. However, even two poor months due to "learning
the pitchers" would bring down the player's numbers for the whole season. In this way, you'd
expect the overall numbers to show a slight decrease in productivity, which is clearly
not the case. I know, I know. "But hey, Griffey took a couple months to get going." Yeah, he
did, but one player doesn't prove anything. To you I'd say, "But what about Jim Edmonds?" He
started out as the likely MVP, but then faded in the second half. It all balances out in
the end.
An obvious follow-up study would be to find pitchers who had changed leagues and see how
their ERA changed. Under the conventional logic, the pitchers would have success in the new league
since the hitters are not familiar with them. However, given this research, I'd
expect pitchers to show no meaningful change in their ERA.
copyright 2000 & 2002, Kirk Allen