PITCHf/x is an AWESOME tool, allowing us to analyze velocity, movement, location, and outcomes of every pitch a pitcher throws. This analysis relies on PITCHf/x correctly classifying every pitch, an imperfect science based on mathematical algorithms, which has been improved upon by the likes of Dan Brooks at BrooksBaseball.Net. One pitcher for whom this is especially difficult is Adam Wainwright.

Adam Wainwright has been masterful as of late, ranking 5th in xFIP (2.67) and 3rd in ERA (1.46) for all starters in his last five starts, a run culminating on Tuesday night with his complete game shutout of the Astros. Wainwright struggled at the beginning of the year, but has reestablished himself in the mold of the 2010, pre-Tommy John surgery ace we fell in love with. An obvious possible turning point is the May 22nd complete game shutout of the Padres, after which Derrick Goold wrote an article about how Wainwright brought back the four-seam fastball—a pitch he hadn’t thrown since Double-A. Goold notes that PITCHf/x tallied 34 cutters and 6 four-seams in that start, while Wainwright himself had it more like 4-5 cutters and 12-15 four-seams.

Everyone knows Adam Wainwright has a best-in-league curveball, but it is his ability to consistently locate a fastball low in the zone to both sides of the plate with movement that makes him great. Wainwright’s fastball has sat right around 90MPH this year with good arm-side-run (ASR) and sink. In some cases this may be the one-seam sinker Goold mentions in his article and in some cases it is a classic two-seamer. This fastball is complimented by Wainwright’s cutter, or whatever cutter-like pitch it is that he throws.

For many pitchers the context, hitter, and catcher location determine how he will throw a certain pitch, but for few pitchers is this truer than Adam Wainwright. In watching both Wainwright’s complete game shutouts this year, a few things are apparent. First, a “one” from Yadier Molina does not mean the same thing every time. I counted at least 3 times against the Astros and 15 times against the Padres in which Molina put down one finger and Wainwright threw something 86-92MPH with some cut to his glove side. The rest of the time—the vast majority of the time—Molina gives a “one” Wainwright throws some variation of his sinker, which moves the opposite direction. Some of those cutter-like pitches may be the illusion of the camera angle, but there is very clear cut to the 92MPH pitch Wainwright struck out Yonder Alonso with in the first inning of the Padres game that is mentioned by Goold in his article as an example of the four-seam fastball.

A four-seam fastball with cutter-like action is rare, and indeed PITCHf/x has Wainwright’s four-seamer moving slightly arm side (the other directions), but it is certainly possible that the pitch Wainwright brought back against San Diego does cut glove side. This might explain why PITCHf/x has Wainwright’s cutter velocity jumping from an average of 85.3MPH before May 22nd against the Padres to 86.8MPH in the starts since then. The tool is simply misclassifying harder four-seams as cutters, right? Like most things in life, however, it is more complicated than that, and Wainwright simply adding a pitch is not the whole explanation either for his resurgence or for variations in PITCHf/x data.

Returning to the importance of context and hitter, the second thing that is apparent in watching Wainwright is that which version of “one” he throws depends on the side of the plate on which Molina sets up. Without exception in these two games, when Molina set up inside to righties (away to lefties) Wainwright threw a sinker. The majority—though not all—of the times Molina set up inside to lefties, a “one” referred to something with a varying degree of cut to the glove side (in to the lefty hitter). Some of these may be four-seamers, but some are cutters in the 86-88MPH range with significant movement.

So, Wainwright throws cutters to one half of the plate and sinkers to the other and maybe some of those cutters are the new four-seam, which is a little harder. Case closed, right? Nice try…context, hitter, and location my friend. To right-handers a new monkey wrench is thrown into the mix, a “three” from Molina. Referred to by announcers when they see the sign as Wainwright’s cutter, the “three” from Molina indicates a similar pitch to what we see to lefties. It stays on the same side of the plate (away to righties, in to lefties), with glove side movement. Wainwright has the ability to vary this offering, however, depending on context. To some hitters early in the count it is identical to the cutter thrown to lefties, sitting around 86-88MPH and generating contact and groundballs. Later in the count, however, Wainwright can add depth and make it look more like a slider—a true “three” in normal pitch calling schemes—with a dip in velocity to 83-85MPH. In fact, in this ESPN interview Wainwright even refers to his “slider”, not his cutter.

To demonstrate this, check out the PITCHf/x charts below from BrooksBaseball.Net, which plot velocity on the vertical axis and horizontal movement on the horizontal axis (the view is from the catcher, so anything to the right of 0 on the horizontal axis indicates “cut” to Wainwright’s glove side). To righties (the plot on the right) we see cutters clustered together with good cut and less velocity. To lefties (the plot on the left) we see a combination of similarly moving pitches with a mix of both cutters and supposedly four-seam fastballs like the one to Alonso, that cut a little bit less and are a little bit harder. Whether he is throwing it as a four-seam fastball, a cutter, or even a sinker with different finger pressure, it is his ability to vary the speed and movement of a pitch cutting to his glove side, while still effectively locating it, that makes Adam Wainwright dominant.

With this in mind, I watched one of Wainwright’s ineffective starts from earlier in the year—May 12th against the Braves, just ten days before his complete game shutout of the Padres. There were a lot of differences—his curveball was not sharp and his overall command was poor—but what stood out was how ineffective his “cutter-like” pitches were. The velocity on these pitches was down, with 2/3 of them never reaching 87MPH. More importantly, however, the life and movement that defines his cutter was lacking and a surprising number of them did almost nothing at all.

In Adam Wainwright’s good starts this year—those in which he has given up 3 earned runs or fewer—he has had significantly fewer walks (1.76 BB/9 vs. 4.1 BB/9 in the 7 bad starts), given up fewer home runs (0.2 HR/9 vs. 2.7 HR/9), and had a significantly lower BABIP (.295 vs. .360). These numbers say the obvious, that he has been better in good starts than in bad ones. These numbers are driven, however, by differences in his cutter—the pitch that primarily generates weak contact, and gets him ahead in counts—than any problems in his curveball, which primarily drives his strikeouts; in good starts he has struck out 8.97 per 9 innings, in bad starts a similar 8.52.

What does PITCHf/x say about this theory? In those good starts Wainwright supposedly throws more four-seams and more cutters. Rather than being a true representation of pitch calling, this likely demonstrates both the addition of the four-seam later in the season and the fact that when his cutter is bad, it moves less, leading it to be incorrectly classified as a sinker rather than a cutter. Looking deeper, Pitch Type Linear Weights, which assign a value to a pitch every time it is thrown based on the outcome, demonstrate how important the cutter is. If a pitch’s value is below 0 that is bad, if it is above 0 that is good, meaning the pitch has been effective. In those good starts the Pitch Type Linear Weights value per 100 pitches for the cutter is .522. In the bad ones? -5.404. Translation: fixing the cutter (possibly by adding the four-seamer) has been imperative to Wainwright’s recent success.