Postseason Preview: After Crushing Their Rival, the Red Sox Set Their Sights on the Rays in ALDS

The Boston Red Sox summarily dispensed with the New York Yankees in their winner-takes-all Wild Card matchup on Tuesday night, with Nathan Eovaldi cruisin’ and Gerrit Cole the subject of a bruisin’. Boston now faces the Tampa Bay Rays in a five-game American League Division Series matchup that represents just the third time the teams have ever squared off in the playoffs.

One of the interesting things about this matchup from a projections standpoint is that it features two of the teams with the largest gap between their FanGraphs Projected Standings and the ZiPS-concocted ones. Indeed, if it weren’t for the St. Louis Cardinals, these two squads would have the biggest separation in the two systems’ respective preseason outlooks, with the FanGraphs Projected Standings preferring the Red Sox and ZiPS leaning toward Tampa Bay:

The reasons for these discrepancies are tricky to pinpoint, but I suspect it comes down to how the systems treat injuries and depth. ZiPS doesn’t assume specific amounts of playing time, seeing playing time for regulars as the midpoint of a very skewed distribution (a player projected to get 600 PA will sometimes end up getting 300 PA but will never end up with 900 PA). A team’s cumulative days on the injured list is a very crude way to gauge the impact of losses. For example, the Braves are in the third of teams with the least days lost to injury, but nearly 20% of those days came from Ronald Acuña Jr. and Mike Soroka, massive losses.

In this case, the Rays were second in the majors in lost injury days with 2,544, while the Red Sox were the fourth healthiest team with just 907 lost injury days. My hypothesis — which I’ll try to confirm more vigorously this winter — is that ZiPS may do better with teams that lose key contributors since it’s explicitly assuming that there’s a chance of injury, while the FanGraphs implementation does better when the main cast stays healthy. In fact, this pattern was noted at the start of the year; Boston’s relatively pessimistic ZiPS projection mainly came not from the healthy scenarios but from those simulations in which the Red Sox started losing starting pitchers due to injury. As it turned out, only seven Red Sox pitchers started multiple times compared to 14 for the Rays, 11 for the Yankees, and 10 for the Blue Jays.

Entering the ALDS, there’s a substantial disagreement between FanGraphs and the oddsmakers, with the FanGraphs Playoff Odds 60/40 in favor of Boston beating the Rays, while Vegas is roughly 58%-42% in implied probability in favor of Tampa Bay. You may think that ZiPS will take the opposite approach again, but in this case, you’d be wrong! The ZiPS game-by-game odds are kind of in the middle of this one, seeing the series as nearly a coin flip, with a slight edge to the Rays:

The reason? The playoffs are an entirely different game. In short bursts, depth becomes less critical — though not unimportant — and teams get to leverage their best players more often relative to the regular season. That fifth starter a thin team dreads but can’t do without in July is likely to be buried in the bullpen or watching the game on television in October. The Red Sox, for example, are in the fortunate position of having Chris Sale‘s services, a luxury they only enjoyed for a small portion of the 2021 season. The Red Sox’s greatest strength is their offensive starters, and they get to utilize those guys in every game unless new injuries — I’ll address J.D. Martinez’s ankle in a moment — arise.

There are quite a few fun examples of the differences between regular season and playoff roster construction, and I’m going to highlight two of the most interesting ones. The first are the projections for this series if the Yankees had managed to upend the Sox. I’m going to present them with two different assumptions: the actual universe, in which Gerrit Cole would likely only be available for one start in the ALDS, and a magical one, in which the Yankees got to skip the Wild Card and go straight into a five-game set, allowing Cole to start two games:

Despite the Yankees projecting as the better team overall, they project worse than the Red Sox do against the Rays, simply because they aren’t able to fully utilize their best pitcher. And despite one of the rockiest outings I can ever remember from Cole, let’s not pretend he wasn’t the team’s best postseason option by a mile. While you could say the same thing about the Red Sox losing an Eovaldi start, ZiPS projects the non-Eovaldi Red Sox starters as superior to the non-Cole Yankees starters.

Let’s also warp back to the 2019 NLDS between the Nationals and the Dodgers. Like just about everyone else on the planet, ZiPS thought the Dodgers were a much better team than the Nationals in a regular-season format. But with Washington’s rotation and bullpen depth problems, which were quite severe at times during the regular season, mattering less in the playoffs, ZiPS saw the Nats riding the troika of Max Scherzer, Stephen Strasburg, and Patrick Corbin to a coin flip:

The Nationals, of course, ended up knocking out the Dodgers and everyone else they faced en route to a World Series championship. Throw the Red Sox and Rays together for another 162 games, and ZiPS would have the latter as the firm favorite. In a five-game series? Boston’s roster nullifies some of Tampa Bay’s impressive depth.

Two other narratives surround this series: the status of J.D. Martinez’s ankle and the Rays’ starters lack of playoff experience. One of these is important; the other is not.

Let’s start with the important one. On Sunday, Martinez turned his ankle tripping over second base as he was running out to right field. The injury was serious enough to necessitate leaving him off Boston’s Wild Card roster, a move the Red Sox wouldn’t have even entertained unless it was absolutely necessary. The acquisition of Kyle Schwarber at the trade deadline gave the Red Sox a potent bat they didn’t have at the start of the season, and Xander Bogaerts and Rafael Devers had very productive years at the plate, but that doesn’t mean the team won’t take a hit if Martinez’s ankle limits him. The above projections assume that Martinez is fine. This table, on the other hand, assumes that he’ll miss the ALDS completely:

Given the reports Wednesday that Martinez is feeling better, we’re leaving him on the roster, at least for now. If there’s a significant change to his status, the probabilities will naturally be updated!

But while Martinez’s availability is of great interest, the Rays’ young starters lack of playoff experience doesn’t really matter. That’s good news for Tampa Bay, as the team has already announced that Shane McClanahan and Shane Baz will go in the first two games and the buzz strongly suggests that Shane Rasmussen will get the nod in the third (ok, I know he’s Drew Rasmussen, but I like the idea of an all-Shane rotation).

While you can make the argument that young players with shorter careers are riskier overall, that risk goes both ways; even if we accept that Shane Baz is less of a known quantity than Eduardo Rodriguez, that goes for upside as well as the downside.

Since history rhymes, if not repeats, I made a very similar argument a decade ago when Matt Moore was pegged to start the first game of the 2011 ALDS against the Rangers with all of nine career innings in the majors to his name. At the time, I wrote a piece for ESPN looking at team performance and experience and found little evidence of teams with less playoff experience underperforming their playoff expectations. (Thankfully, Moore did not make me look stupid, scattering two hits over seven shutout innings.)

That finding holds true for individual pitchers as well. I do this every few years, but just to check, I created a simple model of postseason ERA constructed only from two variables: the players’ ERA from the regular season and the career innings of the pitchers involved. The mix that best predicted postseason ERA was >99.9% ERA and <0.1% career innings. I repeated the same using career postseason innings and got the same results: it just doesn’t matter. Sure, we don’t know if Baz will be Shane Youman, Shane Rawley, Shane Reynolds, or Shane Bieber, but there are just as many unexpected good results as bad ones. (I’ll stop saying Shane now, promise.)

Make no mistake, the Rays do have one of the least experienced rotations in playoff history. But placing a lot of faith in their young talent’s ability to succeed is one of the ways the Rays operate, so why mess with success? From Moore in 2011 to Randy Arozarena and Pete Fairbanks, this is what Tampa Bay does. Let’s suppose the series is resolved in three games. In that case, Tampa Bay’s starters will have, on average, the second-fewest career regular-season innings of any postseason team, behind only Cleveland’s 2013 squad, which played a single Wild Card game started by rookie Danny Salazar:

Before 1995, Marty Bystrom of the 1980 Phillies had the least prior experience, with only 36 career major-league innings before he pitched in the postseason. Pitchers with fewer than 100 innings only received 13 postseason starts through 1995 (0.8%), while there have been 65 such starts since (3.7%). Moore still holds the record at 9 1/3 innings. Incidentally, those teams went 29-32 in their playoff series, compared to the 29.4-31.6 record you would have expected based on regular-season records and home-field advantage.

So, who wins the series? One of the 21st century’s most successful postseason teams or the eternal underdogs searching for their first World Series champagne shower? Flip your coin and take your chances.