Are pitchers really overvalued?

“The market right now is kind of silly, and it may continue to be silly.”
- St. Louis Cardinals general manager Walt Jocketty (ESPN.com; November 28, 2006)

So just how silly is the market?

Quite, actually, but particularly when it comes to pitchers.

But doesn’t defense win championships? Isn’t that the idea that we’ve all been taught, weaned on, suffocated with, what have you? Shouldn’t there be a premium on pitching, if it is really that important?

Let’s consider. Wins can be broken down and credited to three major factors: runs scored, runs allowed, and luck. Pitchers have almost no effect on how many runs their team scores, but a substantial hand in how many their team allows. Substantial is the key word here. Most fans and, incredibly, team execs seem to believe that a pitcher is almost entirely responsible for the runs he lets up, hence the reliance on ERA to evaluate performance.

Of course, because of the unpredictability of balls put in play, we know that pitchers are probably only 60-75% responsible for the runs they allow (or don’t), with the eight other fielders accountable for the rest. For the sake of being generous, let’s give pitchers the high end stake, and assume they are 75% responsible. We’ll also assume that run prevention is just a tad more important than run scoring. Here’s how a team’s total value is then broken down:

- Hitting (48%)
- Pitching (39%)
- Fielding (13%)

So we can assume that in a completely efficient market, pitchers would receive about 39% of all player salaries. But according to Sean Lahman’s database, this figure has actually fallen anywhere from 40 to 48 percent since 1985.

So where is this market-wide inefficiency coming from? There is an ever-so-slight positive correlation between year-to-year rises in total salaries and percentage spent on payroll, with a .13 r-squared value. This could mean that there is a small tendency to spend excess dollars on pitching.

A more logical explanation, though, is simply that teams have overvalued the importance of pitching. Just like with any other market, this is likely due to poor information and analysis. Here are a few likely culprits:

1) Focusing on the wrong statistics. The standard story goes as follows: in his walk year, a below average pitcher puts up his normal (or perhaps slightly improved) defense-independent rates (strikeouts, walks, groundballs), but is extremely hit lucky. Teams that need a starter bid each other up, hoping against hope that this hurler’s “new” performance level is for real.

This occurs because too many teams still use ERA and (wait for it) win-loss record to evaluate a pitcher’s performance. Both are terribly flawed metrics that are largely subject to luck, particularly over the course of just one season. And this ends up causing other problems, such as…

2) A perceived shortage of capable pitchers on the market. Consider the nature of ERA. Over many, many innings, a pitcher’s ERA should reflect his defense-independent ERA, adjusted for his home park. One season’s worth of data (or roughly 200 innings for a starter), however, is far too little to create this effect. This, in turn, creates a superficial increase in variance between pitchers.

Let me try to explain that a bit. Looking at all pitchers who threw at least 50 innings in 2006, the standard deviation of all ERAs is 1.21 runs. When looking at xFIP, a defense-independent pitching stat from The Hardball Times, the standard deviation is reduced to .74. Using Nate Silver’s QuikERA (subscription required) yields a similarly low result, at .79. In other words, the real variance between pitcher performances is much smaller than the perceived variance.

So how does this impact the market? A truly league average starter can look very good or very bad, depending on his surroundings. The ones who look very bad will often be pulled from the rotation entirely. Therefore, at the end of the season, it appears that there is only a very small subset of pitchers that can be trusted. Teams are then willing to pay a premium for these guys, since there appears to be a bit of a shortage.

3) Not taking into account a pitcher’s inherent risk factors. When the Brewers signed Jeff Suppan to a four year, $42 million contract this past offseason, general manager Doug Melvin stressed Suppan’s “experience and durability.” Of course, Carl Pavano hadn’t been on the disabled list for three years prior to signing with the Yankees after the 2004 season.

Yes, some pitchers are more likely to get hurt than others. But all pitchers have a significant inherent injury risk due to the violent, unnatural duty they have to perform. Teams seem to be paying a premium for pitchers with minimal injury histories, which is similar to a life insurance company covering Evil Knievel after a couple successful jumps. Pitchers should actually be discounted relative to hitters, who generally bring a more certain level of value. Hurlers with less risk for future injuries (because of better mechanics, a better build, etc.) should then be discounted less.

This logic is unlikely to take hold, though. As usual, most teams seem to be playing follow the leader and taking the easy (and inefficient) way out instead of being a bit more creative and trying to beat the market. We’ll discuss some potential alternative strategies later in the week.

Feedback? Write a comment, or e-mail the author at shawn(AT)squawkingbaseball.com