Investigating Steamer’s Optimism for the Red Sox

On Monday, we released our first projected standings of 2016, with the Cubs unsurprisingly looking like the best team in baseball heading into Spring Training. More controversially, though, the Steamer projections — which is what our Forecasted Standings are based on currently, and will be until we add in the full ZIPS projections — see the Boston Red Sox as the second best team in baseball at this point, forecasting them for a 92-70 record.

That would be a 14 win improvement over last year’s 78-84 mark, which also notably came after the projections expected big things from the club. Understandably, there’s a decent amount of skepticism surrounding the idea that the Red Sox are really the AL’s best club on paper, so let’s look a bit deeper into the nuts and bolts of the forecast to see what’s really driving Steamer’s belief that Boston’s roster is ready to contend.

The first thing worth noting is that Steamer really likes Boston’s offense; they are projected to lead the majors in run scoring at 4.75 runs per game, a total of 770 runs over a full season. But this isn’t really a huge change from a year ago, because the Red Sox scored 4.64 runs per game in 2014 (748 total runs), and most of the core players are expected to do something pretty close to what they did a year ago. Essentially, the projections think the Red Sox are expected to score 22 more runs than they did last year, and it’s based almost entirely on two players. The players generally considered part of the team’s core roster are actually expected to hit a bit worse than they did a year ago.

Red Sox 2015 vs 2016
Player 2015 PA 2015 wOBA 2015 OFF 2016 PA 2016 wOBA 2016 OFF
Mookie Betts 654 0.351 22.5 630 0.361 21.6
Xander Bogaerts 654 0.338 10.9 651 0.339 8.8
Dustin Pedroia 425 0.347 6.5 602 0.332 2.9
David Ortiz 614 0.379 21.5 595 0.358 10.7
Blake Swihart 309 0.312 -2.3 320 0.303 -5.2
Jackie Bradley Jr. 255 0.355 6.2 525 0.320 -1.9
Rusney Castillo 289 0.283 -12.7 427 0.312 -4.6
Totals 3,200 0.344 52.6 3,750 0.336 32.3

Six of the seven guys listed above are projected to be less effective offensive players, with only Rusney Castillo expected to step up from terrible to mediocre. I don’t think anyone can look at those projections and think that Steamer is being unreasonably optimistic. Mookie Betts gets a small wOBA jump but loses half of his baserunning value. Bogaerts basically repeats his season. Ortiz starts to show his age. Swihart doesn’t improve at all. Bradley retains some of his improvements but still regresses heavily. There are no breakouts here, no huge improvements that are driving the team’s offensive gains.

Instead, the offensive jump is all about last winter’s two big free agent misfires.

Red Sox 2015 vs 2016
Player 2015 PA 2015 wOBA 2015 OFF 2016 PA 2016 wOBA 2016 OFF
Hanley Ramirez 430 0.308 -8.3 511 0.353 10.7
Pablo Sandoval 505 0.288 -20.0 560 0.330 0.1
Total 935 0.297 -28.3 1,071 0.341 10.8

Last year, Hanley Ramirez and Pablo Sandoval combined to hit like a league average shortstop; in 2016, they are projected to hit like a league average first baseman. That jump is worth almost 40 runs over 1,100 plate appearances, which not only cancels out the 20 run drop in offensive production from the other starters, but adds the additional 20 runs the team is expected to score in the aggregate. Essentially, the entirety of the offensive bump can be assigned to expected improvements from Ramirez and Sandoval.

Is this a reasonable expectation, based on their track records? I don’t really see why not. Ramirez’s jump back to a .353 wOBA is driven mostly by a BABIP correction, going from .257 last year to a projected .310 next year; his career BABIP is .327, so it’s not like there’s a lot of evidence that Ramirez should be expected to post a below average BABIP again. As a guy with a better-than-average strikeout rate and still decent power, it’s hard to see why Ramirez wouldn’t be expected to be an above average hitter. A .353 wOBA doesn’t feel at all out of line for a 32 year old with a career .371 mark, even though last year was obviously a disaster.

The story is mostly the same with Sandoval. Steamer is projecting his BABIP to jump back up to .303, from last year’s .270, based on his career average .307 BABIP. There’s also a small power spike included — his ISO jumps to .157, which would be his best mark since 2012 — so I could see taking the under on Sandoval’s offensive projection, and it’s worth noting that ZIPS isn’t as optimistic about Sandoval’s bounce back, projecting just a .310 wOBA, but even if you drop him down to the more pessimistic of the two forecasts, the net result is still going to be that the improvements from Sandoval and Ramirez will more than make up for the production losses elsewhere on offense.

Really, I don’t see a lot to quibble with here. The Red Sox offense projects to be the best in baseball because they have a bunch of guys who are good hitters, and their line-up doesn’t look to have any obvious black holes. The bench should be decent as well, especially with the disaster performances from Allen Craig, Deven Marrero, and Sandy Leon removed from the picture. The story of the improved Red Sox offense isn’t so much about rosy forecasts for their best hitters, but instead, replacing train wreck performances with okay ones. So I’m buying the ~770 run projection for the Red Sox; that doesn’t feel out of line to me at all.

Of course, that brings us to the run prevention side of things. Last year, the Red Sox allowed 753 runs; only five teams gave up more than that total. This year, Steamer projects the team to allow just 666 runs, an 87 run improvement. That’s a huge swing, and would take them from 25th to 10th in runs allowed across MLB. So let’s investigate the huge expected improvement on the pitching and defense side of things, and see if we can find any specific projections to quibble with there.

Let’s start with the defense, since that will affect all of the pitchers individual totals. Last year, the Red Sox posted a +3 UZR and a +3 DRS, so the defense was considered basically league average, even with Ramirez butchering left field and Sandoval failing to show any range at third base. This year, our projections have their fielding ratings jumping up to +23, as Steamer likes the Betts/Bradley/Castillo outfield and sees Pedroia as an elite defender second base. Most importantly, though, it is heavily regressing the disastrous performances from Ramirez and Sandoval, wiping out their combined -35 UZR/-30 DRS and replacing it with expected average defense at the infield corners.

For Sandoval, projecting just a -4 fielding rating seems about right to me; he’d been an average or slightly above average defensive third baseman in San Francisco, and when projecting defensive value, you need to regress single season data pretty heavily. For a guy with a career UZR/150 of -1, expecting something dramatically worse than that is probably an overreaction to 1,000 bad innings in the field. So I don’t see much of an issue with that forecast.

The Ramirez projection — expecting him to be a +3.2 defender at 1st base — is one where it seems likely that the limitations of the forecast come into view. In a vacuum, one wouldn’t think that projecting a guy who played shortstop two years ago to be a slightly above average defender at first base would be crazy, but in this case, I think humans know more about Hanley Ramirez’s defensive abilities than the numbers do. Ramirez was a bad defensive shortstop, a bad defensive third baseman when the Marlins tried him there, and the worst defensive outfielder anyone has ever seen; it seems pretty likely that Ramirez is going to be a bad defender at first base as well. It’s not like he’s taken well to position switches before, and his body just isn’t the same shape now as it was when he played shortstop, so I think it’s probably more reasonable to expect him to be a liability at first base, and maybe even a pretty big one.

We’re not into manually changing the projections — what’s the point of even having them if you just go in and “fix” ones that don’t match up with our preconceived ideas? — but this is one area where I think reasonable people can mentally add 10-15 runs to the Red Sox RA total. So instead of the pitching improvements needing to justify an 87 run improvement, maybe it’s more like a 75 run improvement, with some of the optimism resulting from a probably-too-optimistic projection on Ramirez’s defense at first base.

Still, though, a 75 run swing from the pitchers is a big number. Let’s see if we can figure out where most of that is coming from. While they outbid the rest of the league for David Price, and having a frontline starter is definitely going to help their rotation, the big change is actually in the bullpen.

Here’s what our current depth chart is projecting from the Red Sox bullpen in 2016.

#3 Red Sox

Craig Kimbrel 65.0 11.8 3.3 0.7 .301 77.7 % 2.63 2.67 1.8
Koji Uehara 65.0 9.1 2.1 1.0 .292 79.3 % 2.94 3.41 0.8
Carson Smith 55.0 9.8 3.2 0.6 .309 74.2 % 3.04 3.01 0.9
Junichi Tazawa 55.0 8.4 2.5 1.0 .300 76.0 % 3.30 3.58 0.4
Joe Kelly 45.0 7.7 2.9 0.8 .308 72.3 % 3.73 3.75 0.2
Robbie Ross 40.0 7.6 2.9 0.8 .307 73.2 % 3.57 3.72 0.1
Matt Barnes 35.0 8.3 3.2 1.0 .303 73.9 % 3.74 3.89 0.0
Tom Layne 30.0 8.0 4.1 0.8 .305 72.4 % 3.89 3.99 0.0
Heath Hembree 25.0 7.9 3.4 1.1 .298 74.5 % 3.86 4.19 0.0
Noe Ramirez 20.0 7.0 3.5 1.0 .301 72.6 % 4.12 4.38 0.0
Steven Wright 15.0 6.6 2.8 1.1 .302 72.0 % 4.13 4.28 0.0
Edwin Escobar 10.0 7.3 3.0 1.0 .302 73.5 % 3.83 4.08 0.0
Pat Light 10.0 6.9 3.9 1.0 .302 71.8 % 4.31 4.47 0.0
Williams Jerez 10.0 7.4 4.0 1.0 .302 72.6 % 4.13 4.33 0.0
Total 474.0 8.7 3.0 0.9 .302 74.8 % 3.38 3.56 4.1

Like with the hitters, I don’t see any individual projections there that look overly optimistic. Kimbrel remains excellent, but Uehara and Smith are projected to pitch a lot worse than they did in 2015, and Tazawa is forecast to be just about his usual self based on his career numbers. This has Joe Kelly pitching okay-but-not-great out of the bullpen, though with his velocity, it’s not hard to see him being one of the types of pitchers who takes a big step forward in a relief role, if that’s how the Red Sox choose to use him. Ross, Barnes, and all the rest are basically replacement level fodder. There’s just nothing here that I would think is painting way too rosy of a picture about the team’s relief corps.

And yet, this performance would be a monstrous upgrade over last year’s bullpen. A year ago, the Red Sox relievers allowed 258 runs in 501 innings, so this projection of ~200 runs allowed makes up for the great majority of the improvement expected from the Red Sox pitching staff. It’s not just that Kimbrel and Smith were quality acquisitions, but that so many of the depth guys the team tried last year were total garbage, and there’s no real reason to expect that the team is going to end up giving significant innings to so many below replacement level pitchers again. And Kimbrel and Smith will help too, obviously. The Red Sox bullpen was among the worst in baseball a year ago; now it looks like one of the best. And I don’t see anything wrong with believing the forecasts here.

Now, let’s look at the rotation. This is where a lot of the attention is going to be, because so much of the narrative about the 2015 Red Sox surrounded Ben Cherington’s decision to go for depth rather than make a play for an ace starter, and then Dave Dombrowski immediately changed course and spent $230 million to bring in David Price. If the Red Sox are a contender this year, a lot of the credit is going to be bestowed to having an ace to carry the rotation, but that’s actually not where a lot of the improvement is expected to come, at least not according to our projections.

Last year, Red Sox starters put up +12.1 WAR by FIP or +11.1 WAR by RA9, ranking squarely in the middle of the pack regardless of which methodology you want to use. This year, with Price replacing Wade Miley and Roenis Elias being given some of the innings that went to Joe Kelly and Rich Hill, the team is projected for all of +14.5 WAR. It’s an improvement, as you’d expect from any rotation that added David Price, but everyone else is expected to be pretty mediocre; Porcello’s improvement is offset by regression from Eduardo Rodriguez, and Clay Buchholz is expected to be less valuable even while throwing 50 more innings. Price helps, but Miley wasn’t useless — his ERA was artificially high because almost every single run he allowed was categorized as “earned” — and the rotation improvements only account for about 20 to 30 runs.

Even with David Price in the fold, Steamer looks at the Red Sox as having just an okay rotation; it’s certainly not the strength of the club. And the starting pitching wasn’t the disaster that caused the 2015 Red Sox to implode anyway, so the fact that the improvement there is only minimal isn’t actually a huge deal. The rotation should be better than it was, but I don’t see Steamer being overly optimistic about this rotation, and I’d imagine a lot of Boston fans will think that Rodriguez projection is entirely too pessimistic; ZIPS is much higher on the team’s young arms, for instance.

But an okay rotation with a very good bullpen is an above average pitching staff, which is why Steamer is projecting the Red Sox for the 7th most pitching WAR of any team in baseball. We should likely downgrade their run prevention totals to account for the Ramirez fielding projection, but even moving them to something like 4.25 runs per game leaves them roughly average in terms of runs allowed, and this should be offense that scores a good amount of runs.

Steamer has the Red Sox at 4.75 R/G and 4.11 RA/G, which translates out to a 92-70 record. I’d probably adjust those to something like 4.7 and 4.2 R/G and RA/G personally, but that still translates to something like an 86 to 88 win team. Besides a probably-flawed defensive projection for Ramirez at first base, and maybe an overly optimistic power bump for Sandoval, I don’t see that many areas where I think the projections are overly optimistic. I don’t think the Red Sox are the second best team in baseball, but with a very good offense, an average defense, a decent rotation, and a very good bullpen, this Red Sox team really does look like a contender.

We hoped you liked reading Investigating Steamer’s Optimism for the Red Sox 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

How much do the projections take into account strength of schedule or individual opponents? Just in general, I mean.

For instance, being in a division with no obviously terrible team to beat up (as the Red Sox are) on will mean they have fewer actual wins than a team with a couple of terrible division opponents to feast on (like the Pirates have). Even with equally talented teams, the division and who you actually play with impact actual wins and loses heavily.

This does not change who wins the division (since everyone in the division is in the same boat), but it does change wild card winners and overall records.

I’m curious if the projection systems account for this.