Part three of our foray into sabermetrics is pitching. What pitching metrics exist? And how could they better define performance than the statistics we use now? Individual pitching performance is somewhat difficult to ascertain because the traditional stats we’ve been using still remain very defense and bullpen dependant. Win-loss record isn’t very telling. We don’t have to go much further than the ace of the New York Mets to realize that. This year’s American League Cy Young Award winner would agree. What we use to evaluate pitchers are a few different DIPS (defense independent pitching statistics)

**FIP**

One metric often used is FIP (Fielding Independent Pitching). One’s ERA can suffer if there’s a poor defense backing him up. Research by a young man named Voros McCracken learned that year after year, there was no consistency between the number of balls that fell in for hits. He concluded from this evidence that once the ball is hit, the pitcher is removed from the equation, and thus has no control over the outcome. McCracken wrote that a truer idea of the pitcher’s performance can only be ascertained by measuring his abilities in the areas defense has no say in; strikeouts, walks and home runs. The next step was to weight these accordingly, as a walk is not as damaging as a home run, amongst other measures of FIP.

Extensive research by none other than leading sabermetrician Tom Tango has shown that the best formula to determine FIP is ((13*HR)+(3*(BB+HBP-IBB))-(2*K))/IP+constant.

The constant is the league-average FIP subtracted by the league-average ERA. It’s roughly 3.20 each year. The reason we use the constant is bring FIP to an ERA scale, for familiarity. And by using league-average metrics, we can normalize FIP to show what the ERA would look like if a league-average defense was playing behind the pitcher. FIP is great because it shows how the pitcher pitched regardless of outside factors that he can’t control.

**xFIP**

xFIP (“x” stands for expected) takes Fielding Independent Pitching one step further and corrects for the luck or park factor involvement in home runs (wall scrapers in Philly or down the right field lines in the Bronx or Boston). Pitchers should have a roughly 10.6% HR/FB ratio, so the xFIP formula looks like this: ((13*(.106*# of fly balls))+(3*BB+HBP-IBB)-(2*K))/IP+constant. Of all the pitching metrics, xFIP is the most telling of future performance.

**tRA and tERA**

tRA and tERA (the “t” stands for true) takes it eve further and gives run and out values to the type of batted balls the pitcher allows, including GB, FB, OFFB, IFFB, and LD in addition to the at-bat outcomes used in FIP and xFIP. Roughly 40 years worth of pitching data was used to evaluate the run values. It’s basically an extension of FIP and xFIP because those stats give all batted balls the same value. tRA and tERA involves lots of numbers and could get complicated, but it doesn’t really have to. To simplify the process, using the number of each outcome of an at bat and the expected values of the runs and outs of those outcomes, you can determine the expected runs and the expected outs a pitcher should achieve. The formula to determine tRA is expected runs/expected outs*27. That number reflects the expected runs a pitcher will give up per nine innings pitched.

The easiest part of this statistic is turning tRA into tERA to scale it to ERA. To do this, just multiply tRA by .92. The reason is because roughly 92% of all runs scored are earned.

Note the following chart of the top ten pitchers in the 2010 National League in all four pitching categories:

What are some determinations we can make from these lists? First, we can determine that Josh Johnson and Adam Wainwright are elite pitchers. Shocking, I know. We can also determine that the Marlins probably had a poor defense, because Anibal Sanchez has a top-ten tERA but finished outside the top twenty in ERA. Lo and behold, the Marlins were bottom five in the league in three of the four components of UZR.

What about the Brewers defense, since Gallardo finished in the top ten in FIP and xFIP and but 30^{th}overall in ERA? The Brewers were actually a middle of the road defense, just about league average. But wait, you say! Where’s Gallardo on the tERA list, if he pitched so well despite his average defense? He’s actually a great example. Gallardo was 45^{th}in the league in LD%. And since tERA takes the individual value of batted balls while FIP and xFIP values all batted balls the same, he misses the tERA cut, finishing 19^{th}.

Hiroki Kuroda is another interesting case. The Dodgers were third-worst in UZR/150 in 2010, explaining his top ten finishes in the DIPS despite his 18^{th }place finish in ERA. And what about Derek Lowe? Why is he in the top 10 in xFIP and none of the other categories? Remember that xFIP takes into account the luck and park factors in home runs. Lowe is a notorious sinkerball, ground ball pitcher and he didn’t disappoint in 2010, finishing with the second-lowest FB% in the league. What’s fascinating about Lowe, however, is his 44^{th} place finish in HR/FB%.

I researched further on why Lowe would seemingly be so unlucky. As it turns out, his 18 homeruns (still a pretty solid total) were nearly split down the middle, eight at home and ten on the road. According to ESPN’s park factors, Turner Field was the 8^{th} easiest park in the NL to hit a home run in, roughly average, behind some of the usual suspects (Colorado, Milwaukee, Cincinnati, Chicago, Philly, Houston and Arizona). So according to his game log, Lowe gave up 13 of his 18 home runs in very hitter friendly parks. Now it makes much more sense.

Conspicuous by his absence in the tERA chart is Doc Halladay. He finished 11^{th}with a 3.35, just missing the cut. Like Gallardo (but not as extreme), Doc’s LD% wasn’t very elite, finishing 30^{th} in that category.

Three things stand out to me while looking at these charts. First, I’ll bet anyone a shiny nickel that either Josh Johnson or Adam Wainwright will win the Cy Young award in 2011. Secondly, it stands out how accurate these DIPS are when you look at the cases of Kuroda, Gallardo, Halladay, Lowe, etc. But the thing that stands out to me most is the Doc Halladay finished first in xFIP, despite pitching in a bandbox. To me, that’s amazing. He also finished first in the AL in xFIP in 2008 and 2009 despite Rogers Centre being in the top half in park factor for homeruns. I know he’s sort of a late bloomer and is still two seasons away from 200 wins (let alone 300) at age 33, but if he’s not a Hall of Fame pitcher when he retires, then the Hall of Fame has lost its way.