The prize for winning the offseason is usually presented … during the offseason.

Such is the case now with the onset of Projection Season, during which it seems likely that many and perhaps most statistical forecasting models will give the Red Sox, coming off a 78-84 campaign, the metaphorical trophy of "Most Improved Team of the Winter." Indeed, Fangraphs has already done just that, with an early projection identifying the Red Sox as a 92-win team, tops in the American League.

Those in the statistical business of projections and forecasts do *not* expect that the season will unfold exactly according to their initial depictions. Forecasts are probabilistic rather than absolute, with sizable swings representing a common occurrence rather than an aberration. Dave Cameron of Fangraphs tells Michael Silverman that it would come as little surprise to see teams fall 10 games on either side of these early projections.

Even acknowledging that the forecasting model has considerable inherent uncertainty, it would come as little surprise to encounter skepticism regarding a 92-win outlook for the Sox. After all, the team outperformed expectations significantly in 2013 and underperformed them drastically in 2010, 2011, 2012, 2014, and 2015.


That said, as Jeff Sullivan of Fangraphs wrote a year ago, teams that projection systems miscast by a sizable margin in one year (10 wins above or below their projections) tend to be forecast accurately in the following year.

There have been, however, some notable exceptions, as Sullivan noted in a more recent post, with the Pirates and Royals emerging prominently as teams that have beaten Fangraphs' projections by enormous margins in each of the last three years.

Travis Sawchik of the Pittsburgh Post-Gazette dove into a number of reasons why projection systems might be missing on the Pirates, and some of the reasons he identifies represent easily overlooked (and hard-to-measure) areas that could steer the Red Sox far from their projected won/loss total. Among them:


Bullpen improvement: Teams with lockdown bullpens in recent years such as the Pirates, Royals, and Orioles have typically jumped past their projected wins totals. With the additions of Craig Kimbrel and Carson Smith to Koji Uehara and Junichi Tazawa, the Sox bullpen has a chance to tack on W's beyond their projection.

Youth outperforming expectations: Could Xander Bogaerts fly past his projections and combine high averages with middle-of-the-order power? Could a Jackie Bradley Jr. or Rusney Castillo either significantly exceed or fall short of his projections? The volatility of youth is considered one factor in the recent misses seen in projection systems

Pitch framing: Will Blake Swihart be an average pitch-framer or perhaps take a step forward and become an above-average strike thief? Or might he prove a subpar receiver? Will the team end up entrusting a significant amount to pitch-framer extraordinaire Christian Vazquez? The ability to steal strikes – or lose them – can swing the performance of a pitching staff drastically relative to its projection.

Defense: There is a great unknown in the Red Sox' defensive universe, chiefly: How will Hanley Ramirez take to first base, particularly as it affects the rest of his infield? A year ago, his left field defense represented a great unknown that impacted the team in a fashion that no forecasting model could have predicted. For that matter, the well above-average defense of Mike Napoli at his transition to a new position in 2013 likely contributed to the Sox vastly exceeding their projected wins totals.


Health: The Red Sox' strong projected wins totals are a product of the track records of their starters. But one or two injuries at key positions could drastically alter the team's outlook. Or, as the team learned last year when losing Vazquez and Ryan Hanigan by the beginning of May, two key injuries at one position can create a black hole that siphons wins away from the projections.

In short: As Cameron noted in his early look at the projections, the forecasting exercise represents an interesting one in terms of establishing baseline expectations – with the understanding that there's a huge degree of variability surrounding them, and several areas where public forecasting models will remain limited.

Projections at this point tell us almost nothing about the season to come. As of now, as Peter Gammons writes, it's probably far more accurate to suggest that there's no dividing line between first and fifth place in the AL East.

Follow Alex Speier on Twitter at @alexspeier