Baseball has always been a numbers game, but players and teams are getting smarter than ever — and fans are, too. Throughout The Ringer’s 2017 MLB preview, in a series we’re calling “How to Baseball,” our experts will explore the developments that stand to change the way the game is played and consumed. We’ve never known more, and while knowledge is power, it’s also a wellspring of questions. We hope to answer some of them — and to remind you all to bunt wisely.
Two weeks ago, when Red Sox starter David Price experienced soreness in his elbow that briefly looked like it might lead to Tommy John surgery, I got an email from a podcast listener named Dan. He wanted to know whether the time saved per pitch from the Red Sox potentially replacing Price for a full season would amount to more or less than the time saved from eliminating the nonautomatic intentional walk.
Price, in addition to being a very good pitcher, is also a very slow pitcher. In 2016, he tied Hisashi Iwakuma for the slowest pace among qualified starters, averaging 25.6 seconds between pitches, or 3.6 seconds longer than the average starter. (That was an improvement in his pace from 2015, when, in a season with a slightly quicker average pace, he ranked last at 26.0 seconds.) Even so, some simple arithmetic says that the absence of intentional wide ones last year would have saved more time than the absence of Price. Last year, Price threw 3,595 pitches; replacing him with an average starter would have shaved about 3.6 hours of time between pitches (although the average starter also would have allowed more base runners and runs). If we assume that the signaled intentional walk cuts 37.25 seconds per free pass (as The Wall Street Journal calculated), last year’s IBB savings would have totaled 9.6 hours, roughly three times longer than that of the average Price replacement.
Fortunately for Boston’s hopes of winning the AL East, Price says his elbow feels fine (although he’s still likely to start the season on the DL). Unfortunately for commissioner Rob Manfred’s hopes of quickening baseball’s pace, Price will probably still pitch slowly. Manfred’s pace-of-play regulations, which he instituted in 2015, did trim almost a second between pitches, which helped lop off several minutes from games. But both pace and length increased last season, relative to 2015. And given the inexorable lengthening we’ve witnessed across decades, they’ll likely continue to climb unless further measures (such as the pitch clock that’s already worked well in the minors) are introduced.
It would make Manfred’s task easier if pitchers could be convinced that working quickly were worth it — not just to avoid punishment, but to gain an edge. Earlier this month, FanGraphs writer Travis Sawchik reflected on the career of former starter Mark Buehrle, whom Sawchik labeled the Barry Bonds of pitcher pace:
Perhaps! Buehrle was a quick worker, leading all pitchers in pace in seven of eight PITCHf/x-era seasons (and finishing second in the only other year). He was also a well-above-average starter despite pedestrian (or sub-pedestrian) radar readings. It’s not crazy to suggest a connection between his pace and success; as the theory goes, umpires (whether knowingly or not) reward pitchers who don’t keep them waiting, and fielders focus better behind pitchers who leave less time for minds to wander (and, maybe, more time to get tired from standing around). On the other hand, Price has reached higher highs than Buehrle with a pace at the opposite end of the spectrum.
We’d have a hard time evaluating whether working quickly leads to improved pitch selection and execution; even if we separated the same pitchers into fast and slow outings and found a difference in results, we wouldn’t be able to establish whether “fast” pitchers pitched well because they were working quickly, or worked quickly because they were pitching well. But we can examine whether working quickly leads to better defense and/or extra strikes.
“I think [working quickly] works more for improving defensive support,” says Diamondbacks pitching strategist Dan Haren, whose 18.5-second pace in his last season, 2015, ranked second to Jon Niese among qualified NL starters. “In regard to getting called strikes, there’s a handful of umpires that will punish pitchers for working slowly … but most are pretty fair.” Haren adds that it would be “hard to prove” the defensive relationship but that “position players just appreciate it.”
In 2008, then Hardball Times writer (and current Houston Astros analyst) Mike Fast took a preliminary look at whether working quickly improves pitchers’ defensive support. Fast was working with only two-plus months of complete PITCHf/x data, and he acknowledged that the topic was ripe for revisiting. With pace of play a constant source of discussion and a much larger sample of pitches at our disposal, now seems like the perfect time to examine whether pitchers could benefit from working more like Buehrle and less like Price.
We can start, as Fast did, with a team-level look. The scatter plot below graphs team BABIP allowed against team pitcher pace.
As the lack of a slope suggests, there’s no real relationship here; the correlation between the two stats is only 0.07, where 1 represents a perfect relationship and zero indicates no relationship at all. (The extreme-outlier Cubs, who allowed a .255 BABIP last year, tied for 10th in the NL in pitcher pace.)
Let’s drill down to the individual level. This second scatter plot, which includes all qualified starters from 2008–16 who spent a full season with one team, shows the difference between pitcher BABIP and team BABIP, again plotted against time between pitches. If working more quickly were a dependable way for pitchers to solicit stronger defense than their slower-working rotation mates, we would expect to see some relationship here.
Once again, no slope: This time, the correlation is only 0.04. I’ll spare you the sight of another flat plot, but the correlation for one-team qualified relievers is a similarly insignificant minus-0.05. Fast, who in addition to hits included reaches on error and fielder’s choices on which no outs were recorded, also failed to find evidence of a correlation.
Some pitchers’ paces may fluctuate depending on the day or the inning, which could camouflage a relationship in their overall stats. But with help from FanGraphs writer Jonah Pemstein, we can crank up the magnification one more time and study the subject on the level of the individual pitch. The plot below, based on PITCHf/x data from 2013–16, shows the leaguewide BABIP allowed depending on the time elapsed since the previous pitch.
There’s a slight upward trajectory to the trend line, suggesting a slight tendency for pitches after longer delays to have higher BABIPs allowed. However, the apparent effect is small and seems to be concentrated at the smaller-sample extremes; quick pitches may catch hitters off guard and produce more defensive, less successful swings, but there’s essentially no difference in results throughout the 15–30-second range in which the vast majority of pitches cluster. When we consider that there’s also a separate, if minor, tendency for better hitters to both have higher BABIPs and take more time between pitches — which could slightly skew these results in the direction of a real relationship — any existing leaguewide effect would be weak.
If you’re wondering about Buehrle, the pitcher-pace poster boy, his career BABIP, .292, is exactly the same as his teams’ single-season BABIPs weighted by Buehrle’s innings pitched in each corresponding year. In other words, Buehrle’s BABIP is exactly what we would expect any pitchers’ who played for his teams to be, although there’s no way to know that his BABIP wouldn’t have been even worse had he worked more deliberately.
Buehrle did excel at getting extra called strikes, though: Among 312 pitchers with at least 1,000 innings pitched since 1988, the first season for which Baseball Prospectus’s called strikes above average (CSAA) stat is available, Buehrle ranks 62nd. BP estimates that Buehrle’s ability to get extra strikes — after accounting for the location, the catcher, the umpire, the pitch type, and so on — was worth 51 runs over the course of his career. But no such relationship between pace and extra strikes showed up in the larger population of pitchers. Again, I’ll spare you the flat scatter plot, but among pitchers with at least 100 IP in a season from 2010–16, the correlation between pace and CSAA was an insignificant minus-0.06.
On the per-pitch level, things look a little more interesting. Here’s another graphic via Pemstein, again based on PITCHf/x data from 2013–16:
This graph seems to show that pitches thrown 30 seconds after the previous pitch get about one less expected strike per 100 calls than pitches thrown 10 seconds after the previous pitch. That’s not enormous, but it’s not nothing. Except that here, too, there’s a potential complication: With two strikes, the zone shrinks and pace tends to slow down. Thus it might appear that the zone gets tighter as pace increases, when really the count is responsible. In fact, if we redo the graph for pitches with zero strikes, one strike, and two strikes, the extra-strikes effect shows up only with zero strikes, as the following GIF reveals. It’s hard to come up with a reason why a quicker pace would improve the odds of a strike on zero-strike counts but not on one- or two-strike counts, so this effect too starts to seem insubstantial.
This topic sounds simple — Is it better to be quicker, or isn’t it? — but in light of all the subtle confounding factors involved, it’s tough to tie the research into a tidy package. What we can say is that based on these results, there’s no statistical smoking gun that proves that working quickly leads to demonstrably better defense or friendlier calls. Sorry, Rob Manfred: There are no numbers here to persuade pitchers to speed up.
But if there’s one lesson we’ve learned from previous sabermetric research, it’s that absence of evidence doesn’t always equate to evidence of absence. When experienced players, coaches, and executives have insisted that something is so, they’ve often been right even when statisticians have at first insisted that they’re wrong, only to later recant — about catcher framing; about batter vs. pitcher matchup stats; about the ability to induce weak contact; about spring training stats. Maybe more rigorous adjustments for opponent, park, or pitch type would help detect what pitchers intuit.
Or maybe there’s a way we can square this seeming null result with pitchers’ insistence that there’s something to see. Although there’s no correlation between pitcher pace and overall success (as measured by park-adjusted FIP) among qualified starters from 2008–16, there is a 0.30 correlation between pace and four-seamer speed: The harder a pitcher throws, the slower he tends to work otherwise. That could be because the harder throwers need more time to recover after their high-effort deliveries. But it could also be because the flamethrowers know they can get away without working as fast: If you have Price’s stuff, you can make up for taking your time between pitches, but if you throw as softly as Buehrle, you have to hurry, pressing the attack whenever the ball isn’t in (relatively slow) motion. Perhaps — there’s that word again — pace is an equalizer, a slight edge that the soft tossers are already exploiting even though it’s disguised in the leaguewide stats. In that case, Buehrle’s quick work may have helped him last as long as he did.
Thanks to Jonah Pemstein of FanGraphs and Rob McQuown of Baseball Prospectus for research assistance.
