Why has Miguel Gonzalez consistently been able to outperform his peripheral statistics?

The 2014 Orioles starting five pitchers ended the 2014 season with an ERA+ above 100 (Meaning, they were all above average, adjusted for park and league). Somewhat surprisingly, Miguel Gonzalez had the best season of the bunch. He posted a 119 ERA+ (meaning his ERA was 19% better than the league average). This was five points better than the next starter, the supposed staff ace Chris Tillman.

Using the Baseball Reference version of WAR, he was the second best Orioles starter posting a 2.1 WAR for 2014. However, according to the Fangraphs version he was only worth 0.6 WAR. The difference? Baseball Reference calculates a pitchers value based on his runs against average while Fangraphs calculates pitcher value using his fielding independent numbers.

To describe it somewhat succinctly, fielding independent pitching (FIP) attempts to take the fielders and statistical luck out of the equation when evaluating a pitcher. This means that strike outs, walks, hit by pitches, and home runs are what is solely is used to calculate FIP. All the batted balls and timing are set to league average rates. Then, a constant is applied to make the number equate similarly to ERA.

The theory behind FIP is that it shows a pitchers true talent, because it only takes into account what a pitcher can truly control. An above average defense or a bit of statistical luck can produce a shiny ERA, but that ultimately tells you little about the pitcher--so the theory goes. In my opinion, FIP is a reductive statistic and takes out way too much of the nuance of baseball to be taken on it's merits alone. It is a sound statistic that is very useful in judging pitcher performance, but in my mind is too simplistic to be used alone. There are other ways to get outs than striking someone out and not giving home runs. Pitching is more complicated than that. Miguel Gonzalez has been an example of this over the past three seasons.

Gonzalez has consistently outperformed his FIP over the 435.2 innings he has pitched for the Orioles. He has posted a career 3.45 ERA (Good for a career ERA+ of 117) and a career FIP of 4.59. Gonzalez's particular batted ball profile has remained consistent over his three seasons in the majors and his particular profile is one that is not particularly FIP friendly. Gonzalez has a career 21.3 percent line drive rate (LD%), 37.4 percent ground ball rate (Gb%), and a 41.7 percent fly ball rate (FB%). For some perspective, for pitchers from 2012 to 2014 with at least 400 innings pitched (92 qualify) Gonzalez is 41st in LD%, 84th in GB%, and 8th in FB%.

Also, his home run to fly ball ratio (HR/FB) is 11.3 percent which is 30th among the same group of pitchers. Lastly, his Infield Fly Ball rate (IFFB%) was 12.2 percent over the same period of time good for 5^th among the qualified pitchers. So clearly, Miguel Gonzalez is effective at producing fly balls and in particular infield fly balls and has shown over the course of his major league carer that he can replicate that result.

The other aspect of Gonzalez's profile that stands out his ability to strand runners. Gonzalez has a career 80.3 percent Left on Base Percentage (LOB%). That is good for 2^nd among starters from 2012 to 2014 with at least 400 innings pitched. Furthermore, he has a career BABIP of .265, which is 6^th lowest among the same qualified starters. Those numbers, for a typical pitcher would be concerning. This is one of the main driving forces behind Gonzalez's poor FIP. League average BABIP tends to fall around .300 and while pitchers can influence this number some, the common belief is that the number will always either fall back or rise up to .300. Furthermore, LOB% also tends to fall around the average of 70 to 72 percent. Any variation, so the theory goes, is explicable mostly by variance out of the pitchers control.

Well, Gonzalez has shown an ability to induce fly balls and infield fly balls at rates much higher than league average. Infield fly balls are especially important because an infield fly ball is an out nearly 100% of the time and it is a good sign that a pitcher is inducing weak contact. However, Gonzalez has a well below average strike out rate in his career at 17.0 percent (72nd among qualified pitchers) and a career 7.7 percent walk rate (23rd highest among the qualified pitchers). So it does not appear that Gonzalez has either great control or great stuff. Yet, he still manages to strand runners, keep his BABIP low, and get fly balls.

Gonzalez does this by utilizing his fastball up in the zone. Throughout his career he has always kept his fastball up in the zone. Mostly around a quarter to a half a foot above the dead center of the zone. When he was at his most successful last year in the second half, he threw his fastball on average a half of foot above dead center. See the chart below for the whole story.

By controlling his fast ball and keeping it up, he induces lots of swings and fly balls. Since fly balls fall for hits less often than any other type of batted ball, this keeps his BABIP low. The high amount of infield fly balls and the stable rates of fly balls over his career show not only his ability to keep generating them, but also his ability to limit strong contact even though he has an above average contact rate overall. This allows him to strand more runners. He still gets in trouble with the home run ball, but for a pitcher with his stuff and how he utilizes it, he keeps it low enough to still remain effective. This has allowed, and I believe will continue to allow, Gonzalez to outperform his FIP.

Just so you believe that I am not making this all up. Two other pitchers have similar statistical and stuff profiles to Gonzalez. Chris Tillman and Jered Weaver both have similarly high fly ball rates, low strike out rates, above average walk rates, and have  outperformed their FIPs over the same period of time. Below is a table comparing all three from 2012 to 2014.

Pitcher	ERA	FIP	LOB%	BABIP	FB%	IFFB%	K%	BB%	HR/FB%
Miguel Gonzalez	3.45	4.59	80.3%	.265	41.3%	12.2%	17.0%	7.6%	11.3%
Chris Tillman	3.42	4.22	77.5%	.260	40.4%	10.1%	19.2%	7.7%	10.1%
Jered Weaver	3.24	3.94	78.6%	.259	45.9%	14.0%	18.9%	6.5%	12.0%

As you can see in the table above they all have very similar profiles and indeed, with BAPIP, LOB%,  and FB% they are all basically within the top 10 for the past three seasons--again--among pitchers who threw at least 400 innings. Clearly, Miguel Gonzalez is not the only pitcher who has shown an ability to outperform his FIP over a lengthy period of time. This is done by producing fly balls and weak contact using his fastball up in the zone. Of course it would be better if Gonzalez started striking more hitters out and giving up fewer walks and home runs, however he has shown that he does not need to do either to still be successful at preventing runs.

The point of all of this is to say that using one number, whatever that number may be, to fully analyze one player is foolhardy. Saying that batter X is better than batter Y because his WAR is .3 higher, or that pitcher X is better than pitcher Y because his FIP says so is far too simplistic and antithetical to everything analysis should stand for. Using one statistic to describe a player, or an entire team for that matter, is a failure of analysis. Miguel Gonzalez stands as one reason why analysis should go deeper than simply one measure.