This Postseason, pitchers have allowed a .311 wOBA and a 3.74 ERA, down from .318 and 4.33 during the regular season. That part’s not terribly surprising. Since the start of the Wild Card era in 1995, the league’s postseason ERA is 3.85, nearly half a run below the regular season ERA of 4.29. The thing that caught my eye was that this year’s .311 wOBA is 21 points lower than its .332 xwOBA. In fact, for as long as we’ve been calculating xwOBA, wOBA has underperformed it in the playoffs:

This year’s gap is the largest, but it’s hardly an outlier. There’s a big gap between ERA and FIP during the postseason. Pitchers have outperformed their FIP 24 times in the last 29 postseasons. Over that period, they’ve run an ERA of 3.85 and a FIP of 4.15. They’re performing better overall, but they’re also outpitching their FIP to the tune of .3 runs per game. I started thinking about the causes that might explain these discrepancies, and I realized that our new postseason leaderboards would allow us to break them down in some cool new ways.

One explanation is that the pitcher pool in the postseason is different (and better). I pulled both regular season and postseason data for every player who pitched in the postseason during the Wild Card era. Next, I prorated their regular season performance to how many innings they pitched during the postseason. In other words, I calculated the stats that we would expect to see if every pitcher performed exactly the same in the playoffs as they had during the regular season that year. That’s what’s on the “Projected” line in the table below:

Since they’re facing better hitters, postseason pitchers saw both their ERA and FIP go up by roughly half a run. Good pitching doesn’t beat good hitting, nor is the inverse true. The two meet somewhere in the middle.

What’s interesting is that the postseason seems to select for pitchers who outperform their FIP. It makes a certain amount of sense. Some are contact managers who make a living beating their FIP, and some just got lucky or had good defense behind them. Either way, they had shinier ERAs than the pitchers who suffered from bad luck or bad defense, so it’s not surprising that they got the ball when it mattered the most. That could likely explain some of the difference that we’re looking for. Still, their already low BABIP dropped further, and the gap between their ERA and FIP widened further.

Let’s start with FIP. Here’s the relevant part of the definition from our glossary, emphasis mine: “FIP uses those statistics and approximates a pitcher’s ERA assuming average outcomes on balls in play.” As we can see in the chart above, outcomes on balls in play are far from average during the postseason. They’re worse. In fact, postseason BABIP has been lower than regular season BABIP in 27 of the last 29 seasons.

Similarly, wOBA is calculated using the actual value of certain results over a long period of time; you multiply a single by 0.888, a double by 1.271, and so on. Lastly, xwOBA is based on the expected wOBA value of balls hit at specific velocities and trajectories. If the game is different in the postseason, those results will be doubly skewed because they don’t reflect the new defensive environment, nor the change in relative value of any particular outcome.

My first thought about what could affect the value of a batted ball was the weather. It gets cold in October and November, which means balls don’t travel as far, which means fewer home runs. I used our splits leaderboard to pull stats in buckets of five degrees — first 40-45, then 45-50, and so on — since the start of the Wild Card era:

Things get weird due to the smaller samples on either side of the graph, but once the temperature drops below 70 degrees, ERA drops below FIP. Using the same buckets of five degrees (but throwing out the lowest and highest buckets, since I worried their small sample sizes affected the data), I ran correlation coefficients between the temperature and four variables:

These are pretty strong correlations. As the temperature rises, batters put the ball in play more, and more of their fly balls go for home runs. I was very surprised to find that BABIP has the strongest correlation to temperature:

When it’s hotter, more balls fall in for hits. In June, the average batted ball event has an xwOBA of .314, and an actual wOBA of .319. However, things pretty much balance out by the time the calendar rolls over to September. In the Wild Card era, pitchers have run a 4.20 regular season ERA off a 4.17 FIP in September and October. Since 2015, batters have run a .316 wOBA off a .314 xwOBA in those months, and their .296 BABIP is exactly the same as the full-season BABIP. While the temperature obviously drops a few more degrees as the game progresses into October and November, the effect of weather is likely smaller than I expected at the outset. It certainly doesn’t explain the huge postseason drop-off:

I’m sure that there have been and will be postseason series where the weather has a big effect. One day, the Twins will make the World Series again, and they will have to play outdoor baseball in Minnesota in November. But there have also been plenty of postseason games in warmer climates or indoor stadiums to balance things out. For now, I’m not all that convinced.

The next variable is that defense is better in the playoffs. We don’t have defensive stats for the postseason, but I pulled our regular season defensive values for every team in the Wild Card era. Teams that made it to the postseason were worth 0.11 runs per game on defense, while teams that didn’t were worth -.07 runs. The average playoff team also ran a slightly lower regular season BABIP, .293 to .295. Not only that, but teams likely play better defense once they reach playoffs. It’s not just because they’re more locked in, more likely to make defensive substitutions in tight games, or because their advanced scouting is more thorough, though I imagine all three of those are true. It’s also because contending teams look to improve at the deadline, sometimes by adding defense. I pulled the defensive value for every team that made the playoffs in the Wild Card era (except 2020), then split it up their seasons at the trade deadline. The difference was small, but it was there. On a per-game basis, the average playoff team was worth 0.121 defensive runs before August, and 0.125 from August onward.

In case you’re curious, the 1999 Mets made the biggest improvement of any playoff team. They were 0.37 runs per game better on defense after the deadline than they were before it. There’s obviously going to be a lot of regression going on with such an extreme outlier, but it no doubt helped that they traded Brian McRae, who in his final season was playing center field despite an alarming -0.26 defensive runs per nine innings in the outfield. The sixth-highest improvement, at 0.22 runs per game, belonged to the 2022 Yankees. The Yanks added Harrison Bader and Andrew Benintendi at the deadline and brought up Oswaldo Cabrera two weeks later. These moves allowed them to shift Aaron Judge to right field and stuff Giancarlo Stanton’s glove down the nearest garbage disposal.

The last variable is the aggressive managing that takes place during the postseason. Managers are funneling more innings to their best pitchers, enabled by the extra rest days in the postseason. That means better relievers, who are less likely to allow inherited runners to score. Second, pitchers get a shorter leash overall. If you’re walking people or getting hit hard, you’re much more likely to get pulled before you can have a true meltdown inning. Since 1995, the average annual strand rate in the regular season is 71.7%. In the postseason the average is 74.8%.

The end result of all these differences? Fewer baserunners, and a smaller percentage of those baserunners coming home to score. Since 2008, 43.6% of regular season PAs have taken place with runners on base. During the playoffs, that number is 41.3%. Having fewer runners on base likely devalues all three components of FIP to some degree. Walks are less likely to push a runner into scoring position. Home runs result in an average of 1.55 runs, down from 1.58 during the regular season. And strikeouts likely lose a tiny portion of their relative value over field outs, since there are fewer runners who could advance on a groundout or by tagging up on a fly ball. Still, the reduced value of balls in play means that the three true outcomes play a larger role in run production:

During the regular season, the league has averaged 1.09 home runs per nine innings. Remember the table at the beginning of this article, where I prorated regular season performance to postseason innings totals? If our postseason pitchers had kept on pitching the way they had in the regular season, they would have allowed 0.89 home runs per nine innings. Instead, they allowed 1.11. Good pitching and good hitting may meet somewhere in the middle in terms of overall value, but that value comes through in different ways during the playoffs. There are fewer base hits, almost exactly the same number of walks, more strikeouts, and slightly more home runs. Ginny Searle wrote about the Guillen Number at Baseball Prospectus yesterday, so I won’t belabor the point, but here’s a quick table that shows the percentage of runs that score on home runs, and more specifically, on solo home runs:

That’s likely why Dan Szymborski found that reliance on the home run during the regular season is at least somewhat predictive of postseason success. It’s a different game during the postseason, but simply by staying the same, home runs loom even larger.

Before we finish, I need to leave you with a mystery. Over the last two postseasons, starting pitcher wOBA isn’t lower than xwOBA. Nearly the entire difference has come from relief pitchers. This postseason, starter wOBA is just two points below its xwOBA, while reliever wOBA is 39 points lower. Relievers have a 2.86 ERA, 1.22 runs below their 4.08 FIP. Starters have a 4.50 ERA, just below their 4.68 FIP. I’m not sure what’s going on here. The differential hasn’t been showing up in the regular season, either in a differential between ERA and FIP, or in a differential between wOBA and xwOBA:

Over the last couple years, reliever ERA and FIP have been in lock step. Why would starters stop outpitching their peripherals in the playoffs, while relievers outpitch them by even more? I don’t have an explanation, aside from the fact that this year we’ve seen a lot of relievers — Aroldis Chapman and Craig Kimbrel come to mind — walking tons of runners, then wiggling out of their own jams. It’s not as if valuing velocity and stuff more highly than command is anything new. We’ve been in the Strike Out the Side Before You Walk in a Run era of pitching for a while now. Whatever is going on this year, the Phillies and Rangers, both of whom own a bullpen ERA that is more than a full run lower than their FIP, better hope that it continues.