3 players whose PECOTA and Steamer projections disagree

Projections are great, insofar as they present a starting point for preseason looks at lineups, and parsing which teams look poised for the postseason and which look doomed for the doldrums. But what happens when two major projections disagree?

Baseball Prospectus' PECOTA and FanGraphs' Steamer projections are released every season, and give a glimpse into what players may accomplish in the upcoming year. Both projection systems use previous performances as well as traditional aging curves and regression to the mean to make an educated, unbiased guess about what each player might do. However, the two differ in subtle ways. For instance, PECOTA uses comparisons to other players, and has a higher valuation of what "replacement level" means.

Curated below are three players - a hitter, starting pitcher, and reliever - the projections don't agree on, and what those discrepancies mean.

Rick Porcello, SP - Boston Red Sox

With a difference of 1.4 wins above replacement, Steamer thinks Porcello will be 64 percent better than PECOTA's estimations. Let's break down the difference even further.

By PECOTA's projections, Porcello is better by ERA and actually has a better strikeout rate as well. Interestingly, though, it seems to attribute a lot of Porcello's value based solely on the size of his workload. PECOTA's relative pessimism has him tossing 13 fewer innings, putting up a slightly worse walk rate, and allowing roughly 10 extra baserunners due to a slightly higher projected WHIP than Steamer's estimations.

It's worth noting that Porcello has averaged just under 200 innings over the past three campaigns. Entering his age-28 season, Steamer's projection would be Porcello's second-best season by FanGraphs' WAR since breaking into the league in 2009. Meanwhile, PECOTA seems to believe Porcello is poised for only his fifth-best season. Out of a possible nine seasons - including the upcoming one - that's not great.

Byron Buxton, OF - Minnesota Twins

This time it's Steamer that has the more pessimistic projection, and it isn't particularly close, either. PECOTA thinks Buxton will be nearly twice the player that Steamer projects this year.

PECOTA has Buxton set for 17 homers and 16 stolen bases with a .244/.296/.434 slash line. Meanwhile, Steamer has the 23-year-old center fielder slated for a .243/.298/.410 line with the same number of stolen bases and just one fewer home run. So, if the offensive numbers are that close, why such a large discrepancy in WAR? Turns out Steamer doesn't believe in Buxton's defense as much, projecting just 4.4 runs saved. PECOTA, on the other hand, thinks he'll provide 16 runs of defense, or in simpler terms, almost exactly 1.2 wins above replacement more than Steamer's estimate.

Sam Dyson, RP - Texas Rangers

Among closers, Steamer and PECOTA agree quite a bit. At face value, that makes sense. When you're dealing with 60 innings pitched, the highs and lows likely aren't going to differ that much. However, with Dyson, Steamer thinks the Rangers reliever will be only half as good in 2017 as PECOTA projects.

According to PECOTA, Dyson will post a 3.36 ERA and 1.30 WHIP over 62 innings pitched. Steamer, however, thinks he'll put up a 3.63 ERA and 1.35 WHIP over 65 innings of work. Regardless of what happens, those ERA and WHIP numbers would be the worst of Dyson's previous three seasons, which is where regression to the mean comes in. Projecting ERA for a pitcher is extremely difficult, and there can be a lot of noise involved in the ERA of a player who only pitches 70 innings. In this instance, it doesn't matter which projection system you choose to subscribe to, Dyson is pegged for some regression going into his age-29 season.

(Photos courtesy: Action Images)