Thursday afternoon I spent a few minutes talking about pitch-framing with Michael Baumann and Ben Lindbergh. I was on their podcast for a segment to talk about my entry in this year’s Hardball Times Annual, and in the course of the conversation, Ryan Doumit’s name came up. As a big-leaguer, Doumit mostly stayed under the radar, but pitch-framing research exposed his crippling weakness. The numbers made him look bad. Not just bad-bad. Not just run-of-the-mill bad. Extremely bad. Extraordinarily bad. Doumit, as a receiver in 2008, is charged with -63 runs at Baseball Prospectus.
It wasn’t a one-year fluke. For his career, Doumit’s framing was worth almost -200 runs. If you look at his FanGraphs page, you see 8.2 career WAR. Fine role player, average bat. Add in framing, though, and he plummets to a WAR of nearly -12. Doumit goes from being useful to toxic. All because of something we couldn’t even measure a decade ago.
You’d be justified in wondering whether these numbers are accurate. I have trouble believing in them myself. That’s just so, so much value given away. However, allow me to offer this evidence. Doumit caught more than 4,000 innings. Other catchers on his teams caught twice as much. When Doumit was catching, the pitchers allowed 5.34 runs per nine innings. When someone else was catching, the pitchers allowed 4.90 runs per nine innings. That difference, over Doumit’s innings total: 213 runs. Something bad was happening there.
I’ve gone off course. I’m not here to pick on Ryan Doumit. He earned salaries totaling more than $22 million. He did it! But thinking about Doumit made me wonder. How bad could a pitch-framer possibly be? What would be the lower bound? I can’t give you a realistic answer, but I can give you estimates.
You need to imagine someone very bad. What does a good pitch-framer do? A good pitch-framer maximizes strikes, both in the zone and out of it. It follows that a bad pitch-framer minimizes strikes, both in the zone and out of it. Ryan Doumit cost his pitchers strikes. He gave up strikes over the plate, and he failed to steal many strikes off of it. I imagine he had problems with both posture and technique. I don’t know. He’s done playing, so I’ll leave him be.
Here’s what bad framing can look like. It’s not that these guys are necessarily bad framers, but they can have instances of bad framing, just like Joey Votto can take the occasional bad swing.
Imagine a guy who’s constantly giving up could-be strikes. Look, I don’t know why it’s happening. I don’t know what his deal is. I don’t know who he’s mad at. I don’t know why he’s playing so much! In this hypothetical, he’s playing very much. He’s getting a full season’s worth of action. We have a (1) very bad pitch-framer, who (2) catches most of the games. Let’s say he’s catching for the Cubs. It doesn’t actually matter which team he’s assigned to, but the Cubs could stand to be taken down a notch.
Here is the core of our analysis:
That’s just taken from here. This shows you the league-average called-strike rates, for 2016. There’s not much more we need. I’ll be using a constant run value of 0.14 runs per pitch. That’s derived from good research, and even though constants tend to break down at extremes — and even though this is an extreme imagined scenario — I’m just trying to get us into the ballpark. Let’s see where the math takes us.
Another assumption is that our catcher is receiving pitches with an average location distribution. So, the same distribution as shown above. It’s worth noting that not every taken pitch is a borderline pitch. Nearly 40% of pitches wound up in areas with a 0% called-strike rate. The framer’s lack of talent doesn’t matter there. We can begin with the most extreme case imaginable. Our catcher doesn’t receive a single called strike. Not one! Not even on pitches down the middle. Running all that math, and adjusting for playing time, we’d get a framing value of about -400 runs, compared to average. This would be costing a few runs every single game.
A protest: Even this extreme hypothetical is impossible to imagine. Umpires are still going to call strikes on those pitches down the gut. So let’s give our framer full strike credit for pitches thrown to zones with a called-strike rate of 100%. Now the framer’s value shows up at -350 runs.
We can extend that a little, if we want to. Let’s be nice. Why cut it off at the 100% threshold? Let’s set a new threshold at 95%. Those are all still pretty much pitches over the middle. We’ll give our framer credit for an average performance on pitches within those zones. Still no strikes anywhere else. The value sits around -225 runs.
For the final adjustment, let’s move that threshold one more time, to 90%. Our framer gets credit for average performance on all pitches thrown within zones with a called-strike rate of at least 90%. Those are mostly gimme pitches. Once again, no strikes elsewhere. Now our calculated value settles around -200 runs. That’s 200 runs worse than the average receiver, over the equivalent of regular playing time.
That’s like -20 wins. Now, if you wanted to calculate WAR, you’d have to fold in runs for the replacement-level adjustment. And our framer would get credit for a positive positional adjustment, since catchers get a big boost there. Because this guy is a catcher, he’s probably not much of a baserunner. Nothing to be done there. And just as a hitter, Barry Bonds topped out at +117 runs, compared to average, in 2001. If our guy hit like the best version of Barry Bonds, he’d still be by far the worst player around. He’d still be a handful of games worse than replacement, although given the offense, and given that pitch-framing can be subtle, one wonders how much fans would actually hate him. I bet they’d really hate the umpires.
Anyhow, now we have an idea of the lower boundary. It’s pretty far down there. It’s so much worse than even Ryan Doumit. Thank you for joining me for whatever this was.
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.