Let’s Consider a Few More Questionable Intentional Walks by Ben Clemens October 16, 2019 Last week, I wrote about some intentional walk decisions that merited further scrutiny. One of them was pretty bad and one of them was a little bad, but both managers advanced to the NLCS — and now, like clockwork, we have more intentional walk decisions to analyze. Playoff baseball is predictable like that, even as it’s unpredictable in other ways. Let’s take these two chronologically. The Yankees had their backs against the wall, down 2-0 against Gerrit Cole to start the seventh. It got worse quickly, with two of the first three runners reaching base. That gets us to our situation; second and third, one out, and Alex Bregman at the plate facing Adam Ottavino. Aaron Boone put up four fingers; he chose to face Yuli Gurriel rather than Bregman. The surface level math on this one doesn’t look that bad for Boone. The trail runner here simply doesn’t matter that much; if Bregman scores, it’s 5-0, and three innings to score five runs is a tall task against Gerrit Cole and the best relievers the Astros can muster. To approximate the lower run-scoring environment they’re facing, I lowered the run environment in our WPA Inquirer to 4.5. The decision hardly hurt the Yankees; their odds of winning fell from 12.3% before the walk to 12.0% afterwards. When an intentional walk only costs 0.3% of win expectancy when ignoring batters, it’s more often a good one, and a cursory look at Bregman and Gurriel’s wOBA agrees with that. That doesn’t even add in platoon splits; Gurriel has a reverse split in his major league career, and even regressing it back to the mean leaves him only 3% better against lefties than righties. Bregman, meanwhile, clocks in at 6.8% better against lefties than righties post-regression. Put that all together, and Gurriel’s projected wOBA against Britton (accounting for Britton’s own splits) comes out to .314, while Bregman’s is .372. Given that the Yankees gave up so little win expectancy by putting Bregman on, the back-of-the-envelope math makes the decision obvious. The team would need to face a batter who projected to be at least 31 points of wOBA worse than Bregman to make the decision worth it, and Gurriel is 58 points worse. Good intentional walk! If we dig slightly further and look at each player’s projected individual outcomes, it gets even better. The Yankees had 12.4% odds of winning before walking Bregman, and then 12.6% to win when Gurriel stepped to the plate. The walk actually increased the Yankees’ chances of winning. Some of it is Gurriel being worse than Bregman, and some of it is double play equity, which we’re about to dive into after awarding Aaron Boone our FanGraphs Official Good Managing Decision Badge. Ah, double play equity, the last refuge of scoundrels. When Dakota Hudson allowed three of the first four batters to reach in the bottom of the first yesterday, the Cardinals’ season hung in the balance. Man on second, down 2-0, one out; Shildt intentionally walked Howie Kendrick to face Ryan Zimmerman. If that doesn’t get your bad intentional walk senses tingling, I don’t know what will. Putting a runner on base in the first inning? That’s bad! The only reason it isn’t worse is because the Cardinals were already pretty unlikely to win; the generic win expectancy for the Cardinals only decreased from 26.1% to 24.8% as a result of this walk. Again, we can project each player’s wOBA against Dakota Hudson and compare that to the win breakeven. In this case, the Cardinals would need to face a batter with a wOBA 175 points lower than Kendrick to make up for putting him on first base. Even though Kendrick is a better hitter than Zimmerman, he’s not that much better; accounting for platoon splits and for Dakota Hudson, he only projects to be 20 points of wOBA better. Bad intentional walk! Yet again, however, I need to point out that wOBA isn’t designed for this scenario. It considers things in a context-neutral scenario, and this scenario is very much not context-neutral. The Cardinals needed out of the inning, and they needed it badly. If they could escape the inning with no further runs, they’d have 31.1% odds of winning the game. Allow even a single run, and their odds would drop to 23.2%. Now, while a strikeout plus an out is a valid way to get out of the inning, a double play is much cleaner. Intuitively, it all makes a lot of sense; Dakota Hudson gets double plays, Ryan Zimmerman hits into double plays, you can fill in the blanks. But rather than trust my intuition, I did the math, working out the relative chance of each outcome based on each player’s career rates and league averages for righty-righty matchups as I did above. For example, 62% of groundballs Zimmerman hits in double play situations become double plays, and he hits into an above-average amount of grounders. Hudson, of course, induces grounders at a dizzying rate, and the Cardinals defense converts them at a high clip as well. That’s a big tailwind to facing Zimmerman. Put it all together, and the Cardinals were 25.6% to win the game when they faced Kendrick after accounting for the exact batter-pitcher matchup. Rather than face those odds, however, they chose to face Zimmerman instead. After taking everything into consideration, the Cardinals’ chances of winning after walking Kendrick were… 25.6%. Wait, what? This, too, was a good intentional walk? How can that be? Partially, it comes down to the fact that the fourth run simply isn’t very important. Teams down 4-0 after one inning win only 16.8% of the time. Keep it within three, and that only slightly increases; to 23.2%. The difference between 2-0 and 3-0 is 7.9%, the difference between 3-0 and 4-0 is 6.4%, and the difference between 4-0 and 5-0 is 5%. Another way to think about it is assuming the worst case scenario. If Kendrick hit a home run, the Cardinals would have been only 16.8% likely to win the game. If Zimmerman hit a home run after Kendrick walked, the Cardinals would be 10.9% likely to win the game, barely worse. Because the fourth run is low-leverage, in other words, it’s worth gambling on preventing the third run. Even given that run leverage, the groundball tendencies of the players involved are hugely important. If Zimmerman hit grounders and those grounders were converted into double plays at a league average rate, that alone would flip the decision to bad intentional walk territory; his above-average double play rate was worth 0.4% of win probability to the Cards in this situation. If Zimmerman and Hudson both sported average double play rates, that would cost another 0.3% of win probability for the Cardinals, pushing the decision squarely into bad walk territory. Given that the decision relies so heavily on relative groundball rates and the interaction between Hudson and Zimmerman, I wouldn’t disagree with you if you didn’t want to call for the walk. I am, after all, projecting Zimmerman to produce around 60% grounders against Hudson, which is quite a few grounders, though only slightly higher than Hudson’s career 57.4% rate. While my calculations conclude that this too was an acceptable intentional walk, it’s important to point out how thin the margins are. Whether this walk made sense or not comes down to calculations about how likely the Cardinals are to turn a double play, but there are other assumptions baked in. If you think they’re especially proficient at getting the lead runner on grounders, facing Kendrick gets better. If you think that Hudson can dial in his control and never walk Zimmerman, walking Kendrick gets better. Regress Hudson’s groundball rate more than I did, and facing Kendrick takes the lead again. It’s a close decision either way. Grudgingly, I suppose I will also award Mike Shildt the FanGraphs Official Good Managing Decision Badge. His choice looked poor, but there was more to it than meets the eye, and though I’m not sure I would have done it, the numbers backed him up far more than I expected. In an especially cruel bit of irony, Shildt got his grounder — not a double play grounder due to placement, but close enough that you could imagine a different world — and a Kolten Wong error kept them from recording even a single out. Kendrick and Zimmerman both scored. Even the best-thought-out intentional walks lead to disaster a lot of the time. Writer’s note: The original article double-counted Juan Soto’s run (whoops). The above analysis has been updated to reflect win probability changes, though it doesn’t change the conclusion that the walk to Kendrick was too close to call, merely increases the Cardinals’ chances of winning in each situation.