The arrival of spring training brings with it a number of treats. If you’re a fan, it’s the hope that this might be the year your team wins the World Series. If you’re Yoenis Cespedes, it’s a cavalcade of rides that would make Birdman jealous. And for number crunchers, it’s a fresh batch of projections, ripe for statistical exploration.
Those projections — generated by algorithms such as Baseball Prospectus’s PECOTA1 — endeavor to predict the performance of each major league player and team in the coming season. That’s a tough job, and often a thankless one; if a projection system makes headlines at all, it’s usually for drawing the ire of a scorned team or fan base. (Never underestimate how quickly World Series joy can turn into anger when the computer calls for a sub-.500 record!) But projections are also invaluable because they provide a statistical snapshot, frozen in time, from which we can learn to become more accurate in the future, and suss out potential trends in the game.
And this year, PECOTA’s projections spat out something that may prove even more noteworthy than, say, the last-place finish they predict for the defending champion Kansas City Royals. The numbers suggest that MLB’s brief era of balance may soon be over.
For the last decade or so, the spread of wins in PECOTA’s projection2 has been trending downward. That meant the algorithm was forecasting win totals with less variability and more parity. With most teams clustered around 81 wins — a .500 record — it was becoming more difficult to make precise predictions about each team’s finish. Partly for this reason, team-level projections had their most inaccurate season since (at least) 1996.3 Factors like health and midseason trades, neither of which PECOTA can reliably predict, dominated any differences in talent among teams when it came to determining the outcome of the season.
This year, however, PECOTA is projecting a much larger spread in talent. The standard deviation of its forecast win totals is the highest it’s been since 2013, when (perhaps not coincidentally) its predictions were much more accurate. Since a wider spread in team records implies greater confidence about which teams will be good or bad, it’s fair to wonder what’s driving this newfound certitude after such a bad year for the stats.
Part of PECOTA’s confidence is based on improvements to the algorithm. For the first time, it’s incorporating the effect of pitch framing, or the ability of individual catchers to quietly improve the odds of a called strike through their receiving skills. A pitching staff’s projected earned run average can go up or down depending on the sleight of a catcher’s glove. So teams with good framers, such as Yasmani Grandal of the Los Angeles Dodgers, will see their projected ERA drop, while those with poor receivers, like the Philadelphia Phillies’ Carlos Ruiz, will allow higher ERAs.
But the wider distribution of predicted wins hasn’t been driven completely by tweaks in the algorithms. Las Vegas’s over/under lines tell a story similar to PECOTA’s, forecasting their widest spread in win totals since 2013. Savvy bettors have known about pitch framing for years, so clearly some other factor is driving the change. And the reason may be as simple as a few great teams coming out of rebuilding mode — and several bad ones entering it — at the same time.
Fully five teams project to win 90 or more games this season, compared with only three last year and two the year before. Some are to be expected, like the always-dominant-in-the-preseason Dodgers, and others have risen on the basis of defense (Tampa Bay) or pitching (Cleveland). But some, like the Chicago Cubs and New York Mets, simply represent the recent maturation of prospect-heavy farm systems.
For instance, PECOTA thinks the defending NL champion Mets will win 90 games, largely on the strength of a young, flamethrowing rotation projected to add about 13 wins above replacement.4 And the Cubs figure to win oodles of games no matter which source you use for your projections: PECOTA has them down for 94 wins, while FanGraphs’ projections are even more rosy, calling for 97 victories. Only two other teams in the last four years were projected by PECOTA to win 95 or more games: the 2014 and 2015 Dodgers.5 (By comparison, there were six such teams in the four years before 2014.) Chicago has overwhelmed the projection systems’ conservative streak by following their breakout 97-win season with the best offseason of any MLB team, plus enough reserves of youth, depth and prospect firepower to procure extra wins at the trade deadline if necessary.
The common thread is the transition from “building” to “built.” Graduating many of their most promising players to the majors in the last year or so, the Cubs and Mets are now seeing them play some of the best ball in decades. The concentration of so much young talent on so few teams, during an era where young players are more valuable than they’ve been in many years, has clearly played a role in the widening of win projections this season.
And at the other end of the spectrum, two teams are predicted to win fewer than 69 games: the Phillies and the Atlanta Braves. After failing to reach .500 these past few seasons, both teams have recently hired new general managers with long-term mandates. The new GMs (Matt Klentak in Philadelphia and John Coppolella in Atlanta) have embraced rebuilding as the new way of things, trading skilled veterans such as Andrelton Simmons to restock their minor league systems with prospects. In particular, the Phillies appear to be in for a rough year, with the worst win total projection of any MLB team since the Houston Astros in 2013. Like the Cubs and Mets at the top of the scale, these rebuilding teams are widening the spread of records at the bottom.
Any time teams deliberately move away from .500, in either direction, we should see a more variable league. Projections like PECOTA may be getting more accurate thanks to additions like catcher framing, but a lot of the spread in predicted wins comes down to the old-fashioned distribution of talent across the major leagues. And in a welcome change from recent seasons, this year should give us a handful of great teams battling for playoff spots, instead of a mass of mediocrity.