Pitching the Eighth Is Different Than Pitching the Ninth
I was talking to Nationals reliever Blake Treinen the other day and the topic of command came up. I had to ask what happened earlier this year. He graciously addressed the early-season issues that cost him the closer’s role, and it opened my eyes to how the ninth might be different than the eighth — at least when it comes to typical reliever usage and the way batters act — particularly for a guy like Treinen. Hint: the difference has nothing to do with the pitcher’s demeanor.
First, what Treinen said, in sum:
“When I used to come into situations in my old role, there would be batters on base and the hitters would be aggressive and want to swing. When I came in as a closer, they didn’t swing as much and I didn’t make the adjustment to get strike one in the zone quick enough.”
What’s particular to Treinen here is that he’s a super-sinkerballer with a top-10 ground-ball rate over the last two years. As a reliever, he was brought into situations with men on base late in the game. That’s a recipe for aggressive batters. Take a look at the swing rate with men on base as the game progresses.
| Inning | Swing% |
|---|---|
| 1 | 45.8% |
| 2 | 47.3% |
| 3 | 48.1% |
| 4 | 48.9% |
| 5 | 48.9% |
| 6 | 48.7% |
| 7 | 48.3% |
| 8 | 48.4% |
| 9 | 48.5% |
The average swing rate for the league is 46% across all situations. Every number here except one (the first inning) exceeds that figure, so batters do get more aggressive with men on. They also get more aggressive as the game goes on, especially in extra innings, when everyone seems to want to go home.
| Inning | Swing% |
|---|---|
| 1 | 44.2% |
| 2 | 45.6% |
| 3 | 45.8% |
| 4 | 46.9% |
| 5 | 47.0% |
| 6 | 47.2% |
| 7 | 47.0% |
| 8 | 47.2% |
| 9 | 47.1% |
| 10 | 47.4% |
| 11 | 47.5% |
| 12 | 47.3% |
| 13 | 48.1% |
| 14 | 48.4% |
| 15 | 48.3% |
| 16 | 50.9% |
In the ninth, there’s that little tick downwards in swing rate — as players try to get on base to start a rally, maybe. If you compare the swing rate in the ninth with nobody on (46.7%) to the swing rate in the eighth with runners on (48.4%), you get a general difference in aggressiveness, and that’s meaningful to a pitcher like Treinen.
On the first pitch, the league-wide effect is probably non-existent. Batters swing at 31% of the first pitches they see with runners on in the eighth and 32% in similar situations in the ninth. But for Treinen, he was seeing swings on 36% of his first pitches with runners on last year; early this season, that number dropped to 6% when he came on as a closer with the bases empty. That’s a big difference.
Early this year, he was still using the sinker for strike one against righties, like he did last year. That’s a great tactic for a short at-bat from an aggressive swinger. See how similar his pitch chart was as a closer this year against righties on strike one (right) versus last year in May against righties (left).
Location is similar, pitching mix is similar. Same guy. Different inning, though, and different situations. He mentions making an adjustment to get strike one in the zone. Here’s that adjustment, visually.
It’s most stark when it comes to the pitching-mix change. Look at the addition of all those purple dots. Those are sliders. A first-strike slider is unlikelier than a fastball to generate a swing. If a pitcher throws it in the zone, it can “steal” strike one from a patient batter. Looks like Treinen also is throwing a few more four-seamers in order to ensure the pitch is in the zone.
Early this season, Blake Treinen was pitching like he always has: throwing the sinker on strike one and hoping the aggressive batter at the plate would put the pitch in play in order to try and plate the runs. The problem was that he was now coming into the game as the closer, in the ninth inning. Those batters were less aggressive, and Treinen was rewarded with a 38% first-pitch strike rate.
Since then, the Nationals reliever has realized the difference between coming into a clean-slate situation. He’s changed up his pitching mix and has had a 63% first-pitch strike rate ever since. You know what, he’d probably make a good closer now. And that has nothing to do with his presence on the mound.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
Blake Treinen was once a fairly promising starter as well that would throw way too many pitches early in games, but made the adjustment to the bullpen by keeping the ball low and letting the defense work behind him. It is almost like those starter woes were haunting him, as even when he was successful closing, he was still throwing 20-30 pitches.
It seems to be affecting more than just Treinen and the Nats. There are some crazy bad bullpens out there this season.
Interesting… might help explain Corey Knebel going from 11.0 BB% to 18.6 BB% since becoming closer. Thankfully, his K% also went up.
What is the swing% when the opposing team is down in the 9th by one or two runs (in other words, a save situation). 47.1% doesn’t seem drastically different from 47.2% you see in the 8th, but I would be curious to see if players are even more patient in these situations with no base runners and the tying/winning run is on deck (and similarly more free swinging if their team is up or if one swing can walk the other team off). The conventional wisdom is to be more patient than usual if you don’t represent a tying/winning run in what could be your team’s last few outs.
You mean you’re not impressed by a 1 in a thousand difference in swing rate? Yeah, me neither. The premise of the article seems to make some logical sense, but the numbers simply do not bear it out. Batters will be patient when they know you are afraid to throw strikes and more aggressive when they know you will come after them, so I suspect Treinin’s 9th inning struggles had almost nothing to do with a difference in aggression between the 8th and 9th innings.
There’s a much bigger difference between the eighth with runners on and the ninth with none, unfortunately that’s too hefty for a headline.
That was a 1.7% swing rate difference that only applies with runners on base. So using some round numbers, 4 pitches per plate appearance, makes that one extra swing for every 15 batters with runners on.
Hitters themselves vary from 35-60% swing rates (and maybe more in late-game situations, not sure how to find that), so maybe the matchups worked out in one way for a particular pitcher?
The story is fine and I’m sure Treinen thinks he knows what is happening, but the numbers here just don’t show anything to back up the claims.
[comment deleted because Aweb already said pretty much the same thing below]
This is framed and titled as an 8th inning vs 9th inning comparison, but it’s more a bases empty vs runners-on-base type of thing. None of the evidence here suggests there’s any meaningful difference between the innings themselves.
Also, the reason for the high swing % in the first inning could be because there are high-OBP hitters making up the majority of plate appearances, as opposed to the low-OBP hackers that fill the bottom half of most lineups.
The big jump in the 13th inning could also be due to pitcher aggressiveness more so than hitter aggressiveness, as teams are probably starting to run out of available relievers and want to pitch to contact to save their arms.
Yes, that’s true, but it really is the difference between the ‘closer’ and the rest of the guys — no other inning is ‘owned’, no other pitcher comes on with nobody on as often.
Every pitcher who starts the inning off starts with nobody on. If you are saying that managers prefer mid-inning switches in the 8th inning – isn’t that a managerial issue? I’m not sure why pitchers skill sets should be aligned with excessive managerial brain farts.
The batter swing rate and aggressiveness may make sense as the game goes on – and may call for a different pitcher type – but that is a small difference and may have more to do with batters getting more aggressive the 2nd or 3rd time they face the starter and relievers being more fastball-oriented.
I’m saying that pitchers other than closers traditionally come in at the beginning or middle of innings, since they aren’t married to an inning the same way as a closer. So in that way, yes, the ninth is different, and probably due to our love of the ‘closer’ role. In the end, it’s just a personal story about what was different for Treinen that has a little bit to do with how situations are different from a perspective other than leverage/base-outs state.
man what a great day for fangraphs so far. keep’em coming!
So is there some profile you’re looking for, based on this case study? Someone who can get really good called strikes? Someone who can make them swing anyway?
Clearly, the solution is to have Jeff Mathis on the bench to be the 9th inning catcher.
Excellent article. I had never even thought of this possibility for why not all relievers can transition their success to closing.
Honestly, if a pitcher, at the very least, doesn’t understand the different optimal approaches, depending on the inning, score, outs, and runners, he shouldn’t be pitching in the majors.
This is interesting. But I’m always leery of players’ narratives — some may be true, but since they almost always have one, most have to be just rationalizing randomness.
Treinen’s results this year have been bad whether hitters took the first pitch (.910 OPS in 77 PAs) or swung at it (.979 OPS in 39 PAs). The few that hit the first pitch have hurt him more than ever before (5 for 11, HR, double). Even after an 0-1 count, they’re hitting .306 (15-49), whereas he was always below .185 in that situation.
He knows better than I what he’s seeing from batters, but methinks he’s also just pitching worse than last year.
This is something I’ve been wondering about – how is the batter behaving in terms of the type of contact they’re trying to create. Bring a guy in to a game with runners on base, and you’re likely to see batters focusing on contact with a high % chance of bringing those runners in, if they’re batting in a one run game in the ninth, they at least *seem* more inclined to swing for the fences.
Look at, for example, Tony Watsons HR/FB when he converted to a closer in 2016. It went from single digits for four years running to 14-15% the last two years. Would be interested to see if your average closer saw a similar rise when switching from the 7th/8th innings to the 9th.
Someone like Dusty would surely know this…right?
The more I think about this, the less I buy it as having general impact.
A typical one-inning reliever (like Treinen) throws about 1,000 pitches a year. So even a 1% difference in swing rate between the 8th and 9th innings — which is far more than the overall difference shown above — would mean a difference of 10 swings or takes in a season. And some of those won’t even make a difference. Yes, a handful of changed at-bats CAN impact a reliever’s year, but in the aggregate, the effect would be very small.
The effect could be marginally more for someone like Treinen, who went from often inheriting runners to rarely doing so. But it still shouldn’t matter much.
Lastly, note that Treinen’s overall swing rate this year is significantly higher than his prior 2 years, and his pitches per PA is lower. Now, *maybe* that changed dramatically when he stopped closing, but … I don’t buy it.
Different From*
Viewing this separate from the tiny data results makes it more accessible for me — I’m reading it more as how Treinen internalized and processed a rough patch and how he adapted.
In that context, the actual results and their weight matter less because it was a course that got Treinen to take a step forward.
Yeah I mean for him the effect was monstrous, going from getting a 36%, league average first pitch swing rate to 6% as a closer.
That is a monstrous difference, but there could be a different reason than the slim 1.7% difference in batter aggressiveness – e.g., maybe Treinin got a reputation for being a nibbler in those situations. So the solution may have been correct (be more aggressive throwing strikes both in terms of pitch-mix and location) but due to a different issue. (Another possibility would be if he came in for a lot of 2 and 3 run 9th inning lead situations rather than 1 run leads – the incentives are much greater to try to work a walk in those situations.)
I tend towards Fuego’s explanation here, in part. Perhaps not so much that he got a reputation as a nibbler, but that batters who were already in a spot where they tend to be a little less aggressive see a guy they know may try to get them to chase and those two things came together to snowball on him. Then, in true game theory fashion, he adjusted his strategy to better optimize his results.