At the moment, the really unfortunate thing about the Craig Kimbrel trade is that it was still the right move.
Nick Madrigal and Codi Heuer will produce plenty for the Cubs, but it seems unlikely that either winds up as the crucial, missing piece to a championship squad. Some regression from Kimbrel’s ridiculous 0.49 ERA with the Cubs was inevitable, but there wasn’t much indicating that he might hit the ground this hard.
There’s little doubt that Kimbrel’s success earlier this season was legit. But it’s not just his results that have changed. Some of it has to do with luck, but over the past month, his velocity has plummeted. Sitting in the 500 level this past Saturday, I couldn’t help but notice the radar gun as Kimbrel struggled to put away Red Sox hitters and ultimately allowed the tying run of the eventual 9-8 loss. 95. 95. 95. 95. He might have popped 96 a couple of times, at most. That is not the Craig Kimbrel fans and the team thought they were acquiring. Yet instead of the force of nature we saw up North this spring, he’s turned back into the pumpkin he looked like in the 2019-20 seasons.
The intricacies of Kimbrel’s struggles seem to be manyfold. James Fegan’s feature on Kimbrel in The Athletic last week touches on his struggles in a few different places. Kimbrel and the team attributed much of them to wayward movement on his fastball and other small things that need tweaking. The ever-nebulous question of whether closers pitch differently in non-save situations was also cited.
There’s no reason to disbelieve them. The margins are really small in high-leverage relief pitching. The effects of that are already showing up in both Kimbrel’s numbers and his game-by-game performance. A look at Statcast data tells us something pretty intuitive: Kimbrel caught a decent amount of luck in the first half, even though he was still really good. Since the trade, the pendulum has swung nastily back in the other direction.
Even by underlying numbers he hasn’t been quite as good with the Sox, but he hasn’t been this bad, either. Those little breaks that are the difference between actual and expected stats manifest in ways that stick in our memories and influence our judgments. In retrospect, doesn’t Andrew Romine’s three-run homer to tie the game in Kimbrel’s first post-trade Wrigley Field appearance feel ominous?
The kicker? The exit velocity on that ball didn’t even break 95 mph. It had a .170 expected batting average. But it happened to be in one of the 10 parks around the league in which such a batted ball would’ve been a home run — and that doesn’t even account for the wind blowing out that day.
Andrew Romine vs Craig Kimbrel#CubTogether— Would it dong? (@would_it_dong) August 6, 2021
Exit velo: 94.1 mph
Launch angle: 30 deg
Proj. distance: 365 ft
This would have been a home run in 10/30 MLB ballparks
CWS (4) @ CHC (4)
Remember a few weeks later in Toronto, when Lance Lynn pitched his guts out only for the offense to flop and Kimbrel to wild-pitch himself into a 2-1 loss? That winning run only got on base because Reese McGuire hit one too slowly into the shift and beat out the throw.
Hit probability? 23%. So in a way, luck really does explain some of it. But it doesn’t change the fact that Kimbrel’s velocity has fallen off a cliff (relatively speaking) since arriving on the South Side:
After averaging out faster than 97 mph with the Cubs — and routinely breaking 98 mph in individual games — Kimbrel’s fastball velocity dropped to 96.1 mph in August and has receded even farther to 95.8 mph in September. That would be his lowest monthly average since 2011. This raises a lot of red flags when we see how well Kimbrel’s run prevention has correlated with his fastball velocity in recent years:
As Fegan’s article discusses, Kimbrel’s fastball and unorthodox delivery and release point both have outlier traits that make his fastball almost unhittable when it’s buzzing in the 97-99 mph range. But once again, the margins are minuscule. Visualizing CSW (Called Strike and Whiff) and Swinging Strike rates on all the fastballs Kimbrel has thrown in the Statcast era, we can see how the fastball’s effectiveness has a small-ish drop in effectiveness once it dips below 98 mph, but a much steeper drop once he starts leaking down into the 95-96 mph range.
Credit goes to Esteban Rivera of Pinstripe Alley for the data crunching. Even though he’s never going to reach the velocity peaks of his late-20s again, as long as Kimbrel can keep his fastball in the vicinity of 98 mph, he’s golden. Father Time comes for us all and takes fastball velocity (and effectiveness) with it, yet even when he’s living around 97 mph, his swinging strike rate is still high above league average. After that, things start looking sketchy. Once we get down to 96 mph, the sample sizes get small and noisy, but the trends are not good, as its once-dominant swinging strike rate drops to barely above league average even before we get to the point where he was at this past weekend.
Craig Kimbrel with an average-ish fastball is not something we need to see in October.
I could be wrong, but it doesn’t seem as if there’s been an obvious change to Kimbrel’s mechanics. There are undoubtedly the small tweaks that he discussed in The Athletic, but those aren’t significant enough to explain a drop of nearly two miles per hour in the span of a few weeks. His release point hasn’t changed. What else is there?
There is the bugaboo that’s already bitten into a healthy chunk of Sox pitching since the All-Star break: fatigue. Kimbrel didn’t sign with the Cubs until late in Spring 2019 and didn’t make his first appearance until June 27. After that, 2020 was 2020. He’s already surpassed his appearance and innings numbers from those seasons combined by quite a bit. Kimbrel is 33 years old and has 640 outings worth of mileage on his arm. It’s pretty reasonable to think that the tank might be running low by the sixth month of his first full season in three years.
So, given all of these circumstances, my consternation last Saturday night turned to total bewilderment on Sunday when Kimbrel was brought on to pitch for the third consecutive day, despite registering his lowest average velocity in years the day prior. Because tired arms generally start throwing harder, right? And as we know, for the second consecutive day, it didn’t really work out. The question is begged: Is Kimbrel being overused?
It’s hard to definitively say yes, but it’s probably fair to conclude that Tony La Russa needs to take his foot off the gas at least a little bit. Kimbrel had appeared in 39 of 101 Cubs games leading up to the final outing before being traded, a rate under 40%. Between the trade and his most recent appearance, Kimbrel appeared in 18 of 37 White Sox games, nearly 50%. 11 of those 39 Cubs appearances came on zero days’ rest (28%), while six of 18 with the Sox have come on the second half of back-to-back days (33%).
That’s not a massive spike in usage, but with all the aforementioned about fatigue, the simple fact that there has been an increase in usage is a bit baffling. Put it this way: If La Russa isn’t afraid to hand the ball to Mike Wright Jr. in the 10th inning of a September game against a playoff team, then he shouldn’t be afraid to give Kimbrel a little more rest and let someone else eat those innings.
That’s especially true considering the results of those games he pitches with no rest. This season, Kimbrel has allowed runs in just two out of the 19 appearances he’s made with three days’ rest or more. It gets a little bit worse on one or two days’ rest, as he’s allowed runs in four of 21 games falling in that category. And as of last Sunday’s outing, Kimbrel has allowed runs in six of the 17 appearances he’s made with no rest at all.
Any individual chart or statistic can lie to you shamelessly. Sometimes, groups of charts and statistics can do the same. But when all of the above comes together, it’s hard to say that the numbers are lying. This season, there’s been a more or less linear correlation between how rested Kimbrel is and how frequently he allows runs. Over the past several seasons, there’s also been a more or less linear correlation between his average fastball velocity and his run prevention. And using a half-decade’s worth of information, we can find the precise points where his fastball stops being effective — and observe how he’s falling below that threshold more and more frequently these days.
Put all of that together, and you should be pretty unsurprised at what we saw this past weekend, which seems to have put many people over the edge into the “is his option even getting picked up?” school of thought. In his last three appearances without any rest, Kimbrel’s average fastball velocity has been 95.3 mph, 95.1 mph, and 95.2 mph. Returning to the graphs from earlier, those numbers are very literally off the chart, and not in a way that suggests anything positive.
I’m not sure what the solution to these problems is, but I’m more than sure that it’s not to keep throwing him into games haphazardly. Again, if the plan is to leave an extra-inning game in the hands of Wright, as it seemed to be for both games against Boston, then why drain Kimbrel’s already-limited energy?
These are questions that La Russa and his staff will have to figure out before Houston (presumably) comes to town in the first weekend of October. Kimbrel and Sox coaches seem to have confidence that he can get back to where he needs to be in terms of his movement and location. But there’s evidence that it’s a moot point unless he gets at least some of that fastball velocity back.
Whether La Russa will give Kimbrel space to do whatever’s necessary to get it done remains to be seen.