Projections show the Red Sox as a team of extreme possibilities, for good and bad

Most projections seem to agree: the Red Sox probably are not going to be very good this season. Over at FanGraphs, after making some recent additions, the Sox now have the 14th-most projected wins above replacement for this season. That’s not a team that’s any sort of lock to play meaningful games in October. Without Xander Bogaerts, without Nathan Eovaldi, and with Trevor Story out indefinitely with an elbow injury, the Red Sox just have too many questions at too many key positions to be considered reliable playoff contenders, much less postseason favorites.

Advertisement

But that word probably leaves some wiggle room, and if there’s something in your gut suggesting the Red Sox just might outperform these statistical projections, there’s actually some analytical logic behind your pie-in-the-sky optimism.

Is Triston Casas a Rookie of the Year candidate? Might Chris Sale actually be healthy and productive for the first time in years? Could Masataka Yoshida be one of the game’s best left fielders in his first big league season? Could it all come together at once and make this a better team than we’re expecting?

Sure! There is a reasonably attainable best-case scenario — actually, a bunch of reasonably attainable best-case scenarios — in which the Red Sox are pretty good, maybe even really good.

What’s dragging down their projections is the fact that they also have a lot of worst-case scenarios in which they’re, frankly, awful. The reality is, the Red Sox are far more volatile than most teams, with real uncertainty at many positions — and outcomes that could go in many different directions, for better or worse.

To explain the gap between these optimistic possibilities and cold-shower analytics, we brought in The Athletic’s resident numbers guru, Eno Sarris.

Are the Red Sox really as bad as the projections suggest? Is there any logical reason to maintain a more hopeful outlook? And can we explain the divide between these ho-hum projections and the “but what if!” best-case scenarios?

Take it away, Eno:

Sarris: Projections do have the Red Sox leading the league in one thing right now. They have the most players in baseball with a wide gap between their best-case and worst-case outcomes.

Dan Szymborski’s ZiPS projections offer 80th and 20th percentile outcomes to give an idea of how reliable they are for any single player. The 80th percentile projections are reasonable best-case scenarios (offensive results better than 80 percent of the algorithm’s estimates), and the 20th percentile projections are basically worst-case scenarios (worse than 80 percent of the predicted outcomes).

Advertisement

After counting how many players on each team have a 40-point gap in their league- and park-adjusted OPS (known as OPS+), the Red Sox floated to the top by a large margin.

Some of this is just a random trick of the depth charts. The Sox had a lot of free agents coming and going, and with several positions of upheaval on the depth chart, more players were awarded enough potential playing time to be included in the sample. Those questionable positions had multiple younger prospects associated with them (Marcelo Mayer and Nick Yorke are on this list), but those positions are also why the list features minor leaguers that aren’t necessarily “prospects”  like Niko Kavadas, Stephen Scott and Alex Binelas, who all have a large spread in their projections. And despite garnering some big league time, players like Narciso Crook also don’t give the projection systems that much in terms of major league results to work with.

The same is true for most of the names on this list — they’ll have to hit the upper end of their projections to even be relevant to the Sox this year. So let’s refocus and only show players that are in the first three chairs on the depth charts, to see who the most volatile actual Boston Red Sox players are according to ZiPS.

It’s worth noting here that the league’s median OPS+ gap between 80th and 20th percentile projections is about 38 points, so it’s normal to have a large spread between best-case and worst-case scenarios, but the Red Sox have a gap bigger than usual — sometimes a lot bigger than usual — at basically every position. The chart above lists only the 10 biggest gaps, but the list could have kept going. Right fielder Alex Verdugo also has a 39-point projection gap (132 to 93), as does rookie first baseman Casas (133-94). New center fielder Adam Duvall signed at such a time that he doesn’t show up in the publicly available 80th and 20th percentile projections, but as a 34-year-old coming off an injury, it’s safe to assume he has another larger-than-usual variation.

The most fascinating players with wide projections are young and old, but regardless are crucially important to this team’s outcomes. Turner’s on here because 38-year-olds sometimes can still hit, and sometimes they can’t, and projection systems can’t do much but look at the probability of both scenarios and produce the median outcome. But the reality is that he’s been better than 20 percent better than league average for nine straight years, if he’s still got it, he will likely out-produce the .285/.363/.442 line ZiPS has projected for him. It’s either he does better than that again or begins a quick turn down the aging curve, that actual line is unlikely. That’s why a team like the Red Sox might take a short-deal chance on him, and why he was available.

Advertisement

And then there’s the newest addition to this team, 27-year-old shortstop Adalberto Mondesi, who’s among the 20 players with the largest spread in projected outcomes (minimum 300 plate appearances) in the entire league. He’s been in the big leagues seven years, so there’s a track record, but the injury history has been so severe that he still doesn’t have 1400 plate appearances in the big leagues and the projection system probably is still searching to find his true talent. Strikeouts can also make a hitter harder to project, as strikeout-prone hitters are subject to more streakiness, and only 24 hitters have stepped to the plate 400 times in the last three seasons and had a higher strikeout rate than Mondesi. The shortstop has been as bad as 87 percent worse than league average and 13 percent better than league average, so he fits the bill. He was still an interesting acquisition because his top-end projection has him displaying above-average offense and defense at a position of need for the Red Sox.

Over on the pitching side, the Sox don’t have the same number of volatile players relative to the market, but that’s because these projections all assume a certain level of innings pitched projection. There are a few veterans in this rotation that, if they pitch to their projected innings total, should be fine. But obviously, they are a risk to even hit those innings projections. Here are the rotations with the most injury risk in baseball according to Jeff Zimmerman’s injury percentiles.

Their veteran rotation holds more than a little volatility itself, and will go as their health goes, most likely.

The same story is painted on both sides, in the end. This is a team with a large spread between its good and bad outcomes. Normally, that would be bad news for a Red Sox team, and yeah, it’s not great news overall. Questions at multiple positions, age in the rotation, unproven bats on the way in … the way to the bottom of the standings is pretty obvious in Boston.

The good news is the same as the bad news: this is a difficult team to project! And, as an organization, they must know it, having leaned into that volatility with their most recent additions. There is a universe that exists in which multiple players hit their 8oth percentile outcomes, and the rotation stays healthy, and this team surprises. In a good way.

Thanks Eno! So it’s clear there is some logic — and some actual numbers — to support those who take a more optimistic view of the Red Sox roster. There’s also logic — and numbers — to explain why projection systems aren’t fully embracing those best-case scenarios. Volatility is everywhere across this roster.

On the low end, ZiPS sees Casas as a rookie first baseman struggling to keep up with big-league pitching. On the high end, it sees one of the better first basemen in the game. His 80th-percentile slugging percentage is .495, which would have been better last season than every first baseman except Freddie Freeman, Paul Goldschmidt and Pete Alonso. So, which is it, a Rookie of the Year candidate or a player who might have to be demoted by mid-season?

Advertisement

That level of unpredictability is the norm in this lineup. New left fielder Masataka Yoshida signed out of Japan and has only a 37-point gap between his 80-percent OPS+ projection (148) and his 20-percent projection (111). According to ZiPS, he’s the most stable source offense in the Red Sox lineup, and he’s never played a game in the Major Leagues!

The pitching volatility seems harder to capture because of the workload issue, but do we really need a statistical analysis when Chris Sale and James Paxton haven’t pitched a full season in years, Brayan Bello, Garrett Whitlock and Tanner Houck have never spent a full year in a big league rotation, and Corey Kluber turns 37 in April and has his own injury issues to consider? That volatility speaks for itself. The Red Sox have loaded up on rotation depth to help counter such unpredictability, but that doesn’t make it any easier to predict who’s going to be pitching most of the innings and what those innings are going to look like.

Predictions come with the territory this time of year, and baseball produces so much data that there’s always a certain logic behind the objective expectations that seem carved into stone heading into spring training.

But the numbers also seem to recognize when they don’t have all the answers. This Red Sox team has created a pretty clear set of underwhelming expectations, but with a larger-than-usual caveat saying one thing loudly: When it comes to what they might actually do, we just don’t know for sure.

(Top photo of Alex Verdugo: Michael Dwyer / Associated Press)

ncG1vNJzZmismJqutbTLnquim16YvK57k2pqanFkbHxzfJFsZmlqX2V%2FcL7EnWSsp6hiw7C4wK2gpaGkmnqzu9KtnKtn