Tom Glavine, despite putting the final nail in the coffin of the ’07 Mets, has for the most part been a pretty valuable pitcher for the team. Since coming over from the Dark Side in 2003, Glavine has pitched at least 180 innings every year with a 4.52 ERA or better. While that’s not exactly Tom Seaver, it’s certainly nothing to sneeze at. In fact, that’s like a number two or three starter.

Anyway, when Glavine declined his $13 million player option on October 5th he became a free agent, and I think most of us expect he’ll go back to Atlanta. So how much are the Mets going to miss having him around? Well, the answer to that has a lot to do with how rare you think a number three starter is, and that has a lot to do with how talent is distributed in Major League Baseball.

Take a look at this graph here. It’s the number of players of various heights on the Mets 40-man roster as of September 30th. Like most things that are randomly distributed, there are a couple of outliers at each end of the graph (Anderson Hernandez is short! Mike Pelfrey is tall!), but most of ‘em are pretty close to the middle. You may remember this from high school as the bell curve.

However, talent in the major leagues is not normally distributed – there’s a selection bias. That’s a statistical way of saying that the only people who get to play in the big leagues are at the extreme end of the talent curve, which seems pretty self-evident. Obviously, people who are terrible, or even mediocre at baseball don’t ever get to play in the big leagues. They say that only one out of 100,000 people who try will ever make the major leagues. Those are some pretty long odds.*

If you could graph the baseball skill of everybody in the whole world, it would probably look like this:

Most of us would be clumped somewhere close to the middle, but this talent graph includes both the incredibly good (Babe Ruth) and the terrifyingly bad (the blind, deaf, mute, armless and legless dude from Metallica’s video for One). See that teensy little green section way over on the right? That represents the tiny fraction of the population that’s good enough to play big league baseball.** So how exactly does this help us determine how much the Mets will miss Tom Glavine? Let’s take a closer look.

Last year, Tom Glavine pitched 200.1 innings with a 4.45 ERA (I wish he’d just stopped at an even 200), which is almost exactly league-average. That’s actually pretty valuable, since league-average pitching isn’t easy to come by. Here’s where we run into the difference between ‘average’ and ‘median,’ which I’ll explain like this: if you put the 1996 US Olympic Women’s Gymnastics team in a room with Shaquille O’Neal and Jerome James, the average height of the people in the room would be 5-5”, but the median height would only be 5-0”. In this room, somebody who’s five feet tall would be taller than one half of the people, and shorter than the other half. The presence of the two NBA pituitary cases throws off the average.

So while Glavine’s ERA was league-average, it was still better than 55% of the league, which, in a way, is slightly above average (well, above the median, technically). That’s because in the major leagues, good pitchers throw a disproportionately large number of the innings, while bad pitchers get sent back to Triple-A. Again, this seems self-evident. And so while there are many more bad pitchers than good, each of them only pitch a fraction of the innings a good pitcher does. Consider the case of Dan Serafini, who pitched a whole third of an inning at a 54.00 ERA pace.

This sounds like bad news for the Mets, losing 200 innings of better-than-most production. Sure, Pedro’s back, but the Mets will be lucky to get 30 starts from him and El Duque combined. So who will replace Glavine? Pelfrey, who may or may not ever learn a secondary pitch? Or Humber, whose one start this year was, shall we say, less than inspiring? Well, it turns out that Glavine was really only average when measured by his ERA, which as we all know, isn’t always the best measure of how well a pitcher pitched, and it looks like Tom got a bit lucky in 2007; his FIP (Fielding-Independent Pitching) was more like 4.86. Still not bad, but it’s definitely below average.

And on the flipside, it turns out that Mike Pelfrey was similarly unlucky. Pelfrey threw 72.2 innings with an ERA of 5.57 which, granted, hardly inspires confidence. But his FIP indicates that his ERA should probably have been more like 5.04. Yes, that’s still worse than what Glavine threw last year, but not a whole lot worse. Given that Glavine, at age 42 next year, will almost certainly be a bit worse, while it’s at least possible that Pelfrey or Humber will be a bit better, I think the Mets can replace Glavine’s production pretty easily with in-house options.

So the answer to our question is that the Mets will miss having Glavine around a little bit, but probably not as much as you might think. In fact, Glavine is pretty unlikely to be worth the $13 million from the player option that he just declined. At any rate, the Mets should be able to find someone close to the middle of the league without mortgaging the farm system for one year of Johan Santana. They should save that money to spend on A-Rod. Or not.

*Another way to look at it is, even the crappiest major leaguer is better at his job than I will ever be at anything. I actually have no idea if that one-in-100,000 figure is true or not, but you hear it a lot, so let’s just go with it.

**That tiny green section is actually much larger than the real-life .001% of people who have the talent to play Major League ball, but I wanted it to be visible. An all-red graph is boring.

Thanks to the Hardball Times, David Pinto’s Baseball Musings, Baseball-Reference.com, and the Worldwide Leader for making my inaccurate and uninformative writing possible.