Tony La Russa Is at It Again
I thought this week couldn’t get any better. I got to write about bunts, one of my favorite things to do, and about the Giants picking up tiny edges, another personal favorite. I got to write about Yordan Alvarez and how people underrate him; now I can cross that off my yearly to-do list. But Thursday took the cake. Have you seen this nonsense?
"When was the last time you saw somebody intentionally walked on 1-2?"
"Doesn't happen often."
*Boom* pic.twitter.com/LNR52guBkG
— White Sox Talk (@NBCSWhiteSox) June 9, 2022
I love writing about bad intentional walks. I love writing about bad managerial decisions. But I can’t really wrap my head around this one, hard as I try. Let’s try to do the math, such as it is, while keeping in mind that no amount of math is going to make this make sense.
Let’s start at the top. Trea Turner is an excellent hitter, and Bennett Sousa is a lefty. Turner boasts average platoon splits for his career. Sousa has hardly pitched in the majors, so let’s just consider him an average lefty. With a runner on second and two outs, passing up an excellent righty hitter against your lefty pitcher is standard operating procedure.
Only, the runner wasn’t on second base at the start of the plate appearance. You can see Freddie Freeman standing on second looking confused in the clip above, but he’d advanced there on a wild pitch. In the meantime, Sousa got ahead of Turner 0-2 before bouncing said wild pitch.
Put a runner on second base on an 0-0 count, and an intentional walk to Turner makes good sense. Turner and Max Muncy, the hitter on deck, have roughly equal career lines and projections for the rest of the year in terms of wRC+, but that’s not a great way to look at it. Turner gets to his production via hits, which are much more dangerous than Muncy’s blend of walks and dingers to a team that really wants to avoid allowing another run to score. Muncy’s also been bad this year and may still be hurt, to say nothing of the platoon advantage.
Again, though: there were two strikes. Do you know how hard it is to hit with two strikes? After a 1-2 count, the league as a whole is hitting .167/.220/.258. Hitting is hard! Pitchers are good! Batters strike out 42.5% of the time after a 1-2 count. Trea Turner is hardly immune; he’s a great hitter, and in his career after a 1-2 count, he’s hitting .231/.280/.349. In his career against lefty pitchers – and in a sample so small you should regress it heavily anyway – he’s hitting .258/.310/.390 after a 1-2 count. Take any great hitter, put them down in the count, and they’ll be worse than league average.
Muncy, meanwhile, isn’t exactly helpless against lefties. In his career, he has a 132 wRC+ against lefties and a 125 wRC+ against righties. He bats for a much higher average against lefties, even, which matters when you’re talking about a runner on second. Regress his splits toward the mean, and he’s a hair better against righties than lefties, but more or less a split-less hitter.
That’s the situation as it stood on 1-2: face a diminished Turner or take a fresh crack at Muncy. I fired up my trusty intentional walk decision matrix to crunch the numbers. It works quite simply; you give it the likelihood of each outcome for the hitter against a generic pitcher of the correct handedness, the likelihood of each outcome for the pitcher against a generic hitter of the correct handedness, and league average for that situation. Plug in those numbers and the game situation, and you can find the likelihood of a White Sox win after each possible outcome. Do that for both Turner (assuming no intentional walk) and Muncy, and presto, I can estimate the likelihood of the South Siders winning with and without the walk.
To simplify things even further, I’m removing Sousa from the equation. With so little time in the majors, his projections are probably pretty noisy. Instead, we’ll just be considering Turner and Muncy facing a generic left-hander, which is essentially what the White Sox use Sousa as.
A few more details on the specifics I used: I took Turner’s career production after 1-2, then applied platoon splits to those numbers, rather than use his production against lefties after 1-2, because the sample size is so small in the second case that I’d have to regress it pretty heavily back towards average platoon splits anyway; the end result would be roughly the same in each case. Fortunately, I didn’t have to apply any further adjustments to match his rest-of-season projections; he has a career .362 wOBA, and a projected rest-of-season .362 wOBA. I love when math works out like that.
Here’s the likelihood of those outcomes:
Outcome | Odds |
---|---|
1B | 15.9% |
2B | 3.5% |
3B | 0.8% |
HR | 2.1% |
BB/HBP | 6.7% |
Out | 71.0% |
On a single, I then used the league average rates for scoring from second with two outs, stopping at third, and being thrown out at the plate. I think that’s a fair estimation of Freeman’s baserunning ability; he’s slower than average but a good baserunner, which works out to roughly average in my book. For the record, that’s a 79% chance of scoring, a 4% chance of being thrown out at the plate, and a 17% chance of stopping at third.
Work out the win probability in each of those situations (which I approximated using our WPA Inquirer), and the White Sox stood to win the game 23% of the time if they pitched to Turner. I then did slightly better and adjusted for Muncy hitting next, which moved it down to 22.9%. If you want to see how that looks for each possible outcome, here’s that earlier grid but with win probabilities added:
Outcome | Odds | CWS W% |
---|---|---|
1B | 15.9% | 16.2% |
2B | 3.5% | 14.0% |
3B | 0.8% | 13.8% |
HR | 2.1% | 9.1% |
BB/HBP | 6.7% | 21.9% |
Out | 71.0% | 25.5% |
Next, we need to do the same thing, only for Muncy facing a lefty in a fresh count. As I mentioned above, Muncy has reverse splits in his career, but I’m regressing that data towards the mean (with 1,000 plate appearances of ballast) and calling him a marginally worse hitter against lefties than righties.
Here are the odds that Muncy would achieve each individual outcome, using those splits and his rest-of-season projections. Again, I used league average scoring rates from second and appended win probabilities after each outcome:
Outcome | Odds | CWS W% |
---|---|---|
1B | 9.9% | 15.3% |
2B | 3.5% | 8.3% |
3B | 0.3% | 8.2% |
HR | 5.3% | 5.2% |
BB/HBP | 16.1% | 20.3% |
Out | 65.0% | 25.5% |
Interestingly, if your only goal was to prevent a run from scoring – if, say, it was a tie game in the bottom of the ninth – you could argue that La Russa’s move was correct. Assuming that the Sox escape from the inning at a roughly 65% clip if they walk Muncy, my math says that the Dodgers score fractionally less often in the inning.
If they do score after the intentional walk, though, they’re more likely to do so in chunks. More runners on base, and a more powerful hitter in a better count, is a tough combination. Net of all of that, the math says that the Sox win 21.9% of the time if they intentionally walk Turner to face Muncy, and 22.9% of the time if they don’t.
What happened next? You’ve probably heard – the White Sox had to go and get it out of the ocean:
Intentionally walk the batter to get to Muncy? Bad idea. pic.twitter.com/L2Z7fh2DKm
— SportsNet LA (@SportsNetLA) June 9, 2022
Ex post reasoning is always dangerous, but c’mon. Tony La Russa cited Turner’s batting average with “oh, one, or two strikes” against left-handed pitching in explaining his decision. It’s true! Trea Turner is a good hitter. The problem is, the Dodgers are just good hitters all the way down, and that means you have to capture every edge possible against them. Going from a two-strike count to a zero-strike count is very much not capturing every edge.
Could you construct situations where walking Turner would be the right move? Sure! It would involve making Muncy worse, naturally enough. But you’d have to make him a lot worse. I plugged in Muncy’s stats from this year, when he’s been awful. Coming into Thursday’s game, he was hitting .150/.327/.263. Give him the same marginal platoon splits, and… the White Sox would figure to win the game 22.7% of the time after walking Turner. That’s still worse than they would have done by taking their chances with him, good hitter and all.
You can obviously tweak the numbers enough to make it work. Assume Sousa has enormous platoon splits, and that helps ever so slightly. Assume Turner is better than his projections, and that helps ever so slightly. Assume Muncy is even worse than he’s been this year – which seems like a bad assumption to me, but whatever, we’ve left reasonable assumptions well behind at this point – and that helps a bit. But walking someone with two strikes almost never makes sense. Just bounce three curveballs and hope he swings, at the very least. This is an unconscionable decision.
I’m not entirely sure what to say about La Russa’s defense of it. To quote him, “that wasn’t a tough call.” I agree with La Russa on that, but I disagree on which call he should have made. I don’t know any analyst who would agree with La Russa on this decision. I’m confident that the Chicago front office wouldn’t agree with this decision. For him to think it’s an easy call – the other way – befuddles me.
I’m not really a hot take kind of guy. I don’t think that managerial decisions ever matter that much. Heck, I just ran the math and came up with 1% of a win. That’s tiny. You’d have to make 100 decisions like this to cost yourself a single win in expectation. But that’s just deadweight loss. You could save that 1% of a win simply by not making a bad decision. The players on the field are trying their hardest to win. The manager not helping them out is just about the worst thing he can do.
I don’t expect La Russa to care what I think. He probably doesn’t read FanGraphs. But if he is reading, let me just say this: be curious, but not judgmental. Let the numbers guide you. If the numbers are somehow pointing the other way, I’d be surprised, but happily admit that I’m wrong. But I don’t think La Russa checked the numbers. His answer didn’t suggest that he was thinking the situation through fully. It wasn’t a tough call for him. If you’re going to do something that seems so counter-intuitive, you should think it through. There’s a reason that the last intentional walk on a 1-2 count was in 2014. That one was in extra innings, with a reliever on deck and no hitters available, after the runner on first stole second. That’s an exceptional circumstance, and a good time to consider going against conventional wisdom. This wasn’t, and yet La Russa didn’t even think about it much. It wasn’t close to him. I’m shocked that a major league manager would have done this 15 years ago – let alone today.
Note: there’s been a more recent intentional walk on 1-2 that my query missed. On April 3, 2021, the Rockies walked Corey Seager on a 1-2 count (after the runner on first stole second) to face Chris Taylor. On April 16, the Twins walked Mike Trout after an 0-2 wild pitch, choosing to face Justin Upton. Both Taylor and Upton delivered extra-base hits.
Ben is a writer at FanGraphs. He can be found on Twitter @_Ben_Clemens.
Maybe it’s just me, but it seems like there’s something wrong with the Muncy outcome table – it shows the White Sox having a better chance of winning after a Muncy HR than a Muncy 1B, which can’t be right.
Pasting issue. I typed up the charts from the data manually and flip-flopped HR and 1B probabilities. Fixing it now.
“Assuming that the Sox escape from the inning at a roughly 65% clip if they walk Muncy” (Should Muncy be Turner here?)
Thanks for the write up. Wild it’s only 1%
Nope, that should say Muncy. As in if he walks to load the bases, I made an assumption about how likely the Sox are to escape the jam without allowing a run.
Given the situation was Dodgers up by 2 with a runner on 2nd in the top of the 6th, the 1% difference isn’t as surprising. It it were tied in 9th, that would be a value worth calculating.
“Home runs kill rallies” was a real thing that people actually used to say
A-Rod still does
Ironic.