Baseball By The Numbers By Michael Salfino

March 9, 2006
Fielding-Indendent Pitching Projections

There are the hits and runs a pitcher actually allows each year and those he should have allowed given league-average luck and defense. Focusing on the latter rather than the former significantly alters pitchers' values. More important, these luck-adjusted stats have proven to be of much more predictive.

The formula focuses on those things the pitcher most directly controls: strikeouts, walks, homers allowed. Hit hit rates are assumed to be league average (about 30.5% of balls put in play). ERAs are recalculated accordingly.

Yes, there are some pitchers who consistently allow a lower percentage of hits on balls put in play (most knuckleball pitchers, for example, though they're a vanishing breed). But the vast majority are all over the map. Stuff guys like Pedro Martinez and Randy Johnson have had low percentages one year and the high the next. The same is true of junk-ballers. So assuming average luck/fielding gives us the most reliable way to predict ERA.

The statistic is known by many names, but we like FIP for Fielding Independent Pitching. The stat, created by Voros McCracken (who quickly became a Red Sox consultant), is simple, reliable and wonderfully intuitive. Baseball fans have always been aware of the luck involved on hits and outs, when bleeders dribble through the infield and line drives are at 'em balls. Now we can objectively quantify this luck.

Of course, some teams have really good defenses and should be expected to generate more outs on balls in play. The iron gloves, significantly less. Last year, the best defenses in converting outs were the A’s (71.7%), White Sox (71.3%), Indians (71.1%), Astros (70.6%), Phillies (70.5%), Cardinals (70.4%), Twins (70.3%), Mets (70.2%) and Mariners (70.2%). The worst, the Royals (66.7%), Rockies (67.1%), Reds (67.8%), Marlins (67.8%), Devil Rays (68.2%), Rangers (68.2%), Red Sox (68.3%), Diamond Backs (68.4%).

If that doesn’t sound like a great range, remember that one percentage point in differential translates into 10 points in batting average.

Here’s a sampling of pitchers who were very lucky or unlucky in hits allowed and ERA in 2006:

Buy

Daniel Cabrera (SP, Orioles): This fireballing giant was about half a run unlucky in ERA last year even while averaging about five walks per start. New pitching coach Leo Mazzone should fix that. O’s lefty fireballer Erik Bedard was unlucky, too (ERA should have been 3.59, not 4.00)

Ryan Wagner (RP, Reds): His ERA should have  been 3.73, not 6.11. This makes the 23-year-old a better candidate to replace 36-year-old closer and journeyman David Weathers than is generally assumed.

Hold

Aaron Harang (SP, Reds): Only mildly unlucky last year but the news is that Harang’s peripherals more than support 2005 results. Yet he’s being discounted everywhere on the assumption ’05 was a fluke, despite the 3:1 strikeout-to-walk ratio.

John Lackey (SP, Angels): Has been striking out better than eight per game for a season and half. In the AL, only Johan Santana, Rich Harden, Roy Halliday and Felix Hernandez had better FIP ERAs than Lackey (3.20 vs. 3.44 actual).

Sell

Tim Hudson (SP  Braves): The K-rates don’t support his ERA (should have been 4.34, not 3.52). Ditto Jorge Sosa (4.52 FIPs vs. 2.55 actual) and new Dodger Jae Seo (3.75/2.59).

Zach Duke (SP, Pirates): Another starter who isn’t dominant enough to support his ’05 stats (FIP ERA of 3.20 vs. 1.80 actual).

Kevin Millwood (SP, Rangers): The AL ERA champ goes from the major’s 26th best hitters’ park (Cleveland’s Jacobs Field) to the fourth best. Also note the difference between the Indians and Rangers in fielding efficiency and expect an ERA of at least 4.00 even if he stays healthy.