keyboard_arrow_uptop

Is there anything more futile than predicting what pitchers are going to do?

As you peruse the boxscores from Thursday’s games, take a look at the people who have the letter “W” next to their names. It’s a fun group. There’s Mike Maroth, who last year became the majors’ first 20-game loser since 1980 with a sub-replacement-level performance. A.J. Burnett won yesterday, and he made just four starts in 2003. That’s four more appearances than Orlando Hernandez–who beat the Rangers last night–made last year. Shawn Estes, probably the worst starting pitcher in baseball in ’03, went to 13-4 (albeit with a high ERA and little else going for him) by beating the Phillies. And Jaret Wright, who wallowed through six years of misery after his star playoff turn in ’97, is suddenly the ace of the Mazzonified, division-leading Braves staff.

Last year, those five pitchers combined for the following line:


   IP    H    R   ER   BB    SO   ERA    REC  SNWAR      ARP
  425  507  303  263  201   261  5.57  19-39   -2.9    -29.1

Actually, Hernandez didn’t pitch, so it was just four pitchers creating that fun line. The Adjusted Run Prevented score is almost entirely Jaret Wright’s doing.

PECOTA’s weighted-mean projections for those five guys in 2004:


    IP    H    ER   BB    SO   ERA
 535.1  574   320  260   377  5.34

In 2004, those five guys are:


    IP    H    R   ER   BB    SO   ERA    REC
 563.2  546  280  262  217   389  4.18  38-21

(Updated Support-Neutral stats temporarily unavailable. These guys were 5.8 wins above replacement through Saturday).

We take a lot of grief for “There’s No Such Thing as a Pitching Prospect,” but the whole concept stems from the idea that there’s no such thing as a reliable pitcher. Sure, the top 5% of the species shows some consistency from season to season, but even in that group, you get some variation. How many fantasy teams did Randy Johnson take down last season? How many of you grind your teeth over the performances of Roy Halladay and Mark Prior this year?.

I’ll concede that this is just a snapshot, brought on by watching five guys who happened to all be lousy or injured or both last year all having pretty good days yesterday. Still, two of them, Burnett and Hernandez, were coming off of major surgery, and a third, Wright, hasn’t been good since 1998. If you can stand up in March and make the call that these guys will all be important parts of their teams, and in some cases, important parts of division winners, than start playing the lottery more often.

Think about last season, when guys like Esteban Loaiza, Brandon Webb and Darrell May emerged from mediocrity, obscurity or the Midwest to be among the top 30 starters in the game. Loaiza and Webb were reasonable Cy Young Award candidates in their respective leagues. This year, all three pitchers have fallen way off, with only Webb providing even average performance.

I have all the respect in the world for Nate Silver and his PECOTA system, which I think is a significant step forward in player projection. I just think that it–and every other projection system–is largely wasting its time in trying to sort out what pitchers will do from year to year. Whether it’s the physical toll, the mental strain or just the way the world spins, the vast majority of pitchers are completely unpredictable. They get good, and they get bad, and they get to all points in between, and they do so randomly.

This is why, when it comes to building a team, I don’t see any need to spend money on the middle of the pitching bell curve. If you can invest in the top tier, then you should do so. Get Johnson or Greg Maddux or Roger Clemens. Money spent on the Sidney Ponson class is money thrown away, because the chances of getting three straight good years from a guy like that is tiny.

So you build a staff around the very best, then fill in around it with low-risk gambles and guys you develop. Be aggressive in the free-talent market, and never overspend because a guy put together 70 or 150 good innings the year before. If you minimize your investment in any one guy, you limit your exposure on the downside and make it easier to create an opening for your version of Jaret Wright.

And yes, there’s another one out there. Just wait a year and see.