20 Feb 19
It’s hard to imagine things going more right for the Boston Red Sox than they did last season. Boston jumped out to a scorching 17-2 start, was 38 games over .500 by the All-Star break, posted the most regular-season wins (108) by an MLB team in 17 years, and then steamrolled through the playoffs with an 11-3 postseason record en route to a World Series title. Statistically, it was probably the most impressive performance any major team had in 2018.[footnote]Among teams for which we had have game-by-game Elo ratings and predictions — in men’s pro and college football and basketball and Major League Baseball.[/footnote]
But now the calendar has flipped to 2019, and as spring training warms up for the Sox in Fort Myers, Florida, Boston must focus on defending its crown — and staving off the inevitable regression that comes in the wake of a season as charmed as the one the Red Sox just enjoyed.
As a rule, clubs that win a crazy number of ballgames in one season tend to come back down to earth quickly in the next. Of the 32 teams that cracked the century mark in wins (per 162 games)[footnote]We’re including teams in strike-shortened seasons — like the 1994 Montreal Expos — whose wins would prorate out to at least 100.[/footnote] since 1990, 28 had an inferior record the next year,[footnote]One team — the 101-win 2003 Yankees — compiled the same number of wins the next season.[/footnote] and 24 failed to return to the 100-win club. (Thirteen failed to break even 95 wins.) On average, these 32 triple-digit winners declined by 9.6 wins the following season.
Teams that won substantially more than 100 games have tended to regress even harder. The 2002 Mariners, for example, won “only” 93 games after the 2001 squad tied a major league record with 116 wins; the 1999 Yankees won 98 a year after the team took home 114. The inescapable truth is that few major league teams actually have 100 wins of “true talent” on their rosters, much less 108. Most of these huge winners were aided by some not-insignificant amount of luck along the way.
And it’s hard to argue that the Red Sox weren’t one of the luckier teams in baseball last season. According to the Pythagorean expectation, a team with Boston’s runs scored and allowed should have won four games fewer than it actually did. Furthermore, a team with Boston’s particular statistical profile (its singles, doubles, walks, etc. — both for and against) should have had a Pythagorean record five games worse than it actually did. Add up those two categories, and the Red Sox benefited from an MLB-high 10 extra wins of luck, whether through prevailing in the relative toss-ups of close games or through stringing hits together (or stranding opposing runners) in an unusually favorable manner.
On top of all that, there’s another way a team can have everything go right for it, and that’s at the player level: Did everyone outperform their expected levels of performance at once? Injuries can often play a role here — though the Red Sox were in the middle of the pack in terms of man-games lost to the injured list. More pertinently, Boston also saw a number of players post career-best seasons last year, from American League MVP Mookie Betts (10.6 wins above replacement)[footnote]Averaging together the versions of WAR from Baseball-Reference.com and FanGraphs.[/footnote] to blockbuster free-agent signing J.D. Martinez (6.1), plus young up-and-comers such as Andrew Benintendi (4.1) and even longtime puzzles such as Eduardo Rodriguez (2.7).
Altogether, 12 of Boston’s 21 regulars (those who played at least 2 percent of the team’s available playing time)[footnote]As measured by plate appearances and (leverage-adjusted) innings pitched, scaled to maintain WAR’s implicit 58/42 split between position players and pitchers.[/footnote] exceeded their established level of WAR, with only Jackie Bradley Jr., Eduardo Nunez and the catching tandem of Sandy Leon and Christian Vazquez significantly undershooting their previous production levels during the 2018 regular season.[footnote]You could also argue that more should have been expected of Rafael Devers, who had 1.1 WAR in a partial season at age 20 in 2017 but produced only 0.5 WAR with more than double the playing time in 2018. But Devers was also only 21, playing his first season as an MLB regular.[/footnote]
And this is to say nothing of the unexpected performances the team received in the postseason from the likes of Steve Pearce — a fizzled-out former prospect who arrived in Boston via a midseason trade and ultimately won World Series MVP — or Nathan Eovaldi, another castoff who had a 1.61 ERA in 22 1/3 postseason innings. (Or, in general, the amazingly fortuitous splits the team had in crucial playoff situations.)
All of those different ingredients explain how a team that won 93 games in 2017 suddenly exploded for 108 and won the championship a year later. But again, the pull of baseball’s gravity is strong. Based on data since 1990, we’d expect a team that improved by 15 games between seasons to give back about 5.2 wins the next season. It’s just another data point to toss onto the heap of statistical indicators that foretell a decline for the Red Sox heading into 2019.
The good news for Boston is that if your starting point is a 108-win team, you have a ton of room to regress and still be one of the best teams in baseball. Even if the Sox didn’t truly have 108 wins of talent on the roster last year, they still played like a 98-win team according to their underlying statistics, and almost all of that team will be back this season (with the notable exception of closer Craig Kimbrel). According to an early preseason version of our 2019 MLB projections,[footnote]Not including Tuesday’s news of Manny Machado signing with the San Diego Padres, although that move has minimal implications for the Red Sox.[/footnote] we rate Boston as the third-best team in baseball, with a 95-67 projected record and a 10 percent chance of repeating as champs, which is also tied for third-best in MLB.
Trouble is, that might make the Red Sox only the second-best team in their own division. Our simulations consider the archrival New York Yankees just as likely as Boston to win the World Series and actually think that New York is ever-so-slightly better talent-wise. Although the Sox got the better of the Yankees last season, winning 13 of 23 games (including an August sweep and a four-game division series victory), for all intents and purposes, our projections have the two teams in an absolute dead heat as we look ahead to 2019:
The Red Sox still have a Yankees problem on their hands
How our preliminary Elo ratings are forecasting the 2019 AL East race
Avg. Simulated Season
Win World Series
Based on 100,000 simulations of the 2019 MLB season
Sources: Baseball prospectus, Fangraphs, Clay Davenport, Caesar’s Palace
And the Red Sox could be running out of time to make the most of their current core. By 2021, Betts, Bradley, Chris Sale, Xander Bogaerts and Rick Porcello (plus potentially Martinez, who has an opt-out clause) will have all hit free agency. And team president Dave Dombrowski built 2018’s champion in part by bucking MLB’s prospect-hoarding trend and emptying out the farm system’s next generation in favor of short-term wins, so reinforcements aren’t exactly on the way.
The result of Dombrowski’s moves was a championship, and one of baseball’s all-time great single season performances, so I’m pretty sure it was worth it. The question now is how steep the drop-off will be in 2019 — and beyond. In many ways, Boston caught lightning in a bottle last season, enjoying the kind of magical year that comes along only once every decade or so. But if history is any guide, the follow-up will have trouble coming close to matching the original.