Game Plan

Risk and Reward

The Padres followed up a 2-1 series win against the Guardians with a convincing sweep of the Nationals. Through the first six games of the post All-Star break push the Padres are 5-1, and we’re getting a clearer idea of the team’s game plan about how to manage risk tradeoffs down the stretch. We also saw them deviate from the game plan when the moment demanded it.

Game 1

The first game of the series saw Randy Vasquez make his 14th start of the season. Vasquez appeared to have a very clear game plan: throw strikes and change pitches/speeds constantly. Vasquez threw 47 of 68 pitches for strikes, threw six different pitch types five times or more (Cutter, Curveball, 4-Seam Fastball, Sinker, Sweeper, Changeup), and didn’t throw the same pitch twice in a row until the fourth inning when he pitched back to back 4-seamers to Jesse Winker. Vasquez hasn’t had swing and miss stuff all season. He’s averaging a 5.7 strikeouts per nine innings. That’s the second lowest mark in all of baseball among pitchers that have made at least 14 starts (Dakota Hudson at 5.0 K/9 ‘leads’ MLB). It’s hard to know why Vasquez doesn’t get more strikeouts. He averaged 7.9 K/9 in his true rookie season last year with the Yankees, and his career minor league average was almost 10 K/9. And he has good stuff:

Courtesy @PitchingNinja

The lack of swing and miss is a genuine concern, but constantly changing speed and pitch type is probably the best way to limit damage when you know a lot of balls are going to be put into play. It’s very hard to feel confident that this approach is going to lead to sustainable success. But it worked Tuesday: Vasquez shut out the Nationals for six innings.

The Padres would provide run support for Vasquez, including back to back home runs from Luis Arraez and Jurickson Profar:

The most interesting decisions came in the final three innings starting with the bottom of the seventh. With the Padres leading 3-0 Vasquez was allowed to start the seventh and gave up a leadoff double to Juan Yepez. Despite Vasquez having pitched six shutout innings on only 68 pitches, the Padres immediately brought in one of their best bullpen arms in Adrian Morejon who was able to get through the inning stranding the runner and preserving the shutout. The Padres added on in the top of the eighth extending the lead to 4-0. The Padres brought in their second best relief arm Jeremiah Estrada who shut down the Nationals in the bottom of the eighth. With the Padres still leading 4-0 in the bottom of the ninth, Robert Suarez was brought in to a non-save situation and recorded the final three outs to complete the shutout. These decisions are interesting because of the juxtaposition with the bullpen decisions made in game one against the Guardians.

To recall: in game one against the Guardians the Padres trailed 1-0 going into the bottom of the eighth, and despite having their best relievers available to them they opted to bring in Stephen Kolek who gave up four runs and let the game get away. This would seem like a classic moment of Padresing; leaving win probability on the table. Many (us!) were repulsed by this decision initially. But something Shildt said after the game shed light on the risk calculus the Padres are using. Shildt was asked:

“For it to be as tight as it was for those seven innings what led to what happened in the eighth inning?”

Shildt responded:

It’s a tight ballgame. You’re down a run so, you know, we can’t chase it with the guys we bring it home with.

Shildt is suggesting that the team’s approach is to focus on using the best bullpen arms to preserve wins rather than preserving chances to come from behind. That is, he’s implying the team intends to use its best arms to manage downside risk rather than chase upside when there’s an opportunity cost involved. There is a calculus in which this makes sense. When Kolek was brought in the Padres had about a 10% chance of coming from behind to win that game. Pitching Kolek instead of Morejon or Estrada was conceding the 10% win probability to some degree. On the other hand the downside risk was also only 10%, so Kolek’s blowup likely cost the team a tenth of a win at most. In game one against the Nationals the Padres had a 3 run lead when they opted to go to their high leverage arms. The Padres had an 82.2% chance to win. There was a lot more to be lost if the Padres had to go to a less reliable arm, something that’s already burnt them this season. As a hypothetical if they’d had to go to Kolek and he imploded for 4 runs the win probability would have swung from 82.2% Padres to 82.3% Nationals, a swing of more than a game and a half of win probability. Keeping the best arms ready for opportunities to lock down wins and limit downside risk seems to be the gameplan Shildt was alluding to. Padres fans should be prepared for the team to make similar tradeoffs in future bullpen decisions: foregoing chasing upside in order to be prepared to clamp down on downside risk, and doing so aggressively in games they lead that are remotely close. Your mileage may vary on that being the correct risk management strategy. Even if there’s a cogent argument that it’s a sound strategy, you don’t have to like it. It certainly has defeatist aspects.

Game 2

The second game would have been when the Padres fifth starter took the mound. But the Padres do not have a major league quality fifth starter on the roster, and they opted to take advantage of having had an off day earlier in the week to skip the fifth starter slot altogether and have Matt Waldron take the mound after four days of rest. Waldron was touched up in the first giving up back to back singles followed by this Juan Yepez homerun:

That ball was hit well, but not exactly crushed: it was only 97 MPH off the bat with a 26 degree launch angle. But it traveled 390 feet. That’s one of the further distances for balls hit with that EV and LA:

A lot of balls hit with that exit velocity and launch angle become outs. Maybe the reason Yepez’ didn’t was the weather conditions. Washington DC was at 96% humidity on a warm night at game time. Baseballs travel further in humid and warm conditions. If you’ve ever lived in Boston, or New York, or DC you know that the entire summer tends to be hot and humid. It raises the question of whether West Coast teams should assume they’ll get the same long ball production from sluggers imported from teams in the East Coast hot houses…

Regardless, the Nationals were up 3-0 after the first inning, but the Padres had answers. In the top of the second Ha-Seong Kim walked and Jackson Merrill doubled. Luis Arraez would get a two out single to drive in both runners and cut the deficit to 3-2. Jurickson Profar came up next and kept the Nationals in the torture chamber:

An interesting moment occurred in the top of the third with the Padres now leading 5-3 with runners on second and third and one out. Bryce Johnson came to the plate and did this:

It’s 2024 so bunting is anathema. But, there’s a very strong argument that this situation is an exception. First, the idea that bunts are very bad and cost teams runs comes from analysis of the RE24 matrix that shows the expected runs scored for the inning based on various base/out states:

You can analyze the changes in expected runs scored from sacrifice bunting by looking at the expected runs scored in the initial base/out state and the expected runs scored in the resulting base/out state after the bunt. For example, you can see that sacrifice bunting a runner from first base to second actually decreases the expected runs scored for the inning:

You can see that bunting a runner over from first actually costs a team in run expectancy. But look carefully at the situation Bryce Johnson was in: runners on second and third and one out. Then look at the base/out state that a successful sac bunt leads to:

Starting with a base/out state with a run expectancy for the inning of 1.352, and ending with a run already scored and a run expectancy for the rest of the inning of 0.413. In other words a successful sac bunt in this case takes you from a run expectancy for the inning of 1.352 to 1.413. A well executed sacrifice bunt increases the run expectancy for the inning. The key here is the execution risk. If the fielder can make a play to get the lead runner out, the play becomes a disaster for the offense. And that’s where another key piece of context matters. Before bunting became baseball heresy defenses used to regularly counter the threat of a sacrifice bunt by having the corner infielders play near the bag and charge in when a bunt was likely. But bunts have become vanishingly rare. And defenses have reacted accordingly. Watch how the defense reacts to Johnson’s bunt attempt from a different camera angle:

They basically don’t. They weren’t ready for a sacrifice bunt because it’s 2024 and everybody knows you should never bunt. And that was Johnson’s opportunity.

Another execution risk is the hitter not being able to get the bunt down. Here it seems Johnson may actually demonstrate a positive skill at the plate. He seems to be really good at bunting. All of these are from this season in which he’s only had 46 plate appearances:

Courtesy: @PadresDataDaily

There’s another piece of context that matters in the decision to have Johnson hit, and it has to do with the implicit assumptions of the type of analysis that has led to the rule of thumb turned Newtonian law of baseball that bunts are always a bad decision. The RE24 analysis came from identifying the average outcomes of an enormous corpus of data. It was very sound analysis and very insightful. It revolutionized how the game was played. But by virtue of being a model based on average outcomes, it implicitly assumes the quality of the players on the field are average. It assumes the hitter at the plate is average. And the hitter on deck is average. And the pitcher is average. And so on. This is an unavoidable implicit assumption when data is modeled in this way. And the problem is that Bryce Johnson is not an average hitter. He is a -dreadful- hitter. He entered the season with a career OPS hovering around .400 and it has actually gone down this year. That’s kind of an impressive feat actually. The point is we should not assume that a model that demonstrates the average outcomes that average hitters achieve applies to well below average hitters. The most likely outcome of Johnson swinging away is a nonproductive out. By choosing to sacrifice bunt he very likely chose his most efficient path to offense given the situation.

This is probably already 10 times the word count of the next closest discussion of a bunt in 2024 but there’s a last piece that bears analysis. Bryce Johnson’s successful sacrifice bunt gave the Padres a three run lead which upped their win probability from 70.42% to 79.01%. When one of the worst hitters in the league adds almost 9% win probability in a single at bat by bunting that’s cause to revisit our assumptions. Situational decision making has always mattered and always will. There are mounting examples that the 2024 Padres understand this.

Waldron would settle down after the rough first inning to shut out the Nationals for the next five enroute to yet another quality start. The Padres would pull away with a 12-3 lead allowing the low-leverage arms of Yuki Matsui, Logan Gillaspie, and Enyel De Los Santos to close out the win.

Game 3

The blowout in game 2 meant the Padres didn’t have to use any of their high-leverage arms. They kept the powder dry for the final game of series. Dylan Cease took the mound and was absolutely dominant. He shutout the Nationals through seven innings but his pitch count was up to 94. The Padres had been able to scratch out some runs early and led 3-0. This would seem the ideal time to stick to the gameplan and clamp down the risk of blowing a narrow lead by deploying the rested high-leverage arms. But there was another consideration: Cease had not given up a single hit. He had a chance for a no-hitter.

AJ Cassavell reported that Shildt acknowledged his intention was to take Cease out after the seventh inning:

There’s no real calculus for the ineffable value of going for something special. Only that these moments are rare. Whether you think that supersedes a laser focus on winning is a matter of personal judgement. In the end Shildt deviated from the gameplan and let Cease take the mound in the eighth.

Cease worked a clean eighth but was up to 103 pitches to start the ninth. Up first was Ildemaro Vargas who worked a grueling eight pitch at bat fouling off five balls before finally grounding out. Jacob Young hit the first pitch of the next at bat sharply on the ground to Ha-Seong Kim who fielded it and threw to first for the second out. The final batter was CJ Abrams, one of the centerpieces of the original Juan Soto trade. Cease of course had been acquired for a package centered around Drew Thorpe who the Padres got as part of the second Juan Soto Trade. There was a strange symmetry to the at bat. Cease threw back to back sliders and on the second Abrams lined it into right field:

Courtesy: @Padres

No-hitters are often seen as the accomplishment of the pitcher alone, and so allowing Cease to pursue a no-hitter might be construed as elevating the needs of an individual above a focus on winning, something Padres teams have struggled with in the past. But if you watched the game that was never the sense on the field. The essential play to preserve the no-hitter was a metaphor:

Cease’s accomplishment was the team’s accomplishment. There was an interesting coda to Cease’s first bid for a no hitter:

Two years after Luis Arraez had broken up Cease’s no-hitter with two outs in the ninth he delivered him the game ball from only the second ever Padres no-hitter:

The Padres completed the series, and the season, sweep against the Nationals in resounding fashion. The three games showcased clear gameplans and gave us insight into how they will manage risk across the stretch run. It also showed us that there is a willingness to deviate from the game plan in pursuit of something special.

Their next game is against the Baltimore Orioles and they’ve announced Adam Mazur will make the start. When last we wrote we wondered if perhaps the team would make a trade for a major league quality fifth start before Mazur’s next rotation spot came due. Here we are and no trades have been consummated. It’s possible the Padres gameplan is to ride with Mazur. We think it’s more likely this is his last start of the season, with a veteran arm taking his place before Tuesday’s trade deadline. More intrigue to what has already been a stressful, but incredible season.