added 6/28/2007 by Scott Barzilla
When the Astros were at their best in 1980, 1986, 2004, and 2005, it was stellar pitching that led the way. The likes of Nolan Ryan, Joe Niekro, Mike Scott, Roy Oswalt, Roger Clemens, and Andy Pettitte led the way. So, the temptation is to chase the big starting pitcher in free agency. Carlos Zambrano and Mark Buehrle certainly fit that bill, but are starting pitchers wise investments?
One of the big problems with selecting pitchers is that the statistical evaluation of pitchers is far behind the curve. Teams are still relying on wins, ERA, and WHIP to make their free agency choices. They get burned more often than not. If you use win shares you will see the folly of giving big money to free agent pitchers. Following are the ten highest paid pitchers and their performance over the past three and a half years (2007 is projected to a full season).
2004 2005 2006 2007 AVG CPWS AJ Burnett 7 12 10 13 11 0.8 Roger Clemens 20 25 12 10 17 1.2 Bartolo Colon 10 19 1 2 8 0.5 Pedro Martinez 17 18 5 0 10 0.7 Roy Oswalt 19 22 21 15 19 1.5 Andy Pettitte 6 22 12 15 14 0.9 Johan Santana 27 23 25 17 23 1.8 Jason Schmidt 19 8 16 0 11 0.7 Javier Vazquez 10 14 11 13 12 0.9 Barry Zito 12 14 18 7 13 1.3
Before we start analyzing we need to look at how this was done. Total baseball comes up with the win shares and ESPN.com had the salaries. ESPN using the current year's salaries, so Zito's numbers are better than they will be. So, essentially only three guys averaged one win share per million dollars. In order to understand how putrid this is we need to look at win shares in general.
We need to remember that three win shares equal one win. Another words, if we were to spend one million per win share then you would need to spend at least 270 million dollars to have a legitimate shot at the playoffs. Maybe the Yankees can afford that, but no one else can. So, investing in pitchers is clearly fool's gold, but we need to spend some time figuring out why.
The problem with pitchers is three-fold. First, when you play in only 32-35 games a season, your opportunity to help your team win games is limited. Obviously, it could be argued that the starting pitcher is the most important player on the diamond, but he is only on the diamond once or twice a week. Compare that with a position player and you see why those 150+ games a season add up in terms of value.
Pitchers are less predictable in terms of health. Bartolo Colon, Jason Schmidt, and Pedro Martinez have all missed most of the season this year or been ineffective because of injuries. Six of the ten have had seasons with less than ten win shares. This is mostly due to injuries. The pitching motion is an unnatural motion, so if you want to invest in a mid-thirties pitcher you better be careful in what you are investing in.
Finally, you get the problem with evaluating pitchers. Most teams still use wins, losses, ERA, and WHIP. Sabermetricians have found DIPS (defense independent pitching statistics) to be much more reliable. Unfortunately, those stats are not nearly as sexy as the traditional stats. However, if we look at the top two free agent arms in 2007 and one mystery pitcher you will see how unpredictable this is.
SO/9 BB/9 SO/BB HR/9 Mark Buehrle 5.94 1.80 3.30 1.17 Carlos Zambrano 7.60 3.80 2.00 1.29 Mystery Pitcher 7.92 2.33 3.40 1.22
There is a lot to like about Buehrle and Zambrano. In particular, Zambrano has a very good strikeout rate and Buehrle's control is impeccable. However, the mystery pitcher is none other than Wandy Rodriguez. Does this mean Wandy is just as good as Zambrano and Buehrle? Probably not, but it means that paying 15 to 20 million dollars a year for a pitcher doesn't make a whole heck of a lot of sense.
Next time we will look at the top ten position players and how they do according to the win share test. The key will be whether they are more reliable than pitchers. It may not sound intuitive, but going with position players through free agency and pitchers through the system makes sense.