Bullpen management in the Wild Card Series

So the Cardinals’ season ends in October just as quickly as it stopped in March...and again in July. Gabe offered his takes on the Wild Card Series against the Padres yesterday.

#20 The offense was not why the Cardinals lost in the first round

Game 3 certainly gave everyone the scapegoat they needed, but the fact is the pitching just did not give the Cardinals much of a chance in this series. They were downright lucky to only give up 4 runs in the two games they did and I suspect would have given up more if a ninth inning happened in Game 3. When pitching is supposed to be the strength and the offense scores 9 runs, and you still lose, the pitching is at fault. Pitching 100 percent failed the Cardinals more than the offense failed, even with Game 3.

I generally agree here. Short of having Back to the Future 2 style time travel ability, any reasonable person would have bet on this Cardinals team to win a game in which they scored nine runs.

It’s safe to assume the Cardinals, pitching staff included, were running on fumes because that’s what you do when you play the most brutal schedule in recent memory. If there were ever a season to give a manager a free pass, this would be the one. But, anecdotally, at least, it felt like the Wild Card Series had its fair share of questionable bullpen decisions. (As an aside, I will maintain waiting until Game 3 to start Jack Flaherty was a bad idea both before the series began and in hindsight.)

I went through the game logs and pulled all of the plate appearances for Padres hitters with a leverage index (LI), a metric defined by FanGraphs below, greater than two.

LI (leverage index): A measure of how important a particular situation is in a baseball game depending on the inning, score, outs, and number of players on base, created by Tom Tango.

Baselines: The average LI is 1 and is considered a neutral situation. 10% of all real game situations have a LI greater than 2, while 60% have a LI less than 1.

This conveniently works out to exactly 20 events over the three games against San Diego. I also included the inning, base-out situation (e.g. _2_, 1 out means the PA began with a runner at second and one gone), pre-play score, result of the play, and win probably added (e.g. a .155 WPA means the Cardinals’ win expectancy for that game increased by 15.5%) in addition to the pitcher and batter names.

By leverage index, the two most important innings (in a vacuum, that is; the winner-take-all Game 3 would have an inflated series LI) were the 8th of Game 1 and the 4th of Game 2. Those two frames account for the six highest leverage plate appearances with a Cardinals pitcher on the mound. North of 3.0 LI, St. Louis had Alex Reyes pitching once, Andrew Miller twice, and Austin Gomber three times.

Let’s start with Game 1. Leading up to the 8th, Mike Shildt’s only noteworthy moves were choosing Génesis Cabrera as the first lefty out of the bullpen and calling on Giovanny Gallegos to face Fernando Tatís Jr. representing the potential go-ahead run in the 6th. Both of these are pretty easy-to-defend choices. The Padres’ 6-9 slots featured two lefties and a switch hitter in Jurickson Profar whose career splits are slightly worse batting from the left side—and we knew this part of the order would get at least one more go later in the game. The LI when Cabrera entered to face Jake Cronenworth at #6 in the order was only 1.65; when Miller came in at the exact same spot in the lineup in the 8th, the LI was 1.70. If we assume Miller was the preferred left-handed reliever, Shildt had ample reason to hold off on using him.

As for the Tatís at-bat with runners aboard at second and third in the 6th (2.58 LI), let’s assume Gallegos was the Cardinals’ second favorite righty in the bullpen behind Reyes. You could maybe make the argument that the only way a higher leverage situation could come about for Reyes is if there were traffic on the base paths and the heart of the Padres’ order was due up in the late innings. As it turns out, of course, this is exactly what happened after Miller hit the leadoff man and yielded a single in the 8th to set up the Reyes vs. Tatís battle with the game on the line.

Now let’s take a look at Game 2. Adam Wainwright was on the ropes in the 4th, allowing four of the first five batters to reach on top of three hits and a walk through three innings. When he loaded the bases with nobody out, the LI swelled to 2.76, and I think there are legitimate questions to be asked about waiting two PAs beyond this point to pull Wainwright. When he was lifted for Gomber, the LI was 3.29 with the bases loaded and one out in a 4-1 game. Cronenworth was set to hit followed by fellow lefty Trent Grisham before the lineup turned over to Tatís and Manny Machado. (The three batter minimum rule complicates this scenario since a double play to end the inning and bypass the minimum is the only way to avoid a lefty-righty matchup.) Cabrera and Miller only threw 11 pitches apiece in Game 1, and unless you like Gomber as your best lefty reliever, I think the data-driven decision would have been to go with one of those two here despite it being so early into the game. For context, the bottom of the ninth in a one-run game begins at around a 3.6 LI. The three Gomber PAs in the 4th were 3.29, 3.56, and 3.79, respectively. But baseball is weird, and this decision worked out in St. Louis’ favor. Their win expectancy climbed from 67.2% to 75.6% after Gomber walked Cronenworth to plate a run before fanning Grisham and Tatís.

Then we have to talk about two very specific ABs that may not have had the highest LI, but certainly felt like the turning point of this series. *Sighs*

The Tatís homer to cut the Padres’ deficit in Game 2 from 6-2 to 6-5 had the largest WPA (.221) of any single play in any of the three games. Not too far behind was the game-tying Machado home run at .191. It was clear Gallegos just didn’t have his best stuff that night, so you could possibly criticize Shildt for leaving him in against Profar even after surrendering back-to-back homers and a double to Tommy Pham, but that’s about it. This was an execution problem more than a decision-making one.

We then get the question of who the 7th inning should have belonged to as Wil Myers stepped in with a 1.51 LI to start the frame. The Cardinals went with Daniel Ponce de Leon, who proceeded to let up a go-ahead homer to Myers and Tatís’ second in as many innings to turn a 6-6 tie into a 9-6 hole. The remaining arms at this point were Ponce de Leon, Reyes, Miller, Tyler Webb, Kodi Whitley, and Johan Oviedo. This probably wasn’t a high enough leverage spot to warrant using Miller or Reyes yet, so I don’t think there’s too much to complain about here besides the unfortunate outcome. There’s also the fact that Myers’ two-run homer off Whitley in the 8th meant Paul Goldschmidt’s homer in the 9th made it an 11-9 game rather than 9-9. I only mention this to further the #NabilCrismattDeservedToBeOnThePlayoffRoster cause.

Finally, we have Game 3. There’s not much to say about this one from a bullpen standpoint thanks to Flaherty’s 6 IP, 1 ER, 8 K, 2 BB line, a line that didn’t look as great as it normally would due to some poor fielding behind Flaherty and zero run support. Perhaps there’s an argument to be made that Miller should have been brought in against the lefties in the 8th (Cronenworth homered off Reyes to extend San Diego’s lead to 4-0), but the non-hindsight version of that claim is somewhat tenuous and ultimately a moot point.

It’s ironic how despite the Cardinals losing this series, the most questionable bullpen decisions were some of the most effective ones. Plain and simple, the Cardinals were overpowered by a ballclub that played better and likely was better. All in all, though, this will be a season I’ll never forget with an outcome (reaching the postseason as the #5 seed) that is very impressive given the unprecedented circumstances.