Why Cubs Loved Colin Rea, Through the Lens of New Baseball Prospectus Pitching Metrics

[Matthew Trueblood]

It was a mild surprise when the Cubs were willing to guarantee $5 million to secure the services of the peripatetic righthander. His style was a clear fit for the team's approach, though, and a new data stream gives us insight into that.

Image courtesy of © Jeff Hanisch-Imagn Images

Though he's diligently worked his fastball up into a range that would have been impressive a decade ago, Colin Rea is not by any means a power pitcher. He sits around 93 miles per hour with his four-seam fastball, and even his 90th-percentile velocity is only 94.3. He doesn't have exceptional shape on that pitch, either. He can't get hitters out by being fastball-forward—at least, not that one fastball.

The wrinkle, when it comes to Rea, is his ability to put a wrinkle in his heaters, in both directions. He has both a sinker and a cutter, to complement the four-seamer, and then adds a sweeper and the occasional curveball and splitter to the wide mix. When he's effective—and for long stretches of each of the last two seasons, he has been—he achieves that status by keeping hitters off-balance, changing both speeds and movement patterns.

Now, for the first time in the history of pitching analysis, we have some tools with which to affirmatively measure that skill and proclivity. Earlier this month at Baseball Prospectus, Stephen Sutton-Brown and the stats team introduced their suite of four new Arsenal metrics, designed to evaluate the interaction factors that augment the effectiveness of a pitcher's arsenal beyond the sum of its parts. Years ago, it was also BP who revolutionized the study and quantification of pitch tunneling—ways to disguise the incoming pitch as a different one or reduce a hitter's confidence and reaction time by denying them any certainty about what they're dealing with for as much of the flight time of the ball as possible. This new framework builds upon that one, but takes on a more complex and nuanced version of that dynamic. The stats are:

1. Pitch Type Probability: This is, roughly speaking, the confidence the batter can feel about diagnosing which pitch is coming at them, based on the release point and initial speed and trajectory of the ball out of the hand. Pitchers who mix pitches fairly evenly, especially if they do so while matching release points well and working in the same initial trajectory range, will score well. This also bakes in the count, because that's a key component of the hitter's estimate of pitch type probabilities.
2. Movement Spread: The size of the area a hitter will feel they must cover, based on the likelihood of a pitch being any of that pitcher's possible offerings and the movement patterns of those pitches. This rewards pitches that look the same out of the hand but don't take similar paths to the plate.
3. Velocity Spread: The same measurement of distributed probabilities, but with velocity. Obviously, this tends to be a good score if a pitcher can repeat their delivery and initial trajectory with two or three different pitches that cover a wide velocity band, like a four-seamer, a cutter, and a slider, or a fastball, a changeup, and a curveball.
4. Surprise Factor: A rating of how surprising observed pitch movement was, based on the distribution derived for Movement Spread. So, instead of measuring the size of a hitter's focal field, this emphasizes the ability to violate their expectations with a big movement away from what they think they see.

In some ways, these numbers tend to move together. In others, they tend to work against each other. We're all familiar with the paradox many pitchers face, but these put it into numbers: You want to make everything look the same, without letting everything actually be the same by the time it reaches the batter. That requires exquisite proprioception, feel for spin, and the ability to throw hard in the first place, as well as to take something off without slowing down the whole operation of the arm. As much as the modern game prizes velocity, there is no set of skills more important than this set. We just haven't been able to articulate, isolate and measure those skills very well, until now.

Colin Rea is a modern master of this stuff—especially Pitch Type Probability and Surprise Factor. In the former, he ranked in the 97th percentile among all MLB pitchers in 2024. In the latter, he was in the 93rd. He was more like average with regard to Movement (59th) and Velocity (57th) Spread, because he leans so hard on his three fastballs and they don't diverge all that widely in those measurements, but he was so assiduous about using each and mixing them evenly that hitters were usually set up for the wrong thing, and had a hard time squaring him up most of the time.

To get a sense of why and how this works, let's use Baseball Savant's 3D animations of Rea's arsenal. Here's (obviously, a digitized, strange abstraction of) what a right-handed batter sees from Rea in any given game:

The pink dots in the foreground, close along the trajectories matching each pitch type, are the commitment point, by which time a batter has to have decided whether and where to swing. The further white dots represent the recognition point, where a hitter can start to pick out a pitch type reasonably reliably against most pitchers. The furthest cluster is Rea's average release point for each pitch.

As you can see, there's not much separation here, even by the commitment point. Against righties, Rea threw his sweeper, sinker, four-seamer and cutter anywhere from 15 to 39 percent of the time last season, so hitters could never sit on any of them with much confidence, and the offerings each moved differently enough to dodge barrels or lock a hitter up for a called strike at times. He's also superb at matching his release points, as you can see if we look down the other end of this barrel. Here's a shot of his average pitch trajectories to righties, from a point just a few feet behind his release point, above the mound somewhere.

Most pitchers give hitters a bit better a cue than that. In fact, here's the same shot from behind Rea's invisible hand, but on pitches he threw to lefties.

As you can see, he works a bit more over the top with his curveball (in blue), which gives them a chance. Because he's trying to shape and locate differently with the other offerings, too, there are slightly larger separations between them, at release and throughout their flight. Here's what a lefty batter sees against him.

Against lefties, though, Rea is even more of a kitchen-sink guy, negating whatever losses he suffers from being easier to pick up out of the hand (or at least mitigating those disadvantages). He threw five different pitches at least 13% of the time against lefties last year, with only the curve being held for truly sparing use—and because that pitch has such a wide velocity differential and movement spread from his other pitches and is thrown so rarely, it contributes to that sky-high Surprise Factor.

Sutton-Brown noted in his introductory post about these metrics that they don't yet improve projections of pitcher performance. That doesn't mean they won't, but more fine-tuning (and perhaps more disaggregating, into platoon settings and such) is necessary to get there. For now, they're just qualitative measurements that describe a pitcher's style. However, in the case of Rea (a pitcher with so little in the way of wipeout stuff, who nonetheless has put together two straight seasons as a workmanlike, helpful backend starter), they can be very helpful, already.

View full article