Tigers' Pitchers: A look at theoretical versus actual performance.

I was interested in looking at how Tigers pitchers performed in actual earned runs allowed compared to FIP and xFIP. Specifically, I was interested in seeing if certain types of Tigers pitchers either exceeded or failed to meet their saber expectations. More after the jump!

Below is a chart that shows how each pitcher who threw more than 20 innings. If a pitcher threw less than 20 innings, then their rate stats are exceptionally unreliable. Also, the chart would be too big. It has columns for Innings pitched, ERA, FIP, fipruns, fipruns-runs,xFIP, xfipruns, xfipruns-runs, and xfipruns-fipruns. Innings pitched, ERA, FIP, and xFIP are all things we're familiar with, and I got the data from fangraphs.

Fipruns and xfipruns are calculated by multiplying the pitcher's innings pitched by their FIP and xFIP respectively, and then dividing by 9. This number is the number of runs a player theoretically should have given up, based on their FIP and xFIP.

The fipruns-runs and xfipruns-runs columns are the difference in the player's theoretical runs given up and actual runs allowed. These show by how many runs a pitcher exceeded his FIP and xFIP expectations. If a player has a negative value that means that their actual performance was that many runs worse than what the saber stats predicted, whereas if a player has a positive value in one of these columns, it indicates that the pitcher's performance was that many runs better than what was calculated. So, for example, Verlander should have given up 16 and 20 more runs last season than he did, based on his FIP and xFIP values, respectively.

The last column, xfipruns-fipruns, shows the difference in the number of theoretical runs calculated from FIP and xFIP, which is the difference between the pitcher having a league average home-run per fly-ball rate and the rate with at which they actually allowed flyballs to become homeruns. So, for example, Scherzer's xFIP is quite a bit smaller than his FIP because his HR/FB rate was high, where Coke's xFIP is larger than his FIP because his HR/FB rate was rather low. As the HR/FB rates of pitchers tend to fluctuate quite a bit, using the average rate is better at predicting future performance. However, the Tigers have several pitchers who appear to be consistently good at preventing fly-balls from becoming home-runs, to the point where it, mathematically, looks like a skill and not just a sustained lucky streak. I'll delve into this more in a bit.

player innings pitched era Earned runs fip fipruns Fipruns-runs xfip xfipruns Xfipruns-runs Xfipruns-fipruns porcello 182 4.75 96 4.06 82 -14 4.02 81 -15 -1 coke 108.2 4.47 54 3.57 43 -11 4.42 53 -1 10 scherzer 195 4.43 96 4.14 90 -6 3.7 80 -16 -10 perry 37 5.35 22 3.94 16 -6 5.22 21 -1 5 penny 181.2 5.3 107 5.02 101 -6 4.77 96 -11 -5 below 29 4.24 14 4.2 14 0 4.77 15 2 2 benoit 61 2.95 20 2.96 20 0 3.29 22 2 2 alburquerque 43.1 1.87 9 2.08 10 1 2.78 13 4 3 furbush 32.1 3.62 13 5 18 5 4.43 16 3 -2 fister 70.1 1.79 14 2.49 19 5 2.75 21 8 2 schlereth 49 3.49 19 5.03 27 8 4.89 27 8 -1 valverde 72.1 2.24 18 3.55 29 11 4.01 32 14 4 verlander 251 2.4 67 2.99 83 16 3.12 87 20 4 Totals: 1312.2 3.76 549 3.79 553 4 3.89 567 18 14

Ultimately, when just looking at ERA, FIP, and xFIP, I was unable to come to any strong conclusions about Tigers' pitchers as a group. No specific group of pitchers was decidedly better or worse in actual performance compared to theoretical expectation. Some ground-ball pitchers exceed the sabermetric expectations, while others fail to meet them. The same is true of the starters, relievers, "good pitchers," "bad pitchers," lefties, righties, veterans, young kids, and every other cluster I can think of. The only thing that really stands out to me is that the pitchers, when taken as a whole, exceeded their FIP expectation by four runs, and their xFIP expectation by 18 runs. When you remove Penny and Perry, the two pitchers on this list who are no longer Tigers, these difference numbers jump to 16 and 30, respectively. Given that FIP and xFIP are often touted for their ability to predict future performance, this would seem to suggest that the Tigers pitching staff, on a whole, is likely to regress.

However (!), xFIP is, as mentioned before, calculated using the league average HR/FB rate as opposed to the player in question's HR/FB rate, and the Tigers have several pitchers on their staff who have long histories of outperforming the league average home run rates (This is especially true if you choose to discount rookie years, which is intuitive, as rookie years have less correlation with career numbers than a player's subsequent seasons). For example, Verlander has never thrown over the league average HR/FB rate, and, thus, xFIP is probably not as accurate a measure of his future performance as FIP is. Valverde has pitched under the league average homerun rate in four of the last five seasons, the lone exception being his first season for the Astros, who, incidentally, play in a park that is consistently in the top 10 in HR Park Factor. Benoit has had a better than average rate in every season since he switched to primarily relief work in 2005. Both Coke and Fister have had two post-rookie seasons of decidedly better than average HR/FB rates, and while it's hard to suggest a career pattern out of two seasons, if they repeat the feat this year it's likely an accurate representation of their skills rather than "luck." Porcello and Scherzer both appear to have more or less average career HR/FB rates, although Max's was quite high last year.

Comerica's HR Park Factor (from ESPN) does not likely play a strong factor in contributing to the Tigers' pitchers who outperform the league average HR rates, as it has been rather close to both the median value (14th, 18th, and 18th in all of baseball over the past three seasons) and average (approximately 96% of the league average rate for the past three years). The Hardball Times 4-year home-run per outfield fly rate also has Comerica at 96% of league average. These minute differences are not enough to suggest that, for example, Verlander's consistently low HR/FB rate is due to pitching half of his games in Comerica.

These pitchers who consistently outperform the league average HR/FB rate (Verlander, Valverde, Benoit) threw approximately 27 percent of the Tiger's inning last season (nearly 40 percent if Fister and Coke are factored in). Having such a significant percentage of our innings pitched by players who are better than average at preventing fly-balls from becoming home-runs will help to mitigate some portion any expected regression.

Lastly pitchers who threw less than twenty innings for the Tigers threw a combined 127 and a third innings (Donnie Kelly was the third of an inning! Pitching Wizard!), or about 9 percent of the total team innings. They allowed 98 earned runs, or over 15 percent of the total team runs allowed. Clearly pitching depth was a huge problem last year.