FRIAS Case Study: Ervin Santana
Ervin Santana’s career has taken a very strange path (but isn’t that true of most pitchers?). An amateur free agent out of the Dominican Republic, Santana was regarded as a decent prospect coming up through the Angels system and broke in during 2005. After throwing some below-average bulk innings from 2005-2007 (xFIP-‘s of 109, 107, 106), he posted a huge 2008 season that was highly regarded from all angles: a 3.49 ERA, 84 xFIP- and 10.8% SwStr%. He also averaged a career-high 94.4 mph on the fastball, which was a big spike from 2007’s mark of 92.2. FRIAS, my projection system that uses previous-year peripherals – notably plate discipline stats – liked Santana’s 2008 enough to project a 3.24 FIP in 2009.
Alas, his entire profile took a step back for the next four seasons, when he was a fairly consistent below-average innings-eater who experienced bouts of terrible HR/FB luck which made him look unplayable at times. Then Kansas City acquired him prior to 2013, he pitched admirably in a walk year and Atlanta got a decent season out of him in 2014, paying $14.1M for 2.8 fWAR. In December, Santana signed with Minnesota for 4/54 in a deal that analysts seemed indifferent to. David G. Temple of Fangraphs and Sam Miller of Baseball Prospectus concluded as much in their analyses of the transaction.
So what does 2015 hold in store for Ervin Santana? It varies wildly depending on who you ask. Here’s a table of what some projections think regarding FIP coefficients.
What stands out is that Steamer, ZiPS and PECOTA do not believe in Santana’s 2014 strikeout jump. In the five seasons prior to 2014, Santana’s highest K/9 mark was 7.01 and these systems believe that 2014’s gains will be only marginally impactful on 2015. These projections incorporate an aging curve and Santana’s generally post-prime age is factored in.
Fangraphs Fans (an aggregation of fan submitted projections) were slightly more optimistic at 7.38, and FRIAS went much higher at 8.12. Why is FRIAS so optimistic about Ervin Santana’s strikeout rate? It bases its strikeout rate projection on the previous season’s strikeout rate, whiffs/swing, infield fly rate, fastball velocity and average horizontal movement. Santana’s 2014 whiffs/swing was far and away a career high and he placed sixth among qualified pitchers. Clearly, something special came together for him last year from a strikeout perspective. Whiffs/swing are fairly predictive on their own, as you can see in the graph below. Combining whiffs/swing and K% provides a model for next-year K% with a .5032 adjusted r-squared (based on the 2007-2014 sample of 20+ IP seasons).
The lesser substantive inputs of FRIAS, average horizontal movement and fastball velocity, were also favourable to Santana.
Walk rates are much more evenly predicted. Santana’s 2014 rate (2.89) was well in line with his career rate of 2.82. His F-Strike% was within 0.2 percentage points of his career rate. As a result, any weighing of past seasons and relevant data points would likely produce an expected per-nine walk rate of around 2.9.
The expected home run rates slant similarly to strikeout rate. FRIAS is the optimist while Fangraphs Fans is in the middle. Early in his career, Ervin Santana was a pronounced fly ball pitcher, posting a GB% below 40% in each of his first six seasons (the league average is consistently around 44%). But in the past four seasons, he’s had a roughly average batted ball distribution. Changes in batted ball distribution generally result from a change in approach or pitch usage and aren’t predictable. Pitchers tend to have fairly constant distributions throughout their career and so regressing towards career rate is generally a good approach. However, Steamer, ZiPS and PECOTA aren’t just using his career GB% to forecast an elevated home run rate, and another influence is his decreased strikeout rate per those metrics. A pitcher who strikes out fewer batters will put more balls in play and a portion of those balls will turn into home runs. ZiPS, at 1.21 HR/9, is surely integrating a specifically damaging factor into their analysis. Perhaps a remnant of 2012, when he allowed an absurd 1.97 HR/9, is involved.
I tend to lean somewhere in the middle when it comes to Ervin Santana, similar to what Fangraphs Fans produced. I believe that his 2014 whiffs rates were legitimate and should be significantly weighted. Santana’s fastball, slider and changeup all experienced career-highs in SwStr% whiff rates in 2014. I believe something changed fundamentally regarding his pitching ability. However, I believe it is sensible to assume a ~40% GB% from him going forward as well as a ~2.9 BB/9.
For a pitcher who generated little fanfare this offseason, I’m very much looking forward to seeing how Santana fares in 2015.