Analytical Breakdown of Zach Davies’ 2019 Season using R

Image for post
Image for post

Right-hander Zach Davies posted some very noticeable splits in 2019, pitching to the tune of much better numbers in the months of April, May, and September, while struggling in June, July, and August. Using the “Baseballr” package in R, this breakdown will attempt to uncover what Davies did differently in the months he struggled in, compared to ones with success.


The first aspects to look into were his pitch velocities, specifically his primary pitch. Davies has a pitch arsenal consisting of a Sinker, Changeup, Cutter, Curveball, and Slider.

Image for post
Image for post

Zach Davies is a sinker baller, and sinker ballers tend to pitch to contact and let their defense do the job, so the average velocity of 88.4mph isn’t concerning, as his pitching style isn’t to blow by hitters. It is interesting that in the three months he was successful, he actually was right on his average velocity, immediately dipped below, then added over a mile per hour in the last month of the season, when pitchers typically get tired.

Image for post
Image for post

Here we find his full arsenal’s velocity throughout the year, and his cutter, slider, and changeup all increase in velocity on top of his sinker as the season unfolds, which isn’t typical for a starter who throws over 150 innings in a given season now days. You can also see he actually stopped throwing his slider after July, which is a pitch he uses very sparingly anyways. There doesn’t appear to be any correlation between his pitch velocity and his success/struggles, but we will keep his arsenal and usage in mind.

Pitch Usage

Classifying April, May, and September as his “Effective Months”, and others as “Ineffective,” Davies, unsurprisingly, was more successful in 2019 when using his sinker over half the time. The main successor of his arsenal when decreasing his sinker usage was his cutter, which was almost used 10% more of the time during his ineffective months. Evidently, Davies really only uses his curve and slider as changes of pace every once and a while to keep hitters honest, so his arsenal essentially turns into a sinker, changeup, and cutter after this observation.

To break this down even further, Davies’ pitch distribution actually lines up with his effective outings when facing right-handed batters, no matter the month. Davies, much like many other righties, struggles more against left-handed hitters, posting an opponent’s wOBA 15 points higher than vs righties. During his ineffective months, his cutter usage against lefties raises all the way up to 20.8%. There does seem to be a direct correlation between his cutter and his struggles.

Pitch Locations

Image for post
Image for post

First, it’s important to note what Davies’ overall tendencies are. As this figure shows, he pounds each corner of the lower half of the strike zone, particularly down and in to righties, which makes sense as a sinker baller attempting to induce ground balls, where hitters from both sides of the plate are intended to roll over on.

With this in mind, we then take a look at the pitch in question:

His cutter.

Davies’ Cut Fastball via
Image for post
Image for post
Each cutter thrown denoted by white points

The first and obvious observation here is that this completely backs up the findings earlier of the fact that he used his cutter much more in his ineffective months. It appears he tends to try and throw it in the same location for batters from each side of the plate, being the first base side of the strike zone. This is meant for righties to roll over, and to jam lefties.

In his effective months, the heat map implies that he hits the black more often, as the pitches vary vertically along the strike zone on the first base side. However, in the ineffective months, the cutter, while still centered in that general area, leaks out over the plate much more and is generally all over the place, leading to more walks and hittable pitches.

As mentioned in the pitch distribution section, he throws the cutter much more to lefties, especially when he struggled.

Image for post
Image for post

Breaking his cutter usage down during the ineffective months even further, he barely threw it to righties (4.49% of the time), but threw it to lefties 20.8% as stated above. You can see that the most of the reason for his poor cutter location overall is due to his inability to locate it against lefties.

Situational Pitching

The importance of starting a hitter off with a strike is often emphasized, because the difference between the amount of hitters and pitchers counts can ultimately decide outings.

Zach Davies’ can be visualized in terms of OPS+:

Image for post
Image for post

Most of this tile plot is as expected, except for the fact that Davies actually performs worse when the plate appearance goes through a 2 ball count. He allows a 72% higher OPS to hitters he faces when the PA goes through a 2–0 count. As expected, a 3–0 count still produces poor numbers, as its the ultimate hitter’s count. However, he is worse in 2–1 counts than he is in 3–1 counts which is surprising. From these observations from his 2019 season, we can conclude for now that Davies needs to avoid 2 ball counts unless he has 2 strikes on the hitter. Also of note, is that there is an 81 point differential in OPS+ when starting a hitter off with a strike rather than a ball.

To further narrow down what factors lead to some struggles, matrices of pitch distribution based on counts can be observed.

Image for post
Image for post
Davies’ pitch use for each count in both effective/ineffective months vs R & L hitters (%)

There’s an overwhelming amount that can inferred from these charts. First, Davies throws his sinker much more to righties, especially during the effective months. He’s a sinker baller, thats his bread and butter. The focal point is the 2–0 and 2–1 counts to see what he was throwing on those to make those his worst counts. When effective, in 2–0 counts, he throws the sinker 64% and 48% to righties and lefties, respectively, while the cutter usage increases 7% to lefties from effective to ineffective months, which aligns with the underlying theme of his cutter being overused, especially to lefties.

In almost any situation, Davies’ struggles directly point to increased usage in the cutter, and less in the sinker, again, particularly vs lefties. In 0-2 and 1–2 pitcher’s counts, the changeup and sinker are clearly his preferred put-away pitches against left-handed batters, yet the changeup usage drops significantly in these counts when ineffective to lefties, leaving room for more error with the cutter. When ineffective, he had a higher percentage of 3–0 counts to righties, resulting in more walks to the side of the plate he is stronger against, and he needs to take care of business vs righties.

Something to point out, according to Baseball Savant, Davies’ cutter actually breaks 6.9 inches (horizontally) less than league average. Thats 300% less than the average MLB cutter, meaning he doesn’t have the most effective cut fastball. His sinker however, breaks 10% more horizontally than the avegrage MLB sinker, and his changeup breaks 1% more and 6% more than average, making these two by far his best pitches.

The Sinker:

Sinker via

The Changeup:

Changeup via

Batted Ball Data

To reiterate, Zach Davies is a sinker baller, and his goal is not to strike every hitter out, but to induce weak contact, specifically on the ground. Davies is in the 2nd percentile in Whiff%, which is bottom of the barrel. His pitches do not get swung and missed at very often, so when his pitch selection falters, hitters are both not chasing anything out of the zone and crushing the hangers.

Image for post
Image for post

In his ineffective months, we again see evidence of less sinker usage, with his ground ball % down to less than 40% and less than 35% against lefties. Aside from popups (which are clearly the rarest batted ball type), his ground balls have the lowest xwOBA and xBA based on their launch angle and exit velocity, meaning he needs to keep the ball on the ground and increase his sinker usage to stick to what he knows. His line drive % goes way up as his does his cutter usage. During his ineffective months, his cutter has an xwOBA of .408 overall, and.459 vs lefties.

On top of this, Davies’ curveball, while only used about 3.5%, tends to get crushed too, given its .649 xwOBA when ineffective, and .436 xwOBA overall, albeit a small sample size.

Taking a look at specifically cutters and curveballs that were put into play throughout the 2019 season, it is evident that he was leaving these pitches right over the middle of the plate to be launched. In the spraychart, it appears that hitters were timing these two pitches very well, as indicated by the flyball contour being dead center, as well as more evidence that he tends to favor lefties with these pitches, given the density of ground balls to the right side. Focusing on left-handed hitters, his average exit velocity off the cutter is 6mph higher in ineffective months than effective, and launch angle jumps up 5.5 degrees, going from low liner to screamer.

His Curveball:

Curveball via

Release Point

As far as Davies’ release point, it’s clear that during the ineffective months that he actually drops his arm slot a bit. When right, he releases the ball a little above 5.6 feet off the ground, and hovers around less than 1.75 ft horizontally. When ineffective, his arm slot dips to below 5.5 ft vertically and stretches out to 2.1 ft. This may be a product of fatigue, among other factors, but there is a clear difference in his arm slot when dealing, and when ineffective.

Another aspect to point out is that one of his less effective pitches, the cutter, is almost always released at the ineffective arm slot, and its possible that it’s all correlated. One more thing to note is that there is always a possibility a hitter could pick up on the arm slot quickly and guess cutter more often than not.

Expected Stats vs Actual

On the surface, Zach Davies posted some of the best numbers of his career, setting a career best 3.55 ERA, and a .250 BAA. To find more about luck vs skill, we take a look at his actual and expected stats:

Image for post
Image for post

Davies actually outperformed all his expected stats, with his ERA having the largest deviation. Just looking at the right-most table, it is evident that luck played a massive role in his success. It appears batted balls that should normally be hits just weren’t falling, and that he may have shaved a full run off his ERA as a result.

Image for post
Image for post

This is a graph showing the relationship between ERA and FIP for pitchers who threw more than 150 innings in 2019. The way it reads is if a pitcher is on the left side of the line, they may have been getting some luck on defense and with batted balls, and they’re on the right side of the line, they were possibly getting unlucky. This basically means the further away you are from that line on the left side, the higher the deviation is between your ERA and FIP. Davies has the second highest deviation from the line, meaning his level of (possible) luck was the top in the league, and could very well regress.

More Stats Deep Dive

Davies saw his lowest K% of his career in 2019, which isn’t devastating given he’s a pitch to contact type pitcher. However, he also saw his GB % plummet to 40%, when he normally hovers in the high 40%-low 50% range, meaning the sinker baller wasn’t inducing as many groundballs as sinkers are intended to. He saw the highest Fly Ball% of his career, meaning his pitches were getting lifted much more either due to location or movement. His Pull% was down, his Oppo% was the highest of his career, and his soft contact% was the lowest of his career as well, aligning with the fact that hitters aren’t rolling over on his pitches as often, and are going with the pitch. His barrel% was way up, and his average launch angle on balls in play increased to 12.8, and unsurprisingly again, were the highest of his career. He set a career high in contact%, and O-contact%, meaning hitters were not missing on balls out of the zone, fouling off pitches that should have put them away, lining up with the low whiff%. One thing that stayed constant were his walk rates. He also saw the most shifts in his career, despite the lower pull% and GB%, but this most likely helped him as roughly 70% of his ground balls were pulled.

Also of note, his changeup is by far his most effective pitch, producing a xwOBA of .312, a whiff rate of 29%, and a putaway rate of 16.7%, all highs of all his pitches. Davies only uses this pitch roughly 30% of the time, and it’s looking like that needs to increase.

The fact that Davies has the combination of outperforming his expected statistics and is setting career highs (and lows) in stats he isn’t supposed to, could mean that he is due for a possible regression, and could watch that solid 3.55 ERA creep north.

It will be interesting to see how the division change does for Davies, going from having half his starts in Miller Park in Milwuakee (park factor of .976 runs, 17th in MLB) to Petco Park in San Diego (park factor of .860 runs, 28th in MLB), although he will see the most hitter friendly park in the bigs more often in Coors Field in Colorado (PF of 1.394 runs).


Things that can help Davies avoid a possible regression is to increase the usage of his changeup. It’s by far his most effective pitch in terms of producing whiffs and outs, with above average movement. Incorporating this even more can eventually translate into even more swings and misses, unlike his slider (which he ultimately stopped throwing half way through the year), and his curveball, which he may want to think about using even more sparingly than he already does, despite its above average spin rate. Davies may want to think about developing a new pitch, however, that is much easier said than done. Another aspect is to limit the amount of cutters he uses, especially to lefties, and avoid those dreaded hitters counts. He can go back to his roots as a true sinker baller and keep the ball on the ground more by increasing the usage of it, particularly against lefties. He can also be a bit predictable on certain counts, so mixing up his pitches is key.

Check out the 2020 follow up analysis here :

Citations and Data Usage

“Baseballr” & “Lahman” packages in R/RStudio

Other Projects

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store