At my old job, my boss occasionally held ideation sessions. He wanted everyone to participate and it wasn’t about finding something actionable, just brainstorming. No suggestion was too ridiculous – sure, it could be picked apart in discussion, but the point was to suggest strange things and see what came of it. Still, I can safely say that none of these non-judgmental ideas sounded to me as crazy as “Let’s intentionally walk this guy in front of Aaron Judge.”

That didn’t stop Stephen Vogt on Tuesday evening. With runners on second and third base and one out in the bottom of the second inning, Vogt did not allow Juan Soto to strike out. He raised four fingers to send Soto first. His reward? A bases-loaded encounter with Judge, the best hitter in baseball. Vogt obviously had a reason for his decision. I did the calculations to see how well this reason matches theory.

In a vacuum, it’s pretty clear why this intentional walk was bad: It loaded the bases with just one out and increased the chance of a big inning, all with the presumptive American League MVP at the plate. But there were two reasons for doing it. First, it took the bat out of Soto’s hands, and Soto is a phenomenal hitter in his own right, especially against right-handers. Second, it opened up the chance for a double play in the bottom of the inning, which would have had a huge positive impact on the Guardians’ chances (they were already down by two). If you squint you can kind of see it; Maybe these two options are the same. It didn’t matter in Game 2 because the Yankees won on the road, but if the Guardians come back and win the series, they’ll be facing New York’s best hitters again on key pitches, which is what Vogt decided Tuesday night to help us guess , what he will do in the future.

I decided to work backwards to figure out what I would have done in this situation. First, I made predictions for Judge and Cade Smith, who was called up from the bullpen for this matchup. I adjusted both forecasts based on regressed turn splits. Smith doesn’t have a ton of major league data to work with, but he’s set on an observed reverse split and I predict he’ll have roughly neutral platoon matchups in the future. The judge is about 5% better against left-handers than right-handers, which is about average for a right-hander.

Using these predictions and a modified Log5 model, I created a grid of modeled results. It looks like this:

From there, I calculated what impact each of these outcomes would have on Cleveland’s probability of winning. For example, a home run would result in the Yankees being 6-0 and the game more or less over (5.4% chance of the Guardians winning). A strikeout would improve things significantly (23.5% win probability). A groundout could either score a run or trigger a double play with about equal probability. This is what it looks like when all the winning probability numbers are filled in:

Overall, I have an 18.1% chance of Cleveland winning the game if Smith faces Judge. This is mainly due to the good chance of eliminating Judge; Most ball-in-play results are pretty bad for the Guardians. In reality, Judge hit a sacrifice throw that was almost exactly the average score – our game odds gave Cleveland a 17.7% chance of winning after the plate came out.

This is of course a difficult situation for the Guardians. But we’re not comparing it to 50% – they were already behind and already had the dangerous part of the Yankees lineup with runners on base. It was already bad. To compare it to the alternative of pitching Soto, I had to do a similar analysis. First, I created a matchup grid for Soto vs. Smith. I chose Smith over a lefty because I can’t imagine Vogt wanting to use a worse backup player in such a big spot; Smith was the only Guardians substitute to warm up, so this seems like a good choice. In any case, here is the outcome probability grid for Soto vs. Smith:

From there I just did some plug and play calculations. For each potential Soto outcome, I adjusted the base/off status and then used my grid of the judge’s potential outcomes from above to further advance the game status. For example, after a Soto strikeout, I replayed the Judge numbers with runners on second and third and two outs. After a Soto double, I ran Judge’s numbers again with a runner on second, one out, and a four-run deficit. I did this for all of Soto’s possible outcomes to find out how likely the Guardians are to win in each case.

Some of these were simple – an unintentional walk is the same as an intentional walk, for example. Some are tricky – a groundout doesn’t always score the runner, so I’m guessing so. Maybe Smith Soto would suggest differently if the bases were open; Maybe he would evaluate Judge differently based on what happened when Soto hit. These are merely generalizations with a lot of room for error. But it’s still worth doing the math, so I did.

Here is the result of all these calculations, the probability of the Guardians winning, taking into account the fact that a Smith/Judge duel will take place after Soto’s attack:

These are pretty intuitive results: if Soto gets hit, the guards are worse off than if they had simply taken him out. When he gets out, they’ll be better off than if they had accompanied him. Thanks, I’ll be here all week. But the really interesting thing is that when you add all of these points together, you get a 19.1% chance of winning the game, which is a full percentage point higher than the predicted probability of winning after an intentional walk.

This feels strange because Soto’s predicted results are strong There. We’re talking a batting line of .267/.425/.428, a WOBA of almost .400. Intuitively, it feels like a smart idea to accompany someone with results good enough to pull off a double play. The problem is that many of Soto’s best results are walks. That’s what’s special about him: Many plate appearances that would normally end in a strikeout or weak contact turn into walks thanks to his elite batting eye. That makes Soto a nightmare in spots where a baserunner is valuable. But if the pitcher’s alternative is to intentionally walk him, then they are U.NIntentional walks simply don’t matter. If you face Soto and accompany him, it’s no problem! This was originally your backup plan.

All in all, the overall increase in Cleveland’s win probability isn’t outrageous. A percentage point in probability of winning isn’t nothing, but it’s hardly a big effect. It’s definitely small enough that other factors might make intentionally walking him a good decision, even if the math doesn’t add up in a vacuum. The Yankees, for example, would almost certainly win anyway. Vogt also had something to offer that I can’t quantify for this exercise: Soto didn’t face Smith, so he’s not building a mental catalog of his pitches. The Guardians plan to use their bullpen a lot, so keeping a little secret there is valuable. With a 2-0 deficit in the series, there isn’t much room for error; Any matchup between Smith and Soto will be meaningful. If you believe Smith will face Soto multiple times with the game on the line, perhaps the added value will make up for what the Guardians lost by forgoing the chance to get Soto out.

As always, when you break something down so precisely, there’s a chance that my math here is wrong. A small mistake could make a big difference given the tiny effects we’re looking for. Additionally, the makeup of both the Yankees and Guardians tends to be risk-taking; The Guardians are underdogs, so I generally like strategies that increase variance for them. But despite this limitation, I am quite surprised by the result. My initial guess was that, thanks to Soto’s penchant for walking even when you don’t want to, throwing to him would be a far better idea than walking with him. But his power and move sharing, as well as the fact that Smith is such an elite reliever that his result against Judge isn’t catastrophic, ensures that the equation works out, or at least that it’s close. So good job, Stephen Vogt. They did something that sounds completely crazy – they intentionally called a walk to literally load the bases for Aaron Judge – and I can’t even say for sure that it was a bad decision.