Wednesday, November 05, 2008

Read This Before League MVPs Are Announced

Simpleton baseball experts can be highly annoying. For example, when I hear an MVP debate in which someone argues for a hitter over a pitcher simply because "a hitter is an everyday player, whereas a pitcher only plays every fifth day," I want to pull my hair out.

I assume the logic behind the argument goes something like this: no matter how good a starting pitcher is, he can only affect the outcome of 1 out of every 5 games. A position player, on the other hand, can impact the outcome of every game. So a guy with roughly one fifth the potential impact of his MVP rival can't possibly be as valuable.

Well, let me ask you this: if I go to work for 30 minutes every day, does that mean I'm contributing more than a co-worker who goes to work only once a week but for 8 hours?

I think you see where I'm going with this. I certainly agree that a starting pitcher can only impact the 1 in 5 games he plays. But if it's potential impact we're concerned with, why are these pundits choosing "games played" as the metric? Why are they ignoring the potential impact within those games played?

Let me break it down:

PA's/ game (both teams) PA's/ game directly involving player % of PA's/ game directly involving player Games player plays in % of team's games player plays in % of team's PA's directly involving player
Hitter 81 4.4 5.4% 158.5 98% 5.3%
Pitcher 81 30.3 37.4% 33.8 21% 7.8%

Let's look at the hitter first. In an average MLB game, there are roughly 81 plate appearances total for both teams1. A typical MPV-caliber hitter averages 4.4 PA's per game2, or approximately 5.4% of the game's 81 PA's. And on average he plays in 158.5 of his team's 162 games2, or 98% of them. This means that for the season, the hitter is directly involved (i.e., at bat) in roughly 5.4% x 98% = 5.3% of the plate appearances of the games in which his team plays.

Now let's look at the pitcher. The typical MVP-caliber starting pitcher averages 6.9 innings per start and 33.8 starts per season3. The 6.9 innings per start means that in a given game, the pitcher is directly involved (i.e., on the mound) in 37.4% of the game's plate appearances. And the 33.8 starts per season means the pitcher starts 21% of his team's games. So across the season, the pitcher is directly involved in 37.4% x 21% = 7.8% of the plate appearances of the games in which his team plays. Compare this to the 5.3% for the hitter and guess what? By this metric it's actually the star pitcher who has significantly more potential impact than the star hitter.

Now, there's at least one big factor unaccounted for4: hitters, more so than pitchers, can substantially impact plate appearances in which they aren't directly involved. Ryan Howard on deck or Jimmy Rollins on the basepaths can certainly affect the current at-bat. And while I can't quantify this impact, I can assure you it isn't enough to boost the hitters' 5.3% factor very far beyond the pitchers' 7.8% figure - and certainly not enough, in my opinion, to give the "everyday player" argument validity. So in the next few weeks when you find yourself in an MVP discussion, if you hear someone base their choice solely on this argument, go ahead and kick 'em in the groin.

In all seriousness, this issue has been a pet peeve of mine for quite some time, as perhaps you can tell. Your thoughts?

1Derived from the MLB avg WHIP of 1.38. Three outs + 1.38 = 4.38 plate appearances per half inning. 4.38 x ~18.5 half innings per game = 81. I'm assuming the net effect of DP's, HBP, FC's, and other possible occurrences is minor.
2PA's per game and # of games played for typical MVP hitter taken from last four MVP seasons: Rollins '07, A-Rod '07, Howard '06, and Morneau '06.
3Innings per game and starts per season for typical MVP-caliber pitcher based on last four Cy Young seasons: Peavy '07, Sabathia '07, Webb '06, and Santana '06.
4Many minor factors, such as the players' fielding skills, NL pitchers' batting abilities, etc are also unaccounted for. I assert that these factors are negligible compared with the players' batting and pitching abilities.


bobmac said...

"Simpleton baseball experts can be highly annoying." I agree...but simpleton statistical analysis is equally annoying. While I have no problem with pitchers being MVP's, this is a very incomplete analysis.

Steve Sax said...

Can I kick 'em in the groin even if they agree with your opinion?

Eric Karros said...

Hey Bobmac - Yes, my analysis is certainly incomplete. While couched in a pseudo-quantitative framework, at heart it's really meant to be a qualitative rendering of how certain points are overlooked.

But I'd like to hear a little more detail on the main points you feel are missing (and preferably quantifiable)?

Sax - groin kicks never need justification.

berkowit28 said...

You're including, naturally, plate appearances for both teams, since pitchers' PA statistics are entirely for the defensive PAs. But you're looking only at offensive PAs - hitting - for the hitters. What about hitters affecting defensive PAs? They certainly play a part in putting out the other team's hitters.

Eric Karros said...

Berko - are you referring to the impact of a player's fielding prowess?

Steve Sax said...

EK, be careful. Bobmac sounds like he's gunning to kick your groin.

Orel said...

I wonder what Cheesymac thinks.

Eric Karros said...

Sax - that's why I wear a cast-iron cup 24/7.

Anyhow I hope Bobmac provides more detail on his thoughts. And I can't help but wonder if in person he bears as much resemblance to Chin-Lung Hu as our readers seem to think.

Steve Sax said...

EK, are you related to this guy? (see Actor movie #6).

(Who knew that he co-wrote the movie as well? What a renaissance man. Kinda like you, EK.)

(Orel, if you're counting, that's "renaissance" with a lowercase "r".)

Eric Karros said...

If Actor #6 were me, I wouldn't need a cup.

StolenMonkey86 said...

Now, there's at least one big factor unaccounted for: hitters, more so than pitchers, can substantially impact plate appearances in which they aren't directly involved.

No, they can't.

Zach Martin said...

Ignoring the fielding prowess is rather short sided. I know defensive ability is hard to judge, but totally ignoring it is like saying Jeter is as good a defender as Rollings or a 1B is as important as a SS or even worse saying that a P is as valuable defensively as a SS.

The argument seems valid, but there needs to be a deeper look into fielding. I don't know. Something like TC % breakdown would be necessary to make all segments of the game have value. I guess making fielding have less a value seems fair as well.

Eric Karros said...

Stolenmonkey - interesting link. I'm a little surprised by the findings, but they support the overall point I'm trying to make so I'm sure they're correct.

Zach – good comment. In fact it inspired me to think through the fielding side a bit deeper. Tomorrow I'll actually make a full post summarizing these deeper thoughts - keep on the lookout!