What We All Saw in the Playoffs
Usually these posts follow some kind of narrative structure. You wouldn’t necessarily say you come to FanGraphs for the storytelling, but any decent article is supposed to tell a story, even if it’s mostly statistical. I’m not going to bother this time. I’m not going to lead in with some manner of gripping anecdote. I just want to show you all a bunch of postseason numbers. I want to show them to you, because some of them are interesting, and postseason numbers aren’t always the easiest things to track down. Certainly not if you want to compare them to the same year’s regular-season numbers. That’s why I’m here today.
I’ve shown some of these plots in the past. What’s different now is that I have a few more plots, and also that the 2018 postseason is officially complete. Below, eight images, and limited commentary. How has playoff baseball compared to regular-season baseball over the years? I’ve gathered a whole host of statistical indicators, mostly with the help of Baseball Reference. Join me, if you will, on a quick analytical journey. We always guess at how the playoffs might play out, or which trends we might observe. We don’t have to guess about what’s already in the books.
I’ve examined the wild-card era, stretching back to 1995. And there’s something important to note, here: The playoffs don’t necessarily feature a representative sample of major-league baseball teams. Playoff teams are all better teams, with better players, and they might have their own characteristics. It would be better to compare all these playoff numbers to weighted regular-season numbers for only the teams that made the playoffs, but I’ve settled here for simplicity. It’s just something to keep in mind as we proceed. Shall we proceed? I don’t know where to begin, so let’s start with the most dramatic-looking plot. For the regular season and for the playoffs, here are the rates of innings thrown by the starting pitchers:
Even if you didn’t know the actual numbers, this is something I imagine you could feel. The regular-season decline is continuing, and this year it was accelerated by the Rays’ adoption of the opener. But the Rays didn’t make the playoffs. Other teams made the playoffs. In the regular season, 60% of all innings were thrown by the starters. In the playoffs, 50% of all innings were thrown by the starters. That is, of course, the low, with the bullpen sharing the load equally. I think this is a trend that can only go so far, but there’s nothing magic about 50%. This could conceivably drop even lower still.
Note that that factors in only each game’s two starters, and it doesn’t account for starting pitchers who throw out of the bullpen, like we saw from Chris Sale, or David Price, or Nathan Eovaldi, or so on. That’s a whole other trend. And a much more difficult one to calculate!
So, related to the above, here’s what’s happened in terms of pitchers used per nine innings:
Historically, in the playoffs, compared to the regular season, we’ve seen an increase here of 10%. This October, we saw an increase of 20%, or to put it another way, we saw the playoff number jump by half of a point. Half of a pitcher! Relative to the 2017 playoffs, that is. Last October, teams used 4.7 pitchers per nine. This past regular season, teams used 4.4 pitchers per nine. This October, teams used 5.2 pitchers per nine. When the starters aren’t working so deep, the bullpen has to pick up the slack. It’s all easily and obviously connected.
Where might you see one consequence of this? Game length:
There’s nothing new about this. The average nine-inning playoff game has lasted half an hour longer than the average nine-inning regular-season game. This year, the gap was 40 minutes, instead of 30. The gap rose by five minutes over last year, even though this time around, playoff offense went down. There are a number of contributing factors here. More pitching changes. Longer commercial breaks. Widespread sign-stealing paranoia! Batters and pitchers simply working slower. Deeper counts. I don’t really like to complain about playoff pace, because the playoffs are important and I’m not as invested as an actual fan of one of the competing teams. But, boy, the playoff pace can really drag, and it’s only part the fault of TV.
What else can we look at? There’s always concern that playoff strikeouts go through the roof. Not so much:
Obviously, the playoff strikeout rate is almost always higher than the regular-season strikeout rate. The pitchers are better, and, while the hitters are also better, the pitchers are more better, relatively speaking. And they throw harder. But at least nothing here is particularly out of control. Historically, the playoff strikeout rate has been higher than the season rate by 12%. This year, the difference was 11%. The playoff strikeout rate didn’t increase at all from 2017, even though the regular-season numbers jumped. That’s promising. Right? I think that’s promising.
Here’s something I like. In some circles, it’s referred to as the Guillen Number. Here are the percentages of runs scored on homers:
Last year was nuts. Last year set a regular-season record for homers, and then in the playoffs, homers accounted for more than half of all run-scoring. This time around, things looked a lot more normal. The regular-season numbers calmed down, and then in the playoffs, the rate settled at 42%. Over this window, the so-called Guillen Number has increased by 9% in the playoffs. For 2018, the increase was just 5%. These playoffs weren’t about dingers; these playoffs were about the Red Sox driving everybody in with two outs for some reason.
I should get to runs. Here are runs:
There was a bit of a slip this October, as the playoffs teams collectively dropped below four runs per nine innings. Historically, in the playoffs, run-scoring has dropped by 11%. This year, in the playoffs, run-scoring dropped by 14%. It follows that, when run-scoring is down, OPS also goes down:
We just saw a collective playoff OPS of .659. For the sake of reference, I don’t know why I’m thinking about Freddy Galvis, but this year Freddy Galvis ran an OPS of .680. Mike Zunino ran an OPS of .669. Historically, in the playoffs, OPS has dropped by 6%. This year, in the playoffs, OPS dropped by 9%. I’ve already mentioned how starters are throwing less and less, meaning relievers are throwing more and more. You can understand why the playoff hitting environment might be increasingly hostile. But for the final image, I want to share a factor that’s easy to overlook. Let’s consider the matter of BABIP:
You might not intuitively think that playoff BABIP would decline. But it declines, and it declines by a lot, even though, again, the playoffs select for better teams, which select for better hitters. Historically, in the playoffs, BABIP has been lower by 6%. This year, in the playoffs, BABIP was lower by 9%. The past two years have been the first years in the wild-card era to have a playoff BABIP under .270. It’s harder to get batted balls to find the ground. It’s harder for a number of reasons.
The pitching is better. The pitchers, individually, mostly throw faster, on account of playoff adrenaline. The defenses are presumably better, and the playoffs afford teams a greater ability to scout their opponents. Temperatures go down! That’s another thing that might kill some fly balls. Playing around with some Statcast data for 2015-2018, the league-average regular-season expected wOBA has been .314. In the playoffs, it’s been .307. But the league-average regular-season actual wOBA has been .317. In the playoffs, it’s been .292. Batters aren’t hitting the ball dramatically worse; they’re just ending up with dramatically worse results. That’s been the pattern, and that’s what happened in 2018.
That’s all I’ve got. Thanks to those of you who stuck around. We’ll do it all again a year from now, I’m sure. This isn’t the last that you’ve seen of this data.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
very cool. i’ve been wondering about this kind of thing for a while… thanks for doing this!
Agreed! Would love to see the next steps, some of which you alluded to (starters used as relievers, impact of the shift, pitches per pitcher per game/series, etc.) and see if there are any other analytic gems!
Everything was more or less expected except the last point about BABIP.
I’m assuming it is mostly better pitchers with better contact management
but I would be interested to see how much of it is due to better defense for the playoff teams.
Just checking: pitchers batting are excluded from these [hitting] numbers right?
Wouldn’t this be almost entirely from shifting?
It’s possible. Do playoff teams shift more?
I don’t have any data to back up my feelings, but it doesn’t seem to me that the playoff teams shift any more than non-playoff teams. I wonder if the weather does have something to do with this, although you have to account for the fact that an enclosed Milwaukee stadium and Houston wouldn’t have any variation due to weather. Even warm weather places like Atlanta and LA are much cooler in mid to late Oct than they are in the summer. Perhaps we need to compare the playoff numbers with regular season numbers in the spring months that have comparable weather, and factor in the enclosed stadiums somehow, to see how much of a factor the weather is. Of course a small sample size of playoff data tends to skew things a bit although the trend seems to be there for a while, making me think the playoff BABIP is no statistical fluke.
Or maybe teams that value pitching and defense are more likely to make the playoffs, which would probably reduce BABIP too.
no wonder I hardly gave a shit about these playoffs
Manfred needs to implement some serious changes to improve the pace of play. I’m a hardcore baseball fanatic, but even I found the pace way too slow in the playoffs. The 20 second pitch limit needs to be enforced. I’d also like to see a maximum of 12 (11?) pitchers on the MLB roster to help reduce pitching changes.
Absolutely agreed on the 20 second limit, it’s in the rules, not having a clock is simply a way of saying that we’re not going to enforce the rules. Similarly for batters stepping out all the time, there’s a limit in the rules, enforce it.
But I don’t like a limit by position, but the fact is, every time they change the rules on roster construction, it’s in a way that makes it EASIER to use more players. 10 Day DL, 10 Days down before you’re eligible to be called back up, and that’s waived if you are replacing someone who’s been DLed (for ten days remember).
So, they don’t carry all that many position players on the bench, because it’s so easy to replace someone if you need to. You only need one long reliever (at most) because if you exhaust him, you send him down and call up another AAAA starter as your long man for the next day, sure you’ve probably only got two of those guys, but remember, if you use them both you can DL the second one and get the first one back any time.
All of these changes increase the number of short relievers carried. Then the playoffs hit, and you never play more than 7 games in 11 days and can completely reset your rosters between series, and you’ve got all these short relievers that throw 100 MPH anyway…
Change the regular season roster construction, make the playoffs 7 games in 7 days (off days between series), even if that means playing a couple of day games. Then you’ll get back to a more traditional pitcher usage, traditional usage didn’t just happen, it was a result of roster constraints forcing people to go long. No one forced them to change the roster rules so every team can have 7+ short relievers on the roster every day.
for regular season- they need to go back to 15 day DL. A lot of the issues we’ve seen the last 2 years are as a direct result of that.
Or 10 day DL for position players and 15 day DL for pitchers.
So for whoever down voted me- you like seeing all the batters pitching? You like teams making up bogus injuries so they can just refill their pitching staffs with relievers, rather than actually manage a 25 man roster? Both of those things are a direct result of the 10 day DL.
I really thought hat the Andrew Miller usage pattern (non-closer multi-inning relief ace) would become more prevalent. But aside from Josh Hader, these playoff teams pretty much just went with more and more relievers for super-short outings. A large part of what is reported i this article.
Plus the fantastic defense of several playoff teams. And, in the WS, the Red Sox OF defense versus the constant pop-ups/fly balls of the Dodgers.
I find it interesting that the usage of RPs in the playoffs has increased dramatically since the Royals’ back to back pennants in ’14 and ’15. I wonder if teams saw their success with Herrera, Davis, and Holland, and shifted towards a more bullpen-reliant model.
it would be interesting to see the numbers with the innings thrown by starters taking out the Brewers…. Just looking- the Brewers had only 33 innings thrown by their “starters” in the playoffs, to 60.2 innings thrown by their relievers. Just thinking they are skewing the number a lot.
so looked really quickly and looks like the number this year for all teams outside of Milwaukee is about 53%. Even including the AL Wild Card game.
And have to remember we had several long playoff games this playoff.
This year we had 18 extra innings to only 8 extra innings last year….. Last year only had 3 extra inning games. This year only 4 but 2 were 13 inning games and then obviously game 3 of the WS with 18 innings…
So I did a bit of looking at the extra innings part. Even keeping Milwaukee in there….. looking at just the 1st 9 innings….. So 18 inning game doesn’t inflate the numbers for the relievers. Looking at if you will the managers intent.
last year starters threw 359 2/3 innings. Relievers just in 1st 9 innings threw 299 1/3 innings. 54.6%
this year starters threw 309 innings. Relievers in 1st 9 innings threw 271 innings. 53.3%
BUT if you take out the Brewers, which is pretty fair as they had 4 of their 10 starts where “starter” only went 2 innings or less- the number this year jumps up to 56.2%.
Not covered in article, but it seems there were more passed balls/wild pitches w/ runners moving up into scoring position or scoring this year during playoffs/World Series. Higher spin rate has downside as control decreases & catchers have difficulty digging balls out of dirt.
What I sadly saw in the playoffs was the Dodgers choking again and becoming baseball’s version of the 1990s Buffalo Bills.
Some ideas for pace of play:
– 20 second pitch clock enforcement (maybe even consider 15 seconds)
– no stepping out of batter’s box unless you get knocked down by an inside pitch
– limit the # of pitching changes per game or per series (add extra changes if it goes to extra innings)
– require pitchers to throw to at least 3 batters or 2 runs scored (exception: injury)
Doesn’t say much here other than what we mostly knew.
I did some weather stuff on the Rockies a few years ago. Interestingly, run scoring at Coors tended to go up over and under a certain temperature at game time, but flatten out between something like 55 and 69 degrees. Never came up with a good idea why though.
Btw part of the more pitchers per nine innings might have to do with the number and length of extra inning games this offseason.