In a previous article, I looked at six pitchers who might be curveballers in the 2017 Lance McCullers tradition. Now, I’m going to take a closer look at the most promising four pitchers I identified last time–Parker Bridwell, Jimmy Nelson, Chris Stratton, James Shields–and see whether each really doec have what it takes to be the next great curveball pitcher.
Quality of Pitch Metrics
Fangraphs pitch values can be a valuable method for evaluating pitch quality. I used them in the last article as a way to identify pitchers whose curves had prevented a significant number of runs despite below average usage. They are descriptive, though, rather than predictive; they can gauge the results a pitch gets, but they can’t always gauge how effective the pitch will be going forward.
But we can try and measure the quality of a pitch based on stuff rather than results. The folks at QOPBaseball have developed proprietary Quality of Pitch statistics that attempt to do just that. QOP (quality of pitch) metrics incorporate several components–velocity, location, vertical break, horizontal break, lateness of break, and rise–into a context-independent statistic that measures pitch quality on a 0-10 scale.
For our sample of pitcher seasons with heavy curveball usage, the curveball QOPA (quality of pitch averages) were both above the league average for those pitches for 2015-2017.
James Shields’s curve, at least by this metric, is lacking. Jimmy Nelson’s and Chris Stratton’s curveballs hover just above the sample average and well above league average. Parker Bridwell’s figure really stands out, though, so it’s worth investigating.
QOP data shows percentile ratings compared to the rest of the league for each component of QOPA. Let’s compare Bridwell with McCullers.
(Note that because McCullers’ pitch is classified as a knuckle-curve and Bridwell’s as a curveball here, McCullers has a higher QOPA percentile even though his average is lower).
Both pitchers have similar quality-of-pitch averages. There are pretty stark differences elsewhere, though, as only their location ratings are similar. McCullers does best in velocity, and Bridwell is rise. Velocity, as I noted last time, is pretty important to heavy curveball usage. But I’m not sure what role curveball rise plays. I also don’t know how similar to McCullers’ a curve needs to be for heavy pitch usage, especially if both pitches get good results.
Even if pitch quality average doesn’t rate Nelson, Stratton, or Shields as highly, it’s worth seeing whether their curveballs are more like McCullers’s.
Honestly, save location and velocity, Shields’s curveball looks the most similar. If he only threw it harder, he might be the best bet here. Stratton’s is also much slower, and these numbers really illustrate how much less movement he gets than McCullers (I believe that lower numbers mean more movement here). Nelson has similar velocity and vertical break, which, as we found last time, contribute to higher strikeout numbers. He also seems to locate the ball even better than McCullers does, which can’t hurt.
Context-independent factors like these are a good way to evaluate whether our candidates’ curveballs will play when used over a third of the time. We noted here and last time that things like velocity and movement are crucial. Still, other context-dependent, and even results-based, metrics are helpful, too.
Ideally, we want a curveball pitcher to induce bad plate discipline numbers from hitters. Very generally, swings outside the zone are preferable inside swings. Same goes for curveballs. League-wide, wOBA drops over .200 points to 0.163 on curveballs swung at outside the zone compared to in.
With that in mind, let’s look at the differential between swing rate inside versus outside the zone.
As you can see, the pitchers in the curveballer sample get a much higher swing percentage on their curves both inside and outside the zone than league average. There’s also a larger differential between inside and outside swings. This holds true with our four candidates.
I don’t really know whether getting more zone swings is a good thing. I guess it means fewer called strikes. The wOBA figures for our curveballers on swung-at curveballs in the zone are quite spread out, so I’m not really sure what to make of it. Maybe hitters are just swinging more because their seeing the pitches more. What matters most, though, is whether they’re missing. Let’s look at whiff rate.
Everyone here gets more whiffs, inside and outside, than league average. The curveballer sample, though, has an average chase rate much higher than league average and way higher than any of the candidates. Only Bridwell had a higher chase rate than any of the curveballers, and only two of them. I don’t know whether a high chase rate is the key to curveballers success, but I expect that it’s a big part.
Stratton and Nelson are noticeably at the bottom in both whiffs and swing rate. Hitters are just taking their curveballs more often and hitting it more when they do swing. What they lack in whiffs they’ll have to make up for in weak contact.
Shields’s differential is low, yet, interestingly, his swing differential is the biggest of the candidates. Batters swing at his curveballs in the zone more often, but they whiff more or less the same as his out-of-zone curveballs. This is good for zone pitches; his swinging strike rate there is the best on the board. His chase rate, though, still isn’t good even if it’s better than Stratton’s and Nelson’s. Keep up the good work in the zone, just get more batters to wave at your outside pitches.
Where are curveballs most effective? Location is especially important when a pitch is being thrown over a third of the time. Movement and velocity are useful, but if a pitcher lobs meatballs over the plate, they’re going to get smacked.
Looking at league-wide data through the 2017 season, curveballs have tended to be most effective when thrown low and away, outside the zone. Putting aside areas where relatively few pitches were thrown, I’ve outlined the best locations for both left- and right-handed batters in the heatmaps below.
While plenty of situational considerations determine optimal location at a particular moment, curveballs located in this area tended to have the lowest slugging percentage per pitch last year. So, let’s take a look at how often our candidates located their curveballs in the ideal area.
All are around or above the league-average mark except Nelson, who tends to throw around the same locations regardless of the batter’s handedness (low and inside to lefties and low and outside to righties). Every other pitcher’s heatmap pretty well lines up with the graphic above: best when thrown in the ideal area, and thrown there often. I won’t inundate you with eight different heatmaps, as the results are good and relatively uniform. That’s encouraging!
Quality of Contact
Relatedly, we’d like our pitchers to induce low-quality contact. Ideally, the pitcher wants curveballs hit softly on the ground. So, we want him to run low average launch angle and exit velocity.
Average exit velocity and BABIP were worse than league average for the curveballer sample, but everything else was better. Exit velocity isn’t as important when more balls are being hit at low angles; it just means you’ll get a bunch of harder ground balls rather than softer ones. It puts more pressure on the defense, but it’s something that can be compensated for. And, judging by regular and expected wOBA, that is the case. The curveballers get many more swings and misses than the league, so the higher BABIP isn’t as big a deal.
Two candidates, Stratton and Nelson, showed excellent launch angles of 0.4 and 3.8 respectively, far lower than both league average and curveballer average. Nelson, on the other hand, runs a fairly high exit velocity. His regular batting average is quite low (.190), so maybe his curve is successful because hitters just don’t put it in play very often. Sure enough, his rate of in-play curveballs was significantly lower than league average and the lowest among the candidates. His home run rate was similarly miniscule. If hitters got more looks at his curveball, I’m not sure he could keep it out of play as often. He’ll have to keep his exit velocity down, as he doesn’t have quite the same swing and miss to make up for it as the curveballers do.
Stratton’s numbers look the best of all, each of them the best in the table. His .193 BABIP looks much too low at first glance. Not so, as his launch angle and exit velocity indicate he mostly induced weak ground balls. I don’t doubt he’d sustain his low wOBA if he kept that up.
Bridwell’s and Shields’s results are less encouraging. It’s interesting that, despite their higher-than-average exit velocity and launch angle, that both pitchers come out with below-average xWOBA. They both ran low BABIP, so maybe there was luck involved. I’m not sure they’d be able to keep this up with heavy curveball usage.
The pitch evaluation tools I’ve been using don’t look at pitch sequencing, of course. But that’s a big reason not to throw a breaking ball half the time. A curveball that’s successful used sparingly may not be successful when batters see it more often and not right after they’ve seen a fastball or some other setup pitch. While this might not be as important as we once thought, the ability to still mix in other pitches is helpful.
Our candidate pitchers don’t have to stop throwing a fastball, of course. In fact, the sample curveball pitcher threw a fastball of some type 53% of the time on average. 2017 McCullers–who threw his fastball the least among our sample–still did so 40% of the time. Drew Pomeranz, on the other hand, threw 61% fastballs while still maintaining a 37.5% curveball rate.
The curveballer sample averages only a 6% rate of non-curveball offspeed pitches, so perhaps having a changeup or slider isn’t necessary to complement the curve. Each, save Zack Godley and his cutter, throws a four-seamer a substantial percentage of the time. Four of the twelve pitchers threw a sinker more than 10% of the time.
Yet it wasn’t always the case that a curveballer’s fastball was that good, at least as far as pitch values are concerned. In fact, most of their fastballs had negative values. No standout numbers; some pitches were good, some were bad. McCullers, both in 2016 and 2017 had no other pitches with positive value.
Pitch values aren’t too predictive, but maybe it’s not even necessary to have a good fastball to complement the curve. According to Quality of Pitch data, the curveballers had a four-seam fastball QOPA of 4.49 (about average) and two-seam QOPA of 4.89 (above average). But not standout. Perhaps, a curveball pitcherjust needs a decent fastball to go to when not throwing the curveball. The point of having a complementary fastball, really, is all about pitch sequencing, which pitch values and QOPA don’t consider.
Now let’s look at the candidates. We can examine what they throw and could get rid of.
Stratton and Nelson throw sliders a decent amount, but they weren’t particularly good (by pitch values or QOPA). They can safely ditch them. All four threw changeups, and all were bad, save Stratton’s (who threw it about 9% of the time). Pitch values don’t like Stratton’s change, but QOPA thinks it’s above average. QOPA likes its rise, vertical break, and location; all else is below average. Maybe don’t abandon it just yet, but the rest should ditch theirs.
Each pitcher should keep using a fastball to complement his curve. Each pitcher threw a substantial number of four-seamers. Pitch values, though, showed only Stratton and Bridwell’s to be above average. QOPA agrees. In fact, it thinks Bridwell’s four-seamer was great. And his two-seamer. And his cutter.
Last season, he threw his curveball about 20% of the time. So if he increased that, say, fifteen percent, what would he throw less? He can get rid of his changeup–there’s ten points. Other than that, he might want to leave things as they are, though maybe ramp up the two-seamer. If his fastballs are this good, he shouldn’t become a McCullers-level curveball pitcher. But he should still throw it more.
James Shields’ fastball is, well, James Shields’ fastball. His four-seamer, sinker, and cutter are all dismal. But hey, that’s an even better reason to throw the curveball all the time. What’s more, he also throws a slow curve around 5% of the time. It’s a real yacker, rolling in at 69 miles an hour, dropping and moving horizontally a whole seven inches. It’s also quite good. I don’t know how much of that is because he throws it so rarely–catching batters by surprise–but I’m skeptical that a 69 MPH pitch can stay in the yard if batters see it often. He also drops his release point a good bit to throw it, so probably pretty easy to see it coming.
Stratton’s four-seamer was good, but not great, last season. It clocks in at around 92 MPH. He should still throw it a lot, but not 61-percent of the time. If he keeps the changeup, a 50-40-10 fastball-curveball-changeup split could work for him.
Nelson’s sinker might be the best non-curveball here. Pitch values says the 94 MPH pitch was excellent, the best in the league. QOPA similarly thinks it well above average. He should definitely hang onto the pitch, and can ramp up its usage along with the curveball. He can likely phase out his four-seamer.
So far this season, none of our candidate pitchers has meaningfully increased curveball usage. Jimmy Nelson hasn’t pitched, of course. Parker Bridwell has thrown his curve 30% of the time, but only for 1.2 disastrous innings before being promptly sent down. Since then, his time as a Salt Lake Bee has mostly been spent on the DL.
Chris Stratton’s hook just hasn’t been there, and he’s thrown it less. Pitch values doesn’t think it’s been good. QOPA thinks it has improved. He’s getting more movement on the pitch overall. His curve drops an average of two inches more this season. His fastball’s been quite good, so maybe he should try sequencing the two pitches together more often. Stratton himself thinksthat his curveball and four-seamer look similar to batters. It’s hard to say what the problem is, if there’s one at all.
The metrics’ opinion of Shields’s overall performance are mixed, but it appears his curveball has been a bright spot once again, though its average usage is about the same. He threw it a lot in three starts to begin the season, so it’s unclear why he’s used it so sparingly since then. His other pitches have been better, but I hesitate to say they are better. His fWAR is the most it’s been since 2014, though, so maybe he should just keep doing what he’s doing until it doesn’t work anymore.
It’s not really fair to call any of these guys the next Lance McCullers since Lance McCullers might not even be the next Lance McCullers. Even longtime curveballer Drew Pomeranz might have to find a new strategy; his 2018 has been disastrous, and his curve sure hasn’t helped. So maybe we’re asking the wrong question. Pitchers like Patrick Corbin and Masahiro Tanaka have become heavy slider users, so maybe the phenomenon is catching on, just not with curveballs. Only four qualified starters have a greater than 30% curveball usage compared to twelve for sliders. Or maybe the real lesson here is that pitchers shouldn’t be afraid to experiment with pitch mix just because it seems heterodox. If your fastball isn’t your best pitch, maybe you shouldn’t be afraid to throw your best pitch more often.
Stats courtesy of QOP Baseball, Fangraphs, Brooks Baseball, and Baseball Savant. Image by Splauch1 via Wikipedia.