Randy Dobnak the unlikely... ace?

By now, readers of this site are quite familiar with the story of Randy Dobnak, unlikely major league pitcher and so far in 2020, unlikely Twins’ ace. I won’t rehash his origin story in this piece, but if you aren’t familiar, I highly encourage you to read about him in these excellent pieces at other sites:

• Nick Stevens at Call to the Pen: Randy Dobnak has emerged as a valuable weapon
• Aaron Gleeman at the Athletic: How real is Randy Dobnak? (subscription required)

What I want to do today is offer a breakdown of Dobnak by putting his results thus far in context, assessing how he’s achieved those results through his arsenal and approach, and looking at what it all might mean going forward.

I think we’d all be forgiven for being cautious and holding our breath a bit when it comes to Dobnak. For as great as the story is and as much as we enjoy the success he’s had, in the back of our minds I think a lot of us wonder things like is this real? And, when will the bubble burst? That’s human nature. This kind of success doesn’t come from players with such low-profile pedigrees very often. But let’s try to be objective and see what the data tells us.

Career to date:

An undrafted minor league free agent signing, Dobnak found almost immediate and sustained success in his minor league career. In just shy of 300 innings, across five minor league levels 2017-2019, Dobnak posted a combined 2.57 ERA, 1.82 BB/9, and 6.54 K/9. His highest ERA at any stop was 3.14 over a full 2018 season at Cedar Rapids. The ERA and BB/9 are great marks but the K/9 is pretty pedestrian. But Dobnak is able to counter his relative lack of strikeouts with an ability to get groundballs – doing so at a rate above 45% at every stop. This rate increased in 2019 as he progressed through the high minors – 59.1% at Ft. Myers (High-A), 58.8% at Pensacola (AA), and 61.1% in Rochester (AAA).

Still, despite the minor league success, Dobnak debuted with the Twins with little fanfare. And all he’s done since coming to the big leagues at the end of 2019 is pitch better than he did in the minors. In 48.1 career MLB innings Dobnak has a 1.30 ERA, 1.86 BB/9, and 6.33 K/9. His MLB groundball rate is 58.5%. Incredibly and improbably, Dobnak’s MLB numbers are essentially the same or better than as his minor league numbers across the board.

Comparing to the rest of MLB

While Dobnak has performed consistently with his past performance we need to understand that performance in the context of other pitchers. In a vacuum, he’s been excellent – but how does he stack up with his peers?

In the last two seasons there are 220 pitchers that have thrown at least 30 innings as a starting pitcher. Dobnak ranks first in ERA at 1.37, a figure that is 72 points better than the pitcher in the 2^nd slot. Of those 220 pitchers Dobnak is the only one not to have allowed a home run. Considering the lowest career ERA in history for qualified starters is 1.82, an ERA that low likely isn’t sustainable. But still it’s impressive stuff!

But these are outcome stats based on a small sample size – only 39.1 IP across 9 starts. How do we know if it’s real or a fluke?

As many others have said before, including Gleeman in his piece linked above, for pitchers to have sustained success they need to be able to do some combination of the following:

• Get swings and misses (indicated by K/9 and swinging strike rate)
• Avoid hard contact, especially in the air (indicated by ground ball rate and HR/9)
• Don’t give away free bases (indicated by BB/9)

If a major league pitcher can do two of these things consistently they will have regular success. These are the main ideas behind defense independent pitching stats, most popularly summarized with Fielding Independent Pitching (FIP).

The most effective way for a pitcher to prevent runs is to prevent batters from putting the ball in play. It’s very hard for a pitcher to be hurt by swings and misses and called strikes. As noted above, Dobnak isn’t a standout in this area. His 6.18 K/9 as a big-league starter ranks just 198^th out of the 220. However, that’s not to say he doesn’t get swings and misses. A more granular data point is swinging strike rate, which casts Dobnak in a more positive light. Dobnak gets a miss on an above average 10.8% (89^th of 220) of the swings batters take against him.

The next best alternative to strikeouts and swings and misses is generating contact on the ground. Fly balls create extra bases (especially homeruns) and lead to quick runs allowed. Ground balls often create only singles; and it usually takes multiple singles strung together to create a run. Generating contact on the ground is our man’s calling card. His extreme groundball rate (59.3%) is third best overall, behind only Dallas Keuchel and Astros’ lefty Framber Valdez.

On the control front, free bases do nothing but hurt pitchers. Base runners are opportunities to give up runs. In this area, Dobnak has posted a 2.29 BB/9 as a starter. While this is higher than he showed in the minors, it’s still very solid and ranks 41^st.

Altogether, that combination of outcomes works out to a FIP of 2.77. Certainly, the FIP number indicates his ERA results have over-performed, but 2.77 is outstanding. In fact, that figure ranks fifth best in all of baseball, trailing only Max Scherzer, Tyler Glasnow, Jacob deGrom, and Gerrit Cole. That’s some impressive company. Even if Dobnak “regresses” as this underlying predictor suggests he should, he’d still be keeping company with arguably the games’ best pitchers.

Fantasy baseball writer Alex Chamberlain pointed out in a post last week on Fangraphs’ Rotographs blog that Dobnak’s combination of skills are rare. Chamberlain found there has been 39 qualified pitcher seasons in the last five years with a better than 50% groundball rate and better than 10% swinging strike rate. If we also account for Dobnak’s good control and add a third criteria of a walk rate less than 5.5% that list shrinks to just three seasons:

• 2019 Hyun-Jin Ryu (2nd NL Cy Young)
• 2015 Clayton Kershaw (3rd NL Cy Young)
• 2015 Jake Arrieta (NL Cy Young Winner)

I expanded the time horizon to the past ten seasons and found only 8 seasons meeting those criteria. The table below details those, and... it’s quite the list of names:

From this list, only Jeff Samardzija in 2014 did not receive Cy Young votes. It’s worth pointing out here that each of the seasons listed above covered more than 180 innings. So Dobnak, with only 48.1 career innings, has a long way to go before we can be sure he belongs in this company. But it’s illustrative of how unique and elite his success has been and an indicator that he has the combination of skills to sustain some of it.

How well equipped is he to continue on this path?

Dobnak’s Arsenal

Something interesting happened with Dobnak in the low minors. In his first three minor league stops in 2017 and 2018 his ground ball rate was in the mid-40s (save for a 7 inning sample at Cedar Rapids at the end of 2017). Since then his groundball rates have been well above 50% and usually around 60%.

Dobnak operates with four pitches from a low three-quarters arm slot: the sinker (46.4%, 91 mph), a slider (31.5%, 83 mph), a changeup (16.4%, 85 mph) and a four-seam fastball (5.7%, 93 mph). Against right-handed batters he’s relied more heavily on his slider (31.5% RHB / 25.4% LHB) and against lefties he’s brought in the changeup more (21.9% LHB / 10.1% RHB).

There’s nothing overly complicated about that approach and the velocity figures are OK but not outstanding (36th percentile per Statcast). Thanks to Statcast, we also have detailed data on how Dobnak’s pitches move relative to his peers. From it we can discern where Dobnak excels is in pitch movement, especially vertical movement on the sinker. As he pointed out to Gleeman, he learned a new sinker grip with the help of the Twins development staff in Cedar Rapids in 2018. And that new sinker is something special.

Of 2020 pitchers that have thrown 100 or more sinkers, Dobnak has the second most downward vertical movement, averaging 33.4 inches. The only sinker with more absolute downward movement belongs to Boston’s Ryan Weber at 33.9 inches, but his averages only 88.1 mph. That downward break is what generates the groundballs – yielding an average launch angle of -7 degrees and a ground ball rate of 78.1%. Batters can’t help but to hit the top half of the baseball and beat it into the ground.

The sinker alone is impressive and it also pairs well with his slider and changeup. On the right side of the chart above you can see the movement profiles of Dobnak’s four pitches in relation to each other. The slider is shown in yellow with 39.5 inches of vertical and 1.6 inches of horizontal break on average. Naturally, the sinker and the slider break opposite from each other. When you consider the sinker’s 15.3 inches of horizontal break to his arm side and the slider’s 1.6-inch break to his glove side, Dobnak can essentially move the baseball the full width of home plate (17 inches) from pitch to pitch.

You might also have noticed above that his sinker and changeup (orange and green dots on top of one another) are extremely similar in movement. This is an interesting point of deception for a pitcher. The two pitches will look the same in terms of shape and break, but the changeup comes in 6 to 7 mph slower in velocity, adding another challenge for the hitter to deal with.

Dobnak can maximize the effectiveness of these pitches’ characteristics and deception by throwing the changeup and sinker in similar locations (arm side, down) around the strike zone and splitting the sinker and slider (glove side, down) opposite of each other. That’s exactly what he has been doing so far in 2020:

The net result of these combinations is that Dobnak gets swings and misses with the off-speed pitches. While the sinker has only a 6.3% whiff rate, his slider generates whiffs 31.6% and the changeup 25%.

Dobnak has another peculiarity that helps make him successful and it’s one that you wouldn’t expect for a sinkerballer. Andrew Bryzgornia first pointed this out over at TwinsDaily last week. In addition to all the grounders, Dobnak also generates a weirdly high number of infield pop flies. According to data at Fangraphs, Dobnak’s career infield fly ball percentage is 13.3%. It doesn’t appear to be a statistical anomaly either – at every stop of his minor league career his infield fly percentage was greater than 13%. The major league average for infield fly balls is generally just below 10%. It’s worth pointing this out, even though it’s counterintuitive and small numbers, because it’s yet another way that Dobnak compensates for his relative lack of strikeouts. When more than 70% of the batted balls against you are on the ground or harmless pop-ups, you can have a lot of success with just a modest number of strikeouts.

2020 Adjustments

As he adjusts to the major leagues and continues to develop he’s continued to tinker with his approach. In his appearances at the end of 2019, his combined sinker and four seam fastball usage was approaching 60%, with about three out of every five heaters being sinkers. This year he’s mixed in his breaking ball and changeup more, reducing his combined usage of fastball varieties to 52%. He’s also modified that mix to rely more heavily on the sinker – nearly 9 out of every 10 fastballs he’s thrown have been sinkers.

That additional reliance on the sinker may be due in part to the fact that he’s found some additional movement on the pitch. The 33.4 inches of vertical break I mentioned above is almost 2 inches more than his 2019 average. The 15.3 inches of horizontal break is an inch more. In investigating that, I found that a likely cause of this additional movement is that he has also lowered his release point slightly:

Per BrooksBaseball.net, Dobnak’s sinker release point ranged from 5.4 to 5.6 vertical feet in 2019. So far in 2020, its been 5.2 and below. His other offerings have had a similar drop in release point. The change isn’t huge and it’s hard to know for sure if it was intentional. Pitchers experience slight variation in their releases all the time. But this may help explain the subtle addition of movement to his sinker and will be worth monitoring as the 2020 season continues.

Looking Forward

Of course, it seems silly to think Dobnak will maintain an ERA of 1.30 for the rest of his career. As the 2.77 FIP and other ERA estimators point out, Dobnak’s ERA will almost certainly increase over time. Those analytic tools all peg him somewhere between 2.77 and 4.09. Fangraphs’ hosted future projections by Dan Szymborski’s ZiPS tool forecasts 4.33, 4.20, and 4.18 ERAs for Dobnak for 2020, 2021, and 2022. Before you come away disappointed by those figures, consider that the Major League average ERA for starting pitchers in 2019 was 4.54. In the American League it was 4.76.

The point is, by almost any measure you can find, Randy Dobnak looks like a legitimate above average major league starting pitcher. That’s an incredibly valuable thing for the club, especially when you consider that Dobnak essentially came out of nowhere and remains under team control until 2026.

And who knows, Dobnak might very well continue to exceed expectations. After all, he wasn’t supposed to be here, or this good, anyways. He’s already demonstrated an ability to make adjustments that maximize his skills and chances for success. His arsenal and approach seem to suggest there is staying power here.

If he continues to coax obscene rates of ground balls, keeps the ball in the park, and avoids giving away free passes, chances are pretty good that he’s going to be good for a while. The track record of the data and the profiles of his pitches make it seem likely he’ll continue to get grounders. If he can continue to throw strikes, perhaps this is a mid-rotation major league starting pitcher. If he continues to develop and finds a few more strikeouts, maybe the ceiling is higher still.

It’s no longer a question if Dobnak belongs in a big-league rotation, it’s now a question of how good he can be.

---

John is a contributor to Twinkie Town with an emphasis on analytics. He is a lifelong Twins fan and former college pitcher.