What Remains of Clayton Kershaw
For major league pitchers, the end always feels depressingly close at hand. In a game increasingly dominated by power and velocity, losing a tick can be the difference between sneaking a fastball past someone and watching a home run trot. Throw in elbow injuries, blisters, and hitters obsessively watching video looking for any exploitable edge, and it’s a miracle that any pitchers can sustain success.
Clayton Kershaw is no exception to this pattern. He may be the greatest pitcher of the 21st century, but that doesn’t give him special immunity from velocity loss or a license to avoid injury. Gone are the days of Kershaw posting sub-2 ERAs regularly. That’s partially due to the changing offensive environment, to be sure, but it’s also a reflection of the fact that Kershaw is aging. His strikeout rate fell last year to the lowest rate since his rookie season, a pedestrian 23.9%. He made less than 30 starts for a third straight season. His fastball velocity declined 1.5 mph. Age comes for everyone, or so it seems.
When Kershaw returned from a season-opening IL stint, he did so with old-man wiles. His velocity was down another tick, now approaching 90 mph. As Ben Lindbergh ably chronicled, Kershaw’s plan was to stick with what worked in 2018 and survive on sliders and first-pitch strikes. He became perhaps the most extreme pitcher in baseball, piping in first strikes nearly three-quarters of the time and afterward departing the strike zone entirely. He accomplished that trick by throwing his fastball 61% of the time to open at-bats and only 33% of the time after that.
However, even these extreme changes couldn’t hide the fact that Kershaw was slipping. At roughly the midpoint of the season, July 1st, Kershaw was 36th in WAR among pitchers, sandwiched between Kyle Gibson and Sonny Gray. He was marginally better, 24th, by RA9-WAR, but even there, it was weird seeing Kershaw’s name next to Yonny Chirinos and Zach Eflin. Some of this came from a limited workload, but his 3.23 ERA and 3.76 FIP told the same story. It wasn’t just the ball, either: his 77 ERA- and 89 FIP- were his highest since his rookie year, and his 22.7% strikeout rate was down even from last year’s low level.
The magic of the halfway point of the season is that there’s still time to change the narrative. Since the midway point, Kershaw has made six starts. In those outings, he’s been vintage Kershaw — a 1.66 ERA, a 2.43 FIP, and a 32% strikeout rate all look like numbers from Kershaw’s long peak rather than from the last few seasons of being merely above-average. This isn’t a trick of getting a few extra games in pitching-friendly parks in a small sample, either: his 56 FIP- is seventh-best among starters over the interval, and his 39 ERA- is fourth.
Six games is a small sample. Reading too much into a month’s worth of results is a sure way to convince yourself of false narratives. Still, Kershaw’s underlying results back up his improvement — this change looks real, in other words. His strikeout rate didn’t just increase 10% out of nowhere — he started missing bats more, and the strikeouts followed. His slider had generated whiffs on only 28.7% of swings this year as of the beginning of July, in line with last year and far below his dominant peak:
Over his last six starts, it’s back: the slider sports a 40.9% whiff rate. How out of character is that for latter-day Kershaw? In the season-and-a-half between the beginning of 2018 and July 1, 2019, he’d made 40 starts. Out of those 40, only five had slider whiff rates above 40.9%. He’s regained a form, in other words, that he struggled to reach for even a game at a time over the last few years.
Kershaw’s slider isn’t just improving for no reason. He’s throwing the pitch differently, and batters haven’t yet been able to adapt. He’s slowed the pitch down after years of throwing it hard. Out of the 46 starts he’s made since the beginning of 2018, the four with the slowest average slider velocity have occurred since the beginning of July. Kershaw didn’t always throw a hard slider, but he dabbled with it around 2011 and switched to it full-time in 2014, coinciding with his peak. He’d kept that slider velocity even as his fastball lost its zip.
But pitches don’t exist in a vacuum, and the slider’s relationship to the fastball changed markedly. First, the velocity gap between the two pitches crumbled. Take a look at the fastball/slider velocity gap, and you can see a clear pattern. The 2019 data here is again through the beginning of July, excluding his last six starts:
Of course, matched velocity isn’t an insurmountable problem on its own. The bigger problem was that the fastball and slider started behaving more similarly. Kershaw’s fastball has had added glove-side movement over the years, moving closer and closer to the break of his slider. At the same time, its lost velocity gave it more time to fall due to gravity, and Kershaw also got less ride on the ball even after accounting for gravity. In other words, the fastball started moving more like the slider. Take a look at how far apart a fastball and slider thrown at the same starting point end up by year (again, 2019 data is through July 1st):
Wonder why Kershaw’s slider was missing less bats? It simply wasn’t getting the separation, either in velocity or movement, that it needed. Those most recent six starts? They changed the pattern. The velocity gap increased by .9 mph, the movement gap by two full inches. The movement differential isn’t quite back to 2016 levels, but it’s heading that way. Two inches is everything when it comes to missing bats, and giving the slider more time to break and drop was an easy way to get that separation.
Want an evocative demonstration of the value of this added separation? From the beginning of 2018 until July 1, Kershaw got 268 swings at sliders outside of the strike zone. That’s exactly what he wanted — for his career, batters whiff on 63.3% of such pitches. But for 2018 and most of 2019, he simply wasn’t getting those whiffs. Batters came up empty on only 49% of their swings, nearly five standard deviations below his old rate. Pitches that should have ended at-bats resulted in foul balls and gave batters another bite at the apple. For the past six starts, he’s generated whiffs on 69% of the same swings, recapturing his old form.
The extra movement doesn’t only help on out pitches, either. One of Kershaw’s biggest problems early this year was leaving breaking balls over the plate. When batters swung at sliders in the strike zone, it ended badly for Kershaw. He almost never missed bats (15.7% whiff rate), and when batters put the ball in play, they tattooed it to the tune of a .375 wOBA and .410 xwOBA. That wOBA was the highest he’d allowed on zone sliders since 2012.
It’s still a small sample, but batters have been far worse against Kershaw’s slider over the plate over the past six starts. He’s generating misses on 21.6% of swings, solidly better than earlier this year, and the contact hasn’t been as dangerous, either. He’s allowed a .290 wOBA and .332 xwOBA when batters put zone sliders in play. Want further confirmation? Those 29 batted balls have an average exit velocity of 90 mph, down nearly 2 mph from the first half of the year.
Pitching is more complicated than making one single adjustment, of course, and Kershaw has made other changes. He’s leaning even more on the slider when ahead in the count, going to it more than half the time with two strikes. He’s throwing more sliders to lefties and more fastballs to righties, maximizing the platoon tendencies of both pitches. He’s completely abandoned his curveball to lefties until he gets ahead in the count. For the most part though, the new slider is the driver of everything.
Nothing ever lasts. Before long, Kershaw will have lost even more fastball velocity. Time wounds all heels, and no one can outrun it forever. For a month, however, Kershaw has turned back the clock. He’s made a simple adjustment that makes batters’ lives harder, and for now that’s enough. If you didn’t look too closely at the radar gun as Kershaw carved up the Cardinals in his most recent start, you could almost imagine it was 2015 again. Batters flailing over sliders? Curveballs locking knees? A fastball that was miraculously hard to square up? It was all there.
Sometime soon, it won’t all be there. Batters will adjust back. Something else will break, and Kershaw won’t be able to fix it. That hasn’t happened yet though, and dwelling on the future is no fun. Right this instant, Clayton Kershaw is dominating again. Even in a career with such dizzying highs, his 2019 stands out. You can’t cheat time, but you can always try to adapt, and watching Kershaw figure it out on the fly is a delight.
Ben is a writer at FanGraphs. He can be found on Bluesky @benclemens.
Is this headline an Edith Finch reference? If so, bravo
I figured it was a Kazuo Ishiguro reference.
It was. I just started playing Outer Wilds and got reminded of that game, and for whatever reason the title came to mind when thinking about Kershaw.
LOL, I asked myself the same thing. I’m like 75 years old but I’ve played “Edith Finch” and am now doubly careful pushing my grandkids on the swings at the park. Thanks so much for this incredibly informative article. I’ve been watching Kershaw thinking, “why don’t they hit that fastball?” Now I have a better idea.
I *love* the analysis of why Kershaw has pitched so much better lately, but I want to push back against the trope of inevitable time-based decline as … a bit strong?
Kershaw is 31. Half of Greg Maddux’s career fWAR came in his age 31+ seasons; Curt Schilling 33+; Randy Johnson 35+. More recently, Justin Verlander has been averaging over 5 fWAR per season age 33+; Scherzer over 6 fWAR in ages 32+.
Age-based decline is the most likely scenario, sure, no question. For each of the players listed above, there are many more whose performance fell off a proverbial cliff. But the greatest players in all sports are those that are remembered because somehow, improbably, they withstood the tests of time to be great for long stretches. Hitters can’t always adapt; knowing Mariano was going to throw a cutter certainly didn’t seem to help.
There’s definitely some wishing on my part that Kershaw ends up being one of those (just because greatness is fun to watch), but it’s probably not wise to write this off as a fun resurgence that will surely end soon. The not knowing is what makes sports compelling.
Different time, different body.
How much DL time did Maddux rack up by 31?
How much back problem?
How much did he fely on velocity?
Kershack ‘s Sword of Damocles is his back.
Odds are the Dodgers refused to move May at least in part because of Kershaw. Tick-tock…
I mean, no team is gonna trade high-end pitching talent for any reason because you need a lot of it – the Dodgers have two pitching free agents after this year (Ryu and Hill). May is essential regardless of what happens to Clayton Kershaw. May is a FV60 prospect: the last Dodger prospect to have a FV60 on them was Corey Seager (not Buehler, Bellinger or Verdugo had that rating). No team is gonna trade a player like that.
Agreed that aging is significantly about injury, but it’s easy to get caught up in recent history and forget that while injuries make future injury more likely, they don’t guarantee it. Since Kershaw missed a few starts to begin the season in a primarily precautionary way, he has pitched 6 innings in every single start. Meanwhile, it’s Max Scherzer – who looked indestructible – missing significant time with a back injury. No one would have predicted that at the beginning of the year, and many people thought Scherzer would age better than Kershaw going forward despite being 4 years older.
Pitching is awful for the body and injury is a constant risk for even the most durable-appearing pitcher.
I’m not thinking this year so much as past years.
Back issues come and go…until they don’t go.
He’s okay for now. But the odds of him having a career into the late 30’s or even 40’s look lower than for most any pitcher his age, good bad or indifferent. It’s the type of injury history that bodes trouble.
The back injury is not great, sure, but most injuries cause increased risk of re-injury to that area, and the back injury does not involve year+ rehab surgeries like most arm problems; The main point in favor of Kershaw pitching into his mid to late 30s successfully is that he’s never had an arm problem of any significance. That’s the thing most likely to reduce his contribution: extended rehab.
Doc Halladay is a case in point.
Terrific article, thank you.
He is currently on pace for his 11th straight season with a sub-3.00 ERA. That is pretty impressive. I know Rivera had 11 straight sub-3 ERA season, but I wonder who the last starting pitcher to do that was? Maddux and Pedro each just had 7 in a row. Clemens never had more than 3 such seasons in a row.
Probably better to track it in ERA-, because of offensive era, otherwise your list of sub-3s will just come from the dead ball era.
It’s worth considering that Kershaw’s present ERA slots in right between 2011 and 2012 in his career ERA- list (66 this year, 67 in ’12, 62 in ’11). His ERA in 2012 was 2.53; this year it’s 2.77. This year is actually slightly better by ERA- because the league-wide ERA is so much higher than it was back then.
This is a great article, but the whole idea of him being “vintage” Kershaw because his FIP- is 7th in the league and ERA- is 4th is a bit strong. Vintage Kershaw had:
FIP-: 45 (2016, 1st), 53 (2015, 1st), 49 (2014, 1st), 64 (2013, 3rd)
ERA-: 42 (2016,1st), 57 (2015, 3rd), 51 (2014, 1st), 51 (2013, 1st)
You can maybe parse this and say that the overall FIP- and ERA- is similar to what it was at his peak, so therefore he’s like vintage Kershaw, but that’s a misleading statement because it’s a bait and switch on the reference group. Saying someone is pitching like vintage Kershaw sounds really special until you realize that he’s been outperformed by Verlander, Jack Flaherty, and Jacob deGrom during that timespan.
But the article is really good aside from that.
You’re comparing rankings of minus stats from a six start period to a full season. That’s silly – just look at the actual minus stats instead. Ryu’s 16 ERA- is impossible for a full season. To compare to that is ridiculous. That 56 FIP- and and 39 ERA- are vintage Kershaw. If you take a random 6 start period over the course of Kershaw’s peak, he probably doesn’t rank 1st in ERA- or FIP-. He was peak Kershaw because he sustained that performance, while the people ahead of him regressed. We’ll see if he can do it again.
You’ve got the details right, but I think you’ve got the interpretation backwards. This piece is claiming that Kershaw is “vintage Kershaw” because he’s performing at the same level over 6 starts as he did over a full year. That’s an inappropriate comparison because the context is totally different, as evidenced by the “impossible” stats you reference.
I posted this below for someone else, but I would love to break this down some more because I have know a great deal about this particular topic!
Clayton Kershaw’s slider and its performance have been an object of some interest in the Dodgers’ blog community and elsewhere since at least 2017. Kershaw experienced uneven performance when he returned from his first back injury in late 2016 that persisted into the early 2017 season. Ben Lindbergh did some digging into this on Effectively Wild and found some interesting data: Kershaw had actually added slider velocity when he returned from the back injury, following a trend he had been on for several years of his slider continuing to gain velocity. Using baseball savant, folks were able to notice differences in the behavior of the slider betwen when it was thrown around 87 and when it approached 90 (Kershaw had not yet suffered his FB velo decline, but the slider was at times cracking 90). Ben wrote about this on The Ringer after a start in 2017 where Kershaw found his old slider behavior right in the middle of the game and it was completely unhittable thereafter.
In 2018 when his FB velo declined, the slider did not decline at the same rate, causing similar shrinking. Even at 87, the velocity gap was now too small between the pitches. The FB problems became more significant, but Dodger bloggers like Moriyama and Stephens were tracking the slider in game posts regularly for most of 17 and 18. Eno asked Kershaw about his slider at the 17 WS (this is what got the ‘I dont know what analytics site you work for’, answer you might remember).
Kershaw has never not been dominant when he has a significant velo separation between the slider and the fastball. Every time he’s found that separation over the last few seasons, he’s had dominant outings, but he’s gone on to lose it again. He has sometimes been able to dominate without it, but usually not consistently. At times, he has found that separation for portions of a game and immediately dominated them (such an example is the one you’ll find in Lindbergh’s article in 2017 that I mentioned above)
This past month marks the first time in his diminished-FB velo phase that he has started to find that separation again, and consequently the slider swstr has spiked (it also marks the first time I stopped tracking it every game, of course, because that’s how this works). Since Kershaw discovered his slider in a bullpen session in 2009, he has never NOT dominated when it had 4-7 mph of separation from his fastball, regardless of the velocity of that fastball.
Kershaw is an anomaly among left handed pitchers, as you’re probably aware: He doesn’t have a changeup at all, which is quite uncommon for a dominant LHP since they need a strong counter to the right handed batter. Kershaw has done this with his slider – it has identical tunneling to the FB and late break, and the right handed batter swings through it. The slider inside to righties, breaking under their hands or towards their back foot, is the key to Kershaw’s strikeouts against the RHB (They don’t whiff on the curve the same way that lefties do). This has become even more true as his FB velo has declined and the high fastball for the K has become less reliable for him: or more specifically, as that fastball has become more likely to be destroyed if it isn’t missed. The slider is effectively a changeup for him that breaks the other way: it has an identical arm slot, identical tunnelling until the last second, and identical arm speed. It just breaks differently than a normal lefty changeup: It’s role in his arsenal vs RHB is the same.
This last month’s numbers aren’t quite the 2014-2016, peak MVP dominance Kershaw, but they’re very similar to results from most of his prime, and they’re consistent with achieving this specific pitch mix that he’s struggled to maintain since the day he went on the DL with the back injury in 2016. First the slider moved up, and then the FB moved down (for the worst part of his 2018, both of these things were true, and sometimes sliders would be as fast as the fastball. I saw one game where he had sliders -faster- than the FB in 2018).
I think a big part of the reason that this isn’t more public is that Kershaw absolutely doesn’t talk about his performance or planning, and that he’s known to be somewhat of a data/analytics skeptic. However, the statement he made last offseason and in spring about wanting to regain FB velocity was directly tied to this concern, and it’s clearly something both the Dodgers and he have been aware of for some time. Again I will reference Eno asking him about this at WS Media day in 2017, and Kershaw blowing off the question with his ‘I dont know what analytics site you work for..’ line; Kershaw knows this is a thing, and his attempt to regain some FB velocity was related. Since that didn’t work, he’s probably spent the past few months instead looking for a way to achieve this with the slider, instead. Perhaps they have found it, or perhaps this will again be another blip where we see what happens when he finds the separation, but he can’t keep it there: That happened a few times in ’17.
Hopefully this comment will let me post a link, but you can play around with this graph of pitch velocities and see this trend play out. If you set it to only 1 game, you can see that the deepest velo drop in 2018 is from a game that has no pitch velocity data (both graphs drop to 0), and even averaged over 3 it causes a weird valley.
https://www.fangraphs.com/graphs.aspx?playerid=2036&position=P&statArr=279,285&legend=1,2,3&split=base&time=game&ymin=&ymax=&start=2014&end=2019&rtype=mult>1=3&dStatArray=
Unfortunately, FG Charts doesn’t support graphing Swstr vs separate pitches. If you overlay his swstr and set number of games higher (swstr over a 2-4 game sample is a useless graph), you can see the trend lines with the converging velocities.
This feels like one of those analytics communicating with the player situations; if the fan community is this aware that Kershaw is reliant on slider/fastball velocity differential, we know that the Dodgers with one of the most analytically driven front offices in baseball, are aware as well. And yet, he hasn’t been able to maintain that differential over a large sample size the past few years. That leads me to two possibilities, 1) that he’s been aware of it because of front office input, and he’s just unable to manipulate his pitches enough to lower his slider velocity consistently, or 2) the front office/Honeycutt hasn’t come to him because he’s Clayton Kershaw and it’s a “don’t call us, we’ll call you” situation. Or they have and he’s told them, I’m Clayton Kershaw, don’t call me, I’ll call you.
Just a random guess, I wonder if he was frustrated with his results through the first part of the season, realized the velocity wasn’t coming back, and did the “call you” thing with the analytics guys. They tell him, look, this is what’s going on, here are our pitch design ideas on how to fix it. And out of frustration, he listened and has been able to implement it in games. Obviously we’ve seen the results, and I’m sure he has too (outside of the weird 5 walks in the Padres game), so maybe he’ll make it more of a permanent change going forward.
Who knows, maybe the 5 walks were a result of him trying to throw the pitch differently. I’m definitely encouraged though, and including his current run in the overall numbers, his strikeout rate is up 1.5% over last year and his ERA- is down to 66, getting closer to the mid-50’s range he was at in 2015 and 2017. Hopefully it continues.
Any elite pitcher is already aware of the impact of changing speeds. It’s just that, like most aspects of pitching, it’s something far easier said than done.
Please sir, i would like to upvote this comment a second time.
I think you’re misunderstanding my interpretation. I’m saying only that looking at ranking in categories in that small of a sample is silly. His ERA- and FIP- are in line with his peak stats in the 2nd half so far. That’s the point. He is pitching like vintage Kershaw. That doesn’t necessarily mean he will be vintage Kershaw over the full year, but maybe it does mean he’s recaptured some of that magic.
Six starts is not a statistically significant sample. Those starts could mean anything, everything, or nothing. Only time can and will tell, and there just hasn’t been enough of that yet.
“Time wounds all heels, and no one can outrun it forever.”
Brilliant.
Just wait. He’s got, like, six years of being Jamie Moyer that we’re not even approaching yet!
6 of CC and then 6 more of Moyer maybe?
Great article. But one thing that might be getting overlooked is Kershaw’s curious increase in spin rate / Bauer units…even as his fastball velo has declined, he’s added a significant amount of spin:
2015: 94.2mph, 2213rpm
2016: 93.6, 2294
2017: 92.8, 2340
2018: 90.8, 2389
2019: 90.2, 2477
Bauer Units have steadily increased from 23.5 in 2015 to 27.5 this year. Perhaps as a result, he’s managed to keep his Whiff percentage more steady than one might expect for a pitcher who lost 4 ticks off the heater.
Bauer units… My new favorite stat!
To clarify, I’m not suggesting that Kershaw is doing anything untoward to increase his spin rate. In fact, if anything, the gradual increase (as opposed to a one-time, immediate, and sticky (no pun intended) bump) suggests the opposite.
I wondered about this too, and thought about folding it into the article. His slider is also gaining rpm over time. The weird part, though, is that as the spin rate has gone up, he hasn’t gotten more movement, even as the velocity decline has given the ball more time to break. It almost seems like the added spin has all been gyro spin, and my uncertainty about what to make of it led me to leave it out.
I do not have any insider knowledge here, but if we theorize that Kershaw’s aware of the velocity gap issue and tried to address it by regaining velocity in the offseason (unsuccessfully), the other way to regain the gap is to figure out how to throw the slider slower .. but without changing anything about how he throws it, visually, since the pitch is used like a changeup (identical arm speed, arm slot, and tunnel until it breaks) vs right handed batters.
The challenge for Kershaw in that situation is to find a way to throw as hard as he was before and get less velocity out of it.
The only way to do that is to change his grip.
Changing his grip in a way that added inefficient gyro spin (doesn’t impact ball movement, but absorbs kinetic energy from the throw) is actually the only way to do that.
I don’t think Kershaw is so analytically-inclined that he is thinking about adding inefficient spin to absorb kinetic energy, but that’s the only way in practice to throw the ball just as hard as before but get less velocity out of it.
Excellent article.
Kershaw has started 20 games this season, and pitched between 6 and 7 innings in every one of them.
As someone who has watched a lot of his starts, I found that curious as well. There were a couple times when his pitch count was insanely low, but they didn’t run him out for the 8th. It must be a given that barring some sort of no-hitter (even then, it is David Roberts) that he will not go beyond the 7th. If you look at pitch count, lots of 90-95 pitch games.
I think the Dodgers are primarily looking at pitch totals and not innings pitched as a measure of controlling Kershaw’s usage. You now have an injury history, but even before that (16-17), you had the Dodgers talking about the need to control Kershaw’s usage during the regular season because of the amount of postseason innings they want him to absorb, and how having him throw 220+ innings in his prime left him not 100% for the playoffs.
I think it’s a goal of the Dodgers to not really have anyone pitch enough innings to qualify for the ERA title, or just barely, even in the best of circumstances, since they expect their best pitchers to accumulate an additional 20-30 IP of playoff baseball.
In fact if you look today you can see that all 3 Dodger pitchers who would for-sure start games in a playoff series (Ryu, Kershaw and Buehler) sit between 130 and 135 IP so far this year, which is already about 20-25 less than you see from top-line starters who have been starting every 5th day since opening day. For example, Cole, Verlander and Greinke are all in the 150s in IP at this point. The Dodgers will introduce rest for those pitchers throughout the last 2 months as well, if the opportunity isn’t given to them by something like Hyun-jin Ryu waking up one morning with a stiff neck (which is the official reason for his most recent DL stint, which should really be read as “this guy hasn’t thrown more than 130 innings in a season since he came back from his arm problems, so we need to rest him”).
Not disagreeing, but also saying his pitch count is extremely low as well. Lots of times he was taken out of the game while cruising in 6th or 7th with 90-95 pitches.
Ben, I’m curious how you went about writing this article. Did you get the idea from watching Kershaw pitch and noticing differences in his slider? Did you just know that Kershaw was pitching well and dove in from there? The idea/wherewithal to look at differences in movement/velocity separation is fantastic and I’m just very curious how you arrived on the meat of this article. Is this area one of many you typically look at for every pitcher you decide to write about? Great stuff!
The changes in performance and shape of Kershaw’s slider based on velocity have been widely written about for years. I think Ben Lindbergh wrote about it originally; he’s written about Kershaw’s struggles differentiating the slider and the fastball since at least 2017, and Eno Sarris has asked him about it on a number of occaisions (including the 2017 WS). Even before his fastball velocity decline, you can see that the velocity gap between the pitches was shrinking, because Kershaw has consistently thrown his slider -harder- as his career has gone on, and it particularly picked up velocity at the end of 2016 when he returned from his back injury the first time. In early 2017, the slider was sometimes cracking 90, so the difference between it and the fastball was very small even though his FB velo had not yet declined. Ben wrote about it in May of 2017 when Kershaw figured it out in a start; Kershaw’s slider in those days behaved differently when thrown around 87 than it did when it was 89-90 (89-90 lost depth and became more cutter-ish). Tracking the Kershaw slider velo and shape is something that various Dodger bloggers (Like Chad Moriyama and Eric Stephen) have been doing for the past several years.
This is an excellent piece from Ben, but it’s effectively an update to something that’s been watched for quite a while now.
This past month is the first time since his significant FB velocity decline that Kershaw has been able to achieve significant velocity separation between his FB and his slider again. Every time in the past that he’s been able to do that, he’s been utterly dominant.
One of the things about Clayton Kershaw is that his slider has traditionally been a swing and miss pitch in the strike zone. That slider cutting under the hands of the right-handed hitter in the strike zone is a key to his success because Kershaw has no changeup to speak of. In 2017 when he would intermittantly find it for a while, he would look like vintage 14-16 Kershaw for a start or two, and then lose it again as the slider velo crept up.
When Clayton has 4-6 mph separation between his FB and his slider, he has never not dominated. When he has not had that separation, he has occasionally dominated, but usually not.
Well said. I don’t think I’m the first on this beat at all, though I’d only read the Ringer stuff, not the Dodgers bloggers. As to what led me to write this, I watched two recent Kershaw starts (Padres and Cardinals) and noticed the slider was looking feisty again, so I went digging to see if anything was different, and the velo and movement immediately popped.
Yeah, it’s great! This one is near and dear to my heart because I have researched it so much over the years, and I thought providing some context for folks who had never encountered it before would be useful, as it’s never really been in-writing at Fangraphs.
I don’t read the Dodger blogs that closely, but I think mostly they were just tracking slider behavior after Ben’s article.
I like the article, but I’d appreciate it if you set the vertical axis for your charts to start at 0. It makes the visual data presented in the charts misleading, especially the slider whiff%.
Yeah I realized that after I’d made the graphs and probably should have gone back and fixed it. I kind of talked myself into the axes being okay as-is but looking back at it again now I don’t know how I convinced myself of that. Definitely something to improve on for next time.
I’d push back and say setting the vertical axis to 0 adds no value or additional context and would make the chart misleading by undermining its visual impact. You would essentially be shrinking the chart and making the substantial changes in whiff% appear less significant.
The purpose of the chart is to visualize the magnitude of the drop and rebound in Kershaw’s whiff% relative to peak Kershaw, not relative to 0 or even to league average for that matter.
That said, if I would add anything I would potentially add a league average line so we can get an idea of his rates relative to the league/environment. But since the article is focused on Kershaw and only Kershaw that would be irrelevant to the narrative.
Hot take warning….the title of greatest pitcher in the 21st century goes to Justin Verlander, with Kershaw second. For my money (about $1.26), Verlander’s edge in durability both within a game and within the season, coupled with his postseason success, places him ahead of Kershaw’s advantage in rate stats.
Verlander has only pitched about 120 innings more than Kershaw over the period from 2010 to present. While Verlander now has a substantial innings advantage per season, earlier in the decade Kershaw consistently lead the league in IP and IP/GS. Indeed, Verlander has only a 120 IP advantage, but spreads that over an additional 26 starts! (that would be less than 5 IP/start for those 26 starts)
The perception you get based on what things look like now in 2019 doesn’t reflect the balance of the decade; Kershaw has averaged more IP per start than Verlander over that frame, and has pitched on average only about 13 innings a year less than JV.
Verlander has really closed this gap over the last couple years since they’re his best and Kershaw’s worst of the time period, but it would take several more years of this before Verlander actually had a substantial advantage in availability and how deep into games he goes.
lol’in at the suggestion that Verlander’s postseason trackrecord is significantly better than Kershaw’s. If Darvish hadn’t blown game 7 in 2017, you wouldn’t be saying this. Verlander has an extremely mixed postseaon record, just like Kershaw. Verlander however has never won a World Series game, unlike Kersh.
Great write up, but I’m left wanting to know what are the velocity and movement gaps between his Slider and Fastball since July 1?
About 4mph and 13.8 inches of total separation. I thought about throwing in two more graphs with those but the article already felt a little more cluttered than I’d hoped.