The New Postseason Plan: Defense Early, Offense Late

In general, when Major League teams have to choose how to deploy one-dimensional players, they go offense first, then defense later. Bat-only players are usually starters, they get their three at-bats, and then they are lifted for defensive replacements late in the game if there’s a lead to protect. This usage generally minimizes the number of at-bats you have to give to the weak hitting defensive specialist, and putting your best defensive unit on the field when you have a lead to protect seems to make sense, since you don’t need to score any more runs at that point, so long as you don’t let the other team score.

But baseball has changed, and postseason baseball has changed even more dramatically, so for the Indians and Cubs, I’d suggest that the best way to utilize their specialists is to start the defenders and sub in the offensive upgrades in the middle innings.

For the Cubs, this manifests itself mostly in the Jason Heyward decision. In writing about Kyle Schwarber yesterday, I noted that Schwarber would be taking DH at-bats that could go to Willson Contreras, and so the only way to get Contreras in the line-up in the games in Cleveland would be to bench Heyward. A lot of people are on board with that idea, given how poorly Heyward hit this year, and how awful he’s looked in the postseason. Right now, Heyward looks like a classic defensive replacement guy, a good glove who you don’t trust to do anything at the plate.

But if you want to maximize Heyward’s value in a part-time role, you want him playing in the first few innings of the game, not the last few. Here’s why.

Cubs SP/RP Breakdown
Cubs BB% K% HR% Ball in Play %
Starters 7% 24% 3% 65%
Relievers 8% 29% 2% 60%

This isn’t any kind of surprise, but Cubs’ relievers in the regular season struck out batters at a much higher rate than Cubs’ starting pitchers. In the postseason, that gap only grows, as the high-K relievers like Aroldis Chapman get expanded usage, and the team is more aggressive playing the match-ups to avoid balls in play.

Simply put, defense matters less when relievers are on the mound, because relievers are more likely to just get the job done themselves. When you use a defensive substitution to bolster the run prevention behind a reliever, you’re intentionally putting a defender into the time of game his skills are less likely to be utilized. If you start Contreras in the outfield, then sub in Heyward in the seventh or eighth inning as a defensive upgrade, there’s a pretty good chance he’s just going to stand there and watch Chapman, Hector Rondon, or Pedro Strop strike everybody out.

And while it sounds reasonable, the idea that defense becomes of greater importance with a lead isn’t really true. Additional runs scored don’t stop being important once you’ve taken a lead; the goal from that point is to simply not let your opponent outscore you, which can be done just as effectively by scoring as many runs as they score as it can be by both teams failing to score the rest of the way. Swapping out offense for defense with a late lead changes the way you’re likely to win, but doesn’t really change your likelihood of winning, especially if the defensive substitute has to hit in a critical situation.

And that’s the other consideration here for the Cubs. If you start Heyward, you basically guarantee he’s going to get his at-bats against right-handed pitching, since Cleveland doesn’t have any left-handed starting pitchers besides Ryan Merritt, who I can’t imagine they’ll trust with a World Series start. By starting Heyward and hitting him towards the bottom of the order, he’ll get essentially one guaranteed at-bat against an RHP, and then you can decide whether you want him to hit a second time based on the circumstance and who is pitching at the time.

If Terry Francona has gone to Andrew Miller in the 4th or 5th inning again, you can send Contreras or Jorge Soler up to the plate to bat with the platoon advantage, knowing there’s zero chance that Francona will play the match-ups there. So not only do you get the right-handed bat against the left-handed pitcher, but you also might have gotten four innings or even five innings of defensive work from Heyward while only giving him one at-bat. In a low-scoring environment where each game could turn on one defensive misplay, the Cubs are probably better off maximizing Heyward’s time in the field, even if he has to hit once against an RHP, then they are maximizing the number of times Contreras bats, especially given Cleveland’s RHP-heavy rotation.

For Terry Francona, this is less of an issue in the first two games of the series, but this will come up as a decision once the games move to Wrigley Field, and the team has to decide what to do with Mike Napoli. In an offense-first world, the team could theoretically move Carlos Santana to third base to open up first base for Napoli, shifting Jose Ramirez to left field, and starting Coco Crisp on the bench. This would get the team’s best hitters in the line-up even without the DH, but sticking Santana at third for even a few innings with pitch-to-contact starters like Josh Tomlin probably isn’t worth the cost.

I’d imagine Francona won’t seriously consider punting defense given how well their team is playing with Crisp playing regularly, but it will be interesting to see what the team does if and when Napoli pinch-hits in the games in Chicago; with very high-K relievers like Miller and Allen pitching at the end of the game, you could potentially argue for keeping Napoli in the game if he pinch-hits for Crisp, and trusting Miller and Allen to get outs without needing the team’s best defensive unit behind them.

But for both Maddon and Francona, I’d suggest the plan should be to go with their best defensive units to start the game, when those defenders are most likely to be put to the test, and then sacrifice offense for defense later when the dominant relievers start trotting in from the bullpen. Because relievers are getting used more heavily than ever before, you’re likely going to get the better hitter in the game for nearly as long as if he started and was lifted for a defensive replacement later anyway, only you’ll be lining up the skills of the players when they are most likely to impact the game.

We hoped you liked reading The New Postseason Plan: Defense Early, Offense Late by Dave Cameron!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Dave is the Managing Editor of FanGraphs.

newest oldest most voted
kevo8
Member
Member
kevo8

“Swapping out offense for defense with a late lead changes the way you’re likely to win, but doesn’t really change your likelihood of winning,”

I actually disagree with this point. Swapping in defense for offense lowers the expected amount of scoring in the rest of the game, which theoretically should lower the variability of each team’s final score. As the team in the lead, this is a good thing. On the other hand, if you swap in the offense, you increase the variability in runs scored by both teams, leading to more chances for the other team to come back.

In short, on average the margin of victory will probably be the same whether you swap out offense for defense or vice versa, but your probability of winning is probably higher with the former.

MPRox
Member
MPRox

Maybe I’m not very smart, but can you explain to me how the average margin of victory would stay the same but the probability of winning decreases? I understand the greater variability, but we’re talking one player for 3-4 innings or so. If the batter is leveraged as suggested ( poor hitter gets 1 AB, 4-5 innings in the field, good hitter gets 2-3 AB, 4-5 innings in field), that seems like a better overall outcome than the good hitter getting the same number of at-bats to start the game, but playing 7 innings in the field, while the bad hitter gets his 1 AB, but plays only two innings

mininimi
Member
mininimi

You don’t really care what the margin is, as a win is a win and a loss is a loss. So even if the expected margin of the game increases, more variability can mean a higher chance of losing. Think of it as two normal distributions: one with a mean of 2 and standard deviation 1 and another with mean 3 but standard deviation 2. The second one is more likely to be below 0 despite having a higher mean.

SirLancelittle
Member
SirLancelittle

I think the easiest way to understand it is this. Imagine two teams that, on average, out score their opponent by 3 points. These teams have the same average margin of victory so you might expect their winning percentage to be the same. But imagine that 1 of these teams is playing soccer. A 3 goal differential is huge because so few goals are scored. That team would be one the greatest teams of all time. Now imagine that the other team is playing basketball. A 3 point differential in basketball is almost nothing. This team would be close to .500. They might win by 6 one night and lose by 3 the next.