The Protection Mini-FAQ V1.0
David Marasco
with thanks to Sean Lahman and David Grabiner
Comments/corrections to marasco@nwu.edu
Intro: From time to time the subject of "protection" rears its head
on rec.sport.baseball. URLs to studies are posted and DejaNews is
suggested, but in order to keep things all in one place I've decided to
present this mini-FAQ on the field. The point isn't to squash discussion,
rec.sport.baseball in general encourages arguments and ideas, but we
want to avoid seeing the same thing over and over again. This document
exists so that new concepts come to the table.
What's Inside:
Q1 - What is Protection?
Q2 - Is strong protection real?
Q3 - What about the STATS study?
Q4 - But when I played in highschool I know I saw protection,
how can it be false?
---------------------------------------------------------------------------
Q1 - What is protection?
A1 - Protection is the idea that a good player will benefit when another
good hitter is batting behind him. This has been passed down for
many years, the reasoning being that the first pitcher will "see better
pitches" because of the next hitter. At this point it is very important to
distinguish between two different kinds of protection, and in general once
this distinction has been made it really clears the air:
Weak Protection - A good batter behind you will prevent you from seeing
four balls in the dirt (or just a plain old intentional walk). In other
words, you are being protected from "intentional-unintentional" walks.
We can examine this by looking at walk rates. When people use the example
of "they walk the #8 hitter in the NL because the pitcher can't hit" they
are providing evidence for weak protection.
Strong Protection - A good batter behind you will allow you to get a better
quality of pitch and this will allow you to hit for more power. Sometimes
the "hitting for more power" can be a confusing term. If protection changes
12 "auto-walks" into real at bats, and the batter uses these opportunities to
hit one more home run, he's not hitting for more power if he normally hits
one home run per 12 at bats. We aren't looking at raw counting totals (eg.
home runs, RBI's), but rates. If strong protection is true, we should be able
to measure it by looking at slugging percentage.
Another way to view the difference between weak protection and stong
protection is to simplify the pitcher's thoughts into one of three modes. No,
this is not an accurate description of what goes through a pitcher's mind, but
it serves to give a "flavor" to the pitcher's attitude towards the hitter.
A: You are my vote for MVP, so I will toss you 4 balls in the dirt.
B: You can hurt me, I'm going to play with the corners, try to get you
out, but not give you anything good.
C: I really want to get you out and I'm going to challenge you. If you
can hit my stuff more power to you.
When Bonds had Shawon Dunston batting behind him, he got lots of A. A
person who believes in weak protection will claim that a pitcher will
switch to mode B when a better hitter than Shawon is betting behind Barry. A
believer in strong protection will claim that protection will shift the pitcher
into mode C and Barry will really be able to go to town.
In essence, it is a question of quantity and quality. Weak protection will
give you more true at-bats (fewer 4-balls-in-the-dirt encounters). Strong
protection will give you not only more opportunities, but better
opportunities.
-------------------------------------------------------------------------------
Q2 - Is strong protection real?
A2 - I've skipped weak protection. Most people are willing to believe that
weak protection exists [Is there a study that shows this?] People tend to
focus on whether or not strong protection exists. David Grabiner did a study
involving AL hitters in the 1991 season. He considered a player capable of
providing protection if he had a slugging percentage of .450 and had a "good
reputation." He then looked at players who had 100 at bats both protected
and unprotected. Of the 25 players who could be tested, 10 of them did
better and 15 did worse. He did not find a statistical link between protection
and slugging average. This would argue for strong protection being false.
The URL for this study is:
http://www.math.lsa.umich.edu/~grabiner/protstudy.txt
I looked at Barry Bonds' 1996 season when he had Matt Williams
fading in and out of the lineup. This study can be found at:
http://pubweb.nwu.edu/~dmarasco/protect.html
-----------------------------------------------------------------------------
Q3 - What about the STATS study?
A3 - Both in the near and far past STATS has printed studies that claim to
show protection. However, these studies suffer from a bad flaw. They look
at the total stats of all AB of good hitters with another good hitter on deck,
and the total of all AB of good hitters without another good hitter on deck.
The problem is that most teams with two great hitters bat their best hitter
third and their second-best hitter cleanup. To use a one team example, the
fact that Frank Thomas (protected) puts up better numbers than Albert Belle
(unprotected) means only that Thomas is a better hitter than Belle. To solve
this flaw you have to do what David did, take the SAME hitters and compare
their stats with and without protection.
------------------------------------------------------------------------------
Q4 - But when I played in highschool I know I saw protection,
how can it be false?
A4 - Protection in any form may exist at the lower levels of baseball.
All the studies have been performed at the major league level. Protection
could very well be real in highschool and the low minors, but the evidence
at the major league level shows this beast to be a myth.