The Seek for the Most Predictable Pitcher

Within the pre-PitchCom period, main league groups had extra rigorous protocols for shielding their indicators than your financial institution has for securing your account. It wouldn’t shock me to study that some groups’ customized PitchCom audio clips are learn in a modified pig latin created by a pitching technique staffer. That the hitter doesn’t know what pitch is coming is taken into account an enormous benefit for the pitcher. And it’s not solely pitchers who suppose so — simply ask the 2017 Astros.

Signal-stealing apart, hitters stand within the field pondering which pitch would possibly come hurtling their method mere seconds later. What that pondering seems like is determined by the hitter. There’s Nick Castellanos and his “glorified batting apply” strategy, wherein he seems for the ball and hits it as laborious as he can. However there’s additionally Carlos Correa, who begins his day finding out pitcher tendencies within the video room.

For his or her half, pitchers set the issue stage on the hitter’s guessing recreation. That phrases like “fastball rely” and “pitching backwards” exist inform us that pitchers observe (and, at instances, purposefully upend) typical techniques to sequence their pitches, and consider that sure pitch sorts are optimum in sure counts. Methods change into customary practices as a result of they’re efficient, however an over-reliance on one or two methods can result in predictability. Turn out to be too predictable and a pitcher successfully units their opponents’ guessing recreation on “simple” mode. However does making it simple for the hitter to take a seat on a sure pitch mechanically make the general process of hitting simpler? Does preserving a hitter guessing at all times guarantee efficient pitching?

To determine how predictability components right into a pitcher’s total technique, we have to know which pitchers are setting the guessing recreation on “newbie” and which pitchers set it on “knowledgeable.” We will measure the issue of anticipating a given pitcher by taking part in a strictly by-the-numbers guessing recreation and seeing the way it goes. Which means that for every pitcher, if each hitter stepped as much as the plate realizing which pitch the pitcher throws most incessantly in all doable counts and guessed accordingly, how usually would the hitter be right? I took the success charges of a hitter utilizing this guessing technique on every pitcher’s tendencies from this season and mapped these charges to a Predictability Rating between 0 and 100, the place a rating of 100 is probably the most predictable and a rating of 0 is the least predictable. Trying simply at pitchers with a minimum of 100 innings pitched up to now this season, the highest 10 most and least predictable are listed beneath:

The rankings spotlight one pretty apparent impact. Pitchers who primarily throw two pitches are far simpler to foretell than those that throw 4 or 5. Justin Steele throws 60% four-seam fastballs and 30% sliders, with the opposite 10% cut up between his sinker, changeup, and curveball. However, Seth Lugo throws a proverbial kitchen sink composed of 26% curveballs, 25% four-seamers, 19% sinkers, 13% sliders, 6% cutters, and 6% changeups. And but, their abstract stats are eerily comparable:

Although they make use of far totally different instruments and techniques, the end result is comparable. The spectacular high quality of Steele’s slider, paired with strong fastball command, gives different means for him to maintain hitters unsure even when they accurately guess the pitch kind. For Lugo, neither of his essential fastball choices grade out notably effectively in accordance with Stuff+, so fairly than keep on with the traditional knowledge that leads most pitchers to throw fastballs a minimum of 50% of the time, he leans on two above-average breakers amid an array of pitches that he locates effectively and throws in quite a lot of counts, successfully establishing a god-mode guessing recreation for the hitter.

However since declaring two-pitch pitchers simple to foretell and five-pitch pitchers extra of a thriller isn’t precisely revelatory, let’s go forward and management for the issue of the preliminary problem, or what number of pitches the hitter is guessing between. As a substitute of trying simply at how usually a pitcher throws every pitch in every rely, I in contrast the precise frequency to the anticipated frequency if the pitcher have been equally prone to throw any pitch in any rely, i.e. most unpredictability. Solely pitches with a ten% utilization charge or greater have been included to restrict the scope to choices a hitter would truly must preserve entrance of thoughts whereas within the field. So though Lugo throws six pitches, the cutter and the changeup don’t get sufficient play to make the lower. With that in thoughts, Lugo is least predictable if there’s a 25% of him throwing any of his 4 most-used pitches in any rely. Taking absolutely the distinction between his precise count-level utilization charges and 25% gives a measure of how far off he’s from optimum hitter confusion. Averaging these variations throughout all counts (weighted by how usually he finds himself in every rely) and mapping to the identical scale used beforehand offers us a metric for total comparability throughout pitchers:

On this framing of the query, Andrew Abbott is probably the most predictable along with his 54% four-seamers, 19% sliders, 16% changeups, and 11% curveballs. Regardless of his predictability, he charges as above common with an 85 ERA- on the season, although his 116 FIP- casts some doubt on the proceedings and makes an argument for rising his slider utilization, each within the identify of preserving hitters guessing and throwing your greatest pitch extra.

Lugo nonetheless charges effectively for his lack of predictability, however maybe extra stunning is two-pitch Dylan Stop slotting in at fifth within the rankings. Stop combines to throw his fastball and slider 90% of the time, however stays powerful to foretell by splitting that 90% utilization virtually precisely 50-50. Clearly, the important thing to Stop’s success with simply two pitches is that each pitches grade out extraordinarily effectively, however throwing them equally usually and tunneling them out of a constant launch level amplifies the affect. That stated, preserving hitters guessing solely goes up to now in case you have a case of late-career Patrick Corbin in your fingers.

In the meantime, though five-pitch Zack Wheeler is roughly as predictable as three-pitch Freddy Peralta, the quantity of data to course of on the way in which to understanding and appearing on these tendencies is larger for Wheeler than Peralta. The scale and form of the sport planning course of varies based mostly on the dimensions of the arsenal. For the reason that variety of choices in a pitcher’s stock units the preliminary issue stage for hitters’ predictions (which they’ll then toggle up or down based mostly on utilization), let’s additionally take a look at the uncooked predictability scores adjusted not for arsenal dimension, however fairly grouped by the quantity pitches within the utility belt:

Steele and Stop bookend the two-pitch pitchers by way of prediction success charge, however the hole between Metal and everybody else is at risk of getting sued for trademark infringement by John Fisher. The pitchers in the course of the two-pitch pitcher leaderboard all situate themselves nearer on the spectrum to Stop than Steele. On the one hand, whereas pitchers with simply two major pitches have made a option to lean closely on these choices and possibly wouldn’t accomplish that in the event that they weren’t assured in each of them, mixing each pitches in frequently, always, is clearly a part of the technique as effectively. The truth is, most of the two-pitch hurlers are much less predictable than their friends with three and even 4 pitches.

Luis Severino leads the three-pitch group in maximizing guessing recreation trickery, whereas Kyle Harrison prefers to depend on stuff over secrecy. The four-pitch crew is led in predictability by Abbott and unpredictability by Lugo. Logan Gilbert is probably the most predictable amongst the five-pitch crowd, whereas the Nick Martinez code is harder to crack.

That the least predictable five-pitch pitchers are extra predictable than their four-pitch counterparts is attention-grabbing, however most likely speaks to the luxurious inherent in having so many choices to deploy. With 5 pitches, one or two choices could also be reserved for particular conditions based mostly on rely and hitter handedness. Such a method makes a pitcher extra predictable, however so long as the pitch stays efficient, the commerce off is price it.

Notably absent from each leaderboard up to now is an apparent divide in high quality when evaluating probably the most predictable to least predictable. Although staying unpredictable is a instrument working within the pitcher’s favor, it’s clearly extra of a “good to have” than a “will need to have.”

We’ve additionally solely measured predictability within the combination to this point, giving us a common thought of a pitcher’s power or weak point in clinging too tightly to a set sample of habits. However typically even the strongest opponents possess that one hyper-specific weak point, the deadly flaw that when exploited permits a mediocre gamer to overcome an unattainable last boss. Are there particular counts the place sure pitchers change into so predictable {that a} hitter may simply exploit their one-note strategy?

For the reason that breadth of pitch choice technique narrows because the rely deepens, it felt logical to start out with three-ball and two-strike counts (ignoring full counts as a result of these are a completely separate beast by way of pitcher strategy). In three-ball counts, the pitcher is restricted to no matter pitches he feels he can land within the zone, whereas a pitcher’s benefit in two-strike counts removes the need to serve up apparent strikes or something the hitter would possibly discover enticing and as an alternative incentivizes throwing flexible stuff simply exterior the zone within the hope of getting a swing and miss. Let’s check out the leaderboards:

When pitchers get predictable in three-ball counts, it’s within the actually apparent and boring method. They pump fastballs at an obscene charge, and it doesn’t appear to harm them total anymore than getting right into a three-ball rely hurts them within the first place. The aforementioned Abbott — a.ok.a., probably the most predictable pitcher in three-ball counts — throws his four-seamer in 100% of 3-0 counts and 92% of 3-1 counts. Among the many high 5 on the leaderboard, most are already throwing their fastball round 55% of the time. The exception is Corbin Burnes, who takes his typical cutter charge and buys it a health club membership and a few protein powder for bulking season, permitting him to develop his cutter utilization from 44% total to 95% in 3-0 counts and 84% in 3-1 counts.

What’s extra attention-grabbing are the pitchers who handle to be much less apparent about their three-ball strategy. Like everybody else, Matt Waldron throws extra fastballs as soon as he’s acquired three balls to his identify, however that really serves as a departure from his common strategy, which makes him much less predictable. Waldron’s most used pitch is a knuckleball he throws 38% of the time, however in three-ball counts, the knuckleball all however disappears and he splits utilization between a four-seamer, sinker, and cutter. His highest utilization charge on any pitch is 58% four-seamer in 3-0 counts.

Although Kyle Gibson, the second least predictable pitcher, doesn’t have a knuckleball to ditch, the remainder of his technique for preserving hitters guessing is roughly the identical as Waldron’s: throwing a number of fastballs. Of Gibson’s 4 most used pitches, three are fastballs. So even when he’s fearful about touchdown his sweeper for a strike, he can nonetheless make hitters guess between a four-seam, a sinker, and a cutter.

When score pitchers on predictability with two strikes, there’s a big hole between first place and the remainder of the sphere. Sonny Grey transforms from a pitcher with 5 pitches that he throws greater than 10% of the time and none that he throws greater than 25% of the time right into a pitcher whose obsession with sweepers rivals the web’s obsession with Shohei Ohtani’s canine. In 0-2 and 1-2 counts, he throws his sweeper roughly 73% of the time. In 2-2 counts, he chills out with the sweeper a bit bit (45% utilization) by reintroducing his sinker into the combo (28% utilization). It’s laborious to argue towards his strategy since in plate appearances with 0-2, 1-2, or 2-2 counts, hitters are posting wOBAs of .124, .139, and .230, respectively:

On the much less predictable aspect of two-strike counts, Max Fried is the least prone to be overly reliant on the traditional knowledge that requires pitchers to nibble on the corners with their greatest breaking pitch. As a substitute, Fried adheres to a different in style baseball cliche and “stays inside himself” by preserving his utilization roughly the identical as his total numbers. He shaves a couple of proportion factors of utilization off his four-seamer and allocates them to his curveball, however in any other case he goes about his enterprise as he would in another rely, utilizing 5 complete pitches a minimum of 10% of the time fairly than letting hitters assume they’re getting a curveball or slider out of the zone.

Outdoors of the counts that affect utilization in an intuitive method, one explicit pitcher in a single explicit rely stands out as being particularly predictable. Joey Estes throws his four-seamer 53% of the time total, however in 0-1 counts, when he’s already gotten one up on the hitter, he nonetheless insists on going to his fastball 66% of the time. It’s his third most frequent fastball rely after 3-0 and 3-1. He makes use of his fastball extra incessantly in 0-1 counts than he does in 1-0, 2-0, 2-1, or 3-2 counts. I’ve no guesses for why Estes does this, however I wager opposing hitters see a observe about it of their pre-game experiences on the nights he’s set to take the mound.

Having now sliced and diced the information a bunch of various methods, we’re on the a part of the article the place it’s time to log out with some concluding ideas. This isn’t the kind of article that’s going to finish with some neatly wrapped piece of actionable recommendation to assist pitchers hone their craft, and I sort of desire issues that method. Typically analytical analysis is responsible of so totally optimizing technique that it will be disadvantageous for a group or participant to behave opposite to the newly found, maximally useful strategy — even when meaning a extra boring, monochromatic model of baseball. However typically analysis reveals that multiple technique can work — that perhaps dozens of methods can work — and that methods seemingly in opposition with each other can work. What makes a selected technique optimum is determined by the particular group, participant, and circumstances in query, and understanding these parts is the important thing to picking one of the best strategy.

How cool is it to look at a sport the place beginning pitchers could be profitable throwing two pitches or 5 pitches? How cool is it to have a sport the place each Justin Steele and Dylan Stop can exist and thrive with their disparate utilization patterns? How cool is it that some pitchers strategy three-ball counts by going all-in on one fastball, whereas others strategy it by going all-in on all of the fastballs? How cool is it that baseball has room for Sonny Grey to go ham with sweepers when he will get to 2 strikes, but additionally house for Max Fried to roll out his total arsenal in those self same conditions? How cool is it to have a unusual little man hiding out on the Oakland A’s, throwing 66% fastballs in 1-0 counts and hoping nobody notices?

For those who ask me, it’s all extremely cool.