If Aaron Judge were merely large instead of Statue of Liberty–sized, Francisco Lindor might have hit a two-run homer in the sixth inning of a scoreless ALDS Game 3, sending the Indians to a series sweep and a spot in the ALCS.
The Indians-Astros matchup that many anticipated when Cleveland and Houston took 2-0 leads in their respective division series would have pitted one of the best pitching staffs ever against one of the best offenses of all time. Indians pitching delivered the highest-ever FanGraphs wins above replacement, thanks to the lowest-ever park-adjusted fielding independent pitching. If park-adjusted FIP is too abstract for your taste, no matter; the Indians’ staff also had the fifth-highest WAR ever based on little old runs allowed, and the highest since World War II. The Astros’ offense, meanwhile, posted the best wRC+ of any team but the Babe Ruth–and–Lou Gehrig–led Yankees of the late 1920s and early 1930s.
A showdown between an unstoppable lineup and an immovable staff would have been the best possible test of the time-honored baseball belief that “great pitching beats great hitting,” which pundits repeatedly trot out at this time of year. The Astros held up their end in the ALDS, hitting .333/.402/.571 in their four-game triumph over the Red Sox—an impressive (if unsustainable) performance that put them on pace for the best postseason OPS of all time. The Indians didn’t, as the Yankees upset them with three consecutive series-saving wins.
Much to the relief of my editor, though, the subject is still pertinent. The Yankees didn’t have a historic staff, but they did have this season’s second-best pitching WAR, separating themselves from the pack of non-Indians staffs in the second half after adding Sonny Gray, David Robertson, and Tommy Kahnle at the trade deadline. The ALCS will still feature great pitching against great hitting. It will just be the Yankees’ rotation and often unhittable bullpen trying to quiet the Astros’ bats.
Despite the 10-win difference in their full-season totals, the Astros and Yankees—who went 18-26 in largely luck-driven one-run games, compared to the Astros’ 19-13—are almost evenly matched according to underlying indicators of their regular-season strength. Helped by home-field advantage, the Astros win the battle of pre-series probabilities, 58-42 at FanGraphs and 57-43 at FiveThirtyEight. But the old saw about great pitching and great hitting would imply that the Yankees have a hidden edge that those projections might miss.
One might think that the Indians’ ALDS defeat disproves the maxim; Cleveland lost despite its unparalleled staff, even though the Yankees’ offense finished a distant second to the Astros’ in the regular season. But the Indians didn’t lose because the Yankees hit them hard; Yankees hitters produced a collective .201/.289/.356 slash line, which is worse than the full-season line Indians pitchers allowed (.236/.294/.380). From the Indians’ perspective, great pitching did beat great (or at least very good) hitting. The Indians were undone by bad defense and their own offensive struggles—another data point in favor of the “great pitching beats great hitting” assertion, given that the Indians’ lineup was the third-best this year.
Of course, we can’t confirm or refute the saying by citing one series. Nor is the fact that scoring tends to be suppressed in the playoffs proof enough; that could be because the weather tends to be colder, and because the postseason schedule enables teams to ride their best pitchers disproportionately, which they can’t do to the same degree with their best hitters. We need a more representative sample.
To obtain one, I asked Russell Carleton, a writer for Baseball Prospectus and the author of the forthcoming book The Shift, to help me examine what happens when the best pitchers face the best hitters during the regular season. Carleton looked up head-to-head matchups between the top 15 hitters (by OPS) and top 15 pitchers (by FIP) with at least 500 plate appearances or batters faced in a five-season span from 2012-16. Players qualified for the list only in seasons when they placed in the top 15, so Miguel Cabrera, for example, made it in 2012 and 2013 but not in 2016.
The second and third columns in the table below list how the 75 top hitters and 75 top pitchers, respectively, did against all opponents during their seasons in the sample in an array of statistics. The fourth column shows the MLB averages over the same period, which the top players (as expected) easily surpass. The fifth column shows the expected results of the head-to-head matchups between the top hitters and top pitchers, using an odds ratio method similar to the Log5 technique published by Bill James, while the sixth and final column shows the actual results of those 2,957 head-to-head plate appearances. The odds ratio method—which is designed to tell us the likeliest batting average of a matchup between, say, a .280 hitter and a pitcher with a .250 average allowed—works the same way for all players. So if great pitching has an extra advantage against great hitting, the real-life results should favor the pitchers more than the predictions dictate.
Best Hitters vs. Best Pitchers, 2012-16
|Outcome||Top Hitters||Top Pitchers||MLB Average||Faceoff Predictions||Actual Faceoff Results|
|Outcome||Top Hitters||Top Pitchers||MLB Average||Faceoff Predictions||Actual Faceoff Results|
That’s not what we see. Unsurprisingly, top hitters hit worse against top pitchers than they do against all pitchers, and top pitchers pitch worse against top hitters than they do against all hitters. (As the decades-old line usually attributed to Hall of Fame manager Casey Stengel says, “Good pitching will always stop good hitting, and vice versa.”) But the results of the head-to-head matchups between top hitters and top pitchers lean more toward the hitters than expected. The top hitters struck out and hit grounders less often, and homered more often, than the math suggested they would, beating their expected on-base percentage by four points and their expected slugging percentage by 31 points. On the individual level, then, great pitching does not beat great hitting. It’s the other way around. That’s consistent with previous research that has revealed that hitters have more control over the outcome of individual plate appearances than pitchers do. (If the hitter had no impact, the pitcher would have the same strikeout rate regardless of the quality of the opposing hitter, while if the pitcher had no impact, the hitter would have the same strikeout rate regardless of the quality of the opposing pitcher; based on in-game results, the latter lies a little closer to the truth.) We also know that the hitter’s swing has more influence on a batted ball’s exit speed than the incoming speed of the pitch. So when Yankees ace Luis Severino faces Astros MVP contender José Altuve, Altuve has a slight advantage relative to what one would expect based on their full-season stats.
When people say that great pitching beats great hitting, though, they’re often talking about the team level, not the individual level. And although it may seem like what applies on the individual level should also apply on the team level, that’s not necessarily the case.
It’s true that the outcome of each Astros plate appearances against Severino is controlled more by Houston’s hitter than by Severino, but Severino still has some influence: We would expect Altuve to hit worse against Severino than he would against a league-average pitcher. And while the hitter has the bigger impact on any one matchup, Severino’s smaller impact is amplified because it extends to every hitter he faces as the lineup tries to chain together enough hits to push a run across. Those lesser impacts add up.
Even in this year of record home run rates and home run reliance, only 42.3 percent of runs were scored on homers (and only 59 percent of homers were solo shots). That means that most scoring requires a multi-hitter sequence of events, such as non-homer hits, walks, hit by pitches, and “productive outs.” A good pitcher makes each of those “positive” outcomes for the hitter less likely. Against a good pitcher, a hitter’s odds of doing something good decrease, as do the odds of the next hitter—and the one after that, and the one after that—doing something to drive in the batter (or batters) who came before. There’s a cumulative edge there that doesn’t show up in a single plate appearance.
There’s a way to detect that edge, too. Again relying on regular-season data from 2012-16, Carleton calculated the correlations between the runs scored by a team in a given game and two factors: that team’s overall average runs per game, and the opposing team’s overall average runs allowed per game. In general, a team’s run total should be a function of those two factors, although on any particular day, many other elements—luck, injuries, park effects, weather, failing to signal for a replay review, and so on—exert some control.
If great pitching beats great hitting on a team level, then the opposing team’s overall runs allowed—which, granted, is a product of both pitching and defense, as are most things that we typically think of as “pitching stats”—should have a stronger correlation with a team’s runs scored in any given game than that team’s runs scored on a seasonal level.
And in fact, that’s what the stats say. According to Carleton, the correlation between the final score and the opposing team’s overall average runs allowed is .10; the correlation between the final score and the hitters’ overall average runs scored is only .08. Both correlations are weak, because most teams are bunched around the same run-scoring average, and because individual scores vary wildly from day to day (and shutout to blowout). But the correlation on the offensive side falls short of the one on the defensive side. As Carleton writes via email, “Both have an influence (the same way that both pitcher and batter have an influence over a PA) that shakes out statistically. The one for pitchers is just stronger.”
Technically, then, the adage is true: On the game level—or in this case, the best-of-seven series level—great pitching does have a leg (or arm) up on great hitting, all else being equal. But although there is some signal amid all the “great pitching beats great hitting” (and “pitching wins championships”) noise, the dictum’s predictive power lags far behind the assurance with which some state it (as Indians fans found out). For one thing, all else usually isn’t equal; in this case, the Astros are probably better at hitting than the Yankees are at pitching (and, with a lineup of high-contact hitters, may have a small edge of their own against good velocity). For another, in the small-sample theater of a single series, random variation plays a larger role than an incremental advantage that accumulates over the course of a season.
As Carleton says, “This is a case where it's not ‘good pitching always beats good hitting’ as much as ‘good pitching ekes out a small but meaningful victory over good hitting in the aggregate.’” In other words, A.J. Hinch shouldn’t forfeit, and Astros fans shouldn’t despair. Yes, the Yankees’ dealing starters and deep bullpen could counter the Astros’ bats. Great pitching might beat great hitting. But if it doesn’t, don’t be surprised.