THE BOOK--Playing The Percentages In Baseball

I posted the following elsewhere, and repeat it unedited.

=======================================
Here is some useful data:
http://www.insidethebook.com/ee/index.php/site/article/switch_hitters/

If you click the link in there, you’ll get my file on Google Docs with the complete handedness data by year.

We see in the very early years of baseball, when no one knows about platoon advantages, the % of LHP is around the 10%, pretty much mirroring society at large.  My guess as well is that these days in little league, 10% of pitchers are LHP.

The % of LHH however was around 25% in the early years.  So, unlike the likely extremely strong relationship between writing and throwing, we find likely just a strong relationship between writing and hitting.

As the platoon advantage became apparent over the years, and in a bid to stop those LHH, more LHP came on board.

But, since there was still so many more RHP (80% RHP), there was still an advantage for bringing in LHH, to the point that in 1900, 50% of at bats were by LHH!

For hitters, it has been pretty much reached as 40% of PA are by LHH, and 60% by RHH.  For pitchers, there is still some bouncing around, but it’s pretty much around 25-30% of pitchers are LHP.

So, in both cases, hitters and pitchers, there is more LH than the population at large.  If we treat the early years as the general society of today, then 10% of people are LHP, but 25% of MLB are LHP.  And 25% of people are LHH, but 40% of MLB are LHH.

What’s also interesting is the rise of the switch hitters, since the early 80s, likely as the pitching staff increased in size, thereby making the importance of a switch hitter greater, for flexibility.

Nice data, Tom. Can you break it a bit further though? Here are some off-the-cuff ideas:

--by league
--by team (and maybe a STDEV of the team avgs. per year; in the 60s the Dodgers had a switch-hitting infield, and that probably shows a huge skew for that time frame, which may have pushed the SH phenomenon into existence)
--by age or age range (the ones I always use are -25, 26-29, 30-34, 35+, but you can do it any way you think makes sense)

What about percent of IPs by LH/RH pitchers? Is there any variance there with respect to simple population count? Are/were there any more lefty starters than lefty relievers, which would account for a higher IP% than pop%?

Question: why are SH PAs leveling off now? Decline in relative quality of the hitters? Increasing platoon diffs even with the expected “embedded platoon advantage”? Or have some teams simply abandoned it (another reason for computing those team-by-team avgs).

Anyway, just curious. If you crank any of this out sometime, please drop me a line and let me know.

Best,
DM

Don, I’ll do something along these lines tomorrow, and publish on google docs.

I’ll also post the necessary SQL, so people can make whatever changes they like.

Handedness file on Google Docs

The first sheet “Original” is exactly from the original blog entry.

The second sheet “Regulars” breaks up the data based on whether the player was a true hitter or pitcher.  Regular=1 means that the player’s primary position was nonpitcher for the SH,LH,RH columns, and that the player’s primary position was pitcher for the LP, RP column.  That is, the Ichiros, Everetts et al’s data counts in the hitter columns and Clemens, Oswalt et al’s data counts in the pitcher columns.  If you go to the bottom half, Regular=0 means you’ll see Wade Boggs’ pitching appearances count in the pitcher columns, and Clemens’ hitting appearances count in the hitter columns.

The third sheet “Age” is a further breakdown of the “Regulars” sheet, this time including an Age class.  There are 3 age classes: 26 and under, 27-30, 31 and over.  Each age class corresponds to roughly 30%-35% of the population.

The headers “knownPA” and “unknownPA” refers to the PA of known hands.  So, if I don’t know what the guy’s batting hand was, it gets added to “unknownPA”.

At the bottom of the “Age” sheet are records labelled “Age Unk”, and those are all the records for players of an unknown age.

PA = AB+BB
BFP = IP*3+H+BB

The explosion of switch hitters happened in the 1977 season.

If we look at the 26-and-under data, we see that through 1962, switch hitters accounted for 0 to 5% of all hitters in this age group.  Then from 1963-1976, it jumped to the 5-10% range.  Since 1977, it’s been in the 15-25% range.

What is interesting is when we also look at the 31-and-over players.  The 15% range started in 1987, which makes sense, since all these 26-and-under in 1977 would reach the 31-and-over age sometime between 1982 and 1988.  However, around 1977, there was also a jump in switch hitters among the 31-and-over crowd.

It’s possible that the old-timer switch hitters had more staying power than their equals who were one-sided hitters.

***

The number of LHP innings thrown, by age class, does not really change.

***

Of players born since 1930, 29% of innings were thrown by LHP (real pitchers).  Of those batters who threw as pitchers, 30% of those innings were thrown by LH throwers.

Of players born since 1930, 21% of the pitchers’ at bats were as LHH, and 5% as switch hitters (compared to 31% and 14% for nonpitchers).

***

In the early years of baseball (1877 and earlier), 21% of batters were LHH and only 2% were switch hitters.  Among pitchers as batters, only 1.5% of them bat left-handed.

In 1885 and earlier, only 8% of innings were thrown by LHP. In these days, they didn’t know about the platoon advantage, and therefore, this 8% figure probably mirrors society as a whole.  I would guess in Little League, 8% of pitchers are lefties.