At the back of the BP Annual, on page 562, among the PECOTA leaderboards, there’s one table for WARP Declines and another for WARP Improvements. The guys on these lists are a hodgepodge of stories, but mostly these players are on the list because PECOTA hasn’t changed its mind on them. Colby Rasmus had a good year last year? PECOTA acknowledges it, but it hasn’t changed its mind about Rasmus. Decline. Dan Uggla was terrible last year? PECOTA adjusts downward some but basically hasn’t changed its mind about Uggla. Improve. The WARP decline/improvements tables are essentially regression leaderboards. These are guys who did something unexpected but, in PECOTA’s estimation, didn’t really change.

But some players did change. Some players had good years that so shook PECOTA’s understanding of them that the system concedes, in fact, that they are different players than they used to be, or different players than we thought they were. We don’t have a table for those guys in the book, but we’ll have one here. These are the players whose PECOTA projections for 2014 are most different than their PECOTA projections for 2013 were.

Player 2013 WARP projection 2014 WARP projection Difference
Yasiel Puig -0.3 3.3 3.6
Jean Segura 0.8 4.2 3.4
Matt Carpenter 2.3 5.3 3.0
Josh Donaldson 1.1 3.5 2.4
Manny Machado 1.1 3.5 2.4
Anthony Rendon 0.4 2.7 2.3
Evan Gattis 0.4 2.5 2.1
Kolten Wong 0.2 2.2 2.0
Khris Davis 0.7 2.6 1.9
Shane Victorino 2.1 4.0 1.9

(Note: PECOTA gets some tweaks from year to year, so this isn't a perfect apples-to-apples comparison, but it's at least like comparing blue-gray crayons to gray-blue crayons.)

Puig was BP's 79th-best prospect in baseball before the 2013 season; immediately following him on that list were Jonathan Schoop, Chris Owings, and Michael Choice, all of whom were within a year of Puig, age-wise, and all of whom actually projected to be more valuable in 2013: 0.2, 0.0 and 0.4 WARP, respectively. Naturally, when all we have on a player are nine games in Rookie ball, 14 in High-A, and a mysterious scrapbook of numbers sent from Havana, his projections are going to be subject to whiplash. PECOTA’s estimation of Puig’s defense and baserunning actually declined, but his projected batting line goes from .224/.272/.329 to .289/.354/.488. From Alfredo Griffin’s career line to Reggie Jackson’s.

Segura is a bit more complex, since he had plenty of history for the system to work with already. But PECOTA’s reassessment mirrors the larger baseball world’s: Even a year ago, there were some (the Angels, for instance) who doubted that Segura would ever be an average defensive shortstop. PECOTA projected him to be a -1 defender in 2013, but after challenging Andrelton Simmons for the FRAAing title, Segura now projects to +12. Overall, Segura’s projection is probably PECOTA’s most aggressive, both on defense and offense, where a .263/.304/.387 projection turned into a .282/.323/.403. While we out here focus on his second-half slide, PECOTA recalls his healthier minor-league walk rate and bundles it with some of his power outburst of 2013.

Rendon and Wong are a bit surprising because neither seems like he had a great 2013 season, or at least a season that made you reconsider him. But Rendon outhit his major-league projection by 15 points of True Average, undid whatever had caused PECOTA to project a miserable .271 BABIP for him, and in a partial repeat of Double-A Harrisburg upped his OPS by almost 400 points, just about to Calvin Pickering levels.

Wong, meanwhile, mostly benefits from a tripling of projected playing time. On a rate basis, his gains are very modest.

Victorino has been around so long and been so steady, that it seems impossible to have a new opinion of him. Indeed, PECOTA’s projected TAv for him is exactly the same as it was a year ago (though his final month bumped his projected hit-by-pitches by three this year). But last year, based on his history as a center fielder, he projected to be a poor defensive outfielder. After a spectacular defensive season in right, he projects this year to have baseball’s second-highest FRAA, behind Andrelton Simmons.

And the losers:

Player 2013 projection 2014 projection Difference
Hiro Nakajima 2.1 -0.4 -2.5
B.J. Upton 3.1 0.7 -2.4
Albert Pujols 6.9 4.7 -2.2
Paul Konerko 2.6 0.7 -1.9
Jesus Montero 1.9 0.0 -1.9
Jimmy Rollins 2.9 1.1 -1.8
Lance Berkman 2.8 1.0 -1.8
Matt Kemp 4.5 2.9 -1.6
Kevin Youkilis 2.9 1.3 -1.6
Justin Morneau 1.8 0.3 -1.5

The 2013 Annual comment for Nakajima concluded that “the multimillion-dollar question is whether the bat will [thrill] Oakland crowds with the swagger of his bat-flip on home runs, or if he will follow in the offensively empty footsteps of fellow NPB middle infielders Kaz Matsui and Tsuyoshi Nishioka.” His PECOTA comps this year show how optimistic even the final 14 words of that comment were: Juan Melo, Brian Barden, Ramon Nivar.

PECOTA has loved Upton forever. His projected WARPs going back to Joe Maddon’s first season:

2006: 4.6 WARP
2007: 4.3
2008: 5.5
2009: 4.2
2010: 3.6
2011: 1.9
2012: 3.2
2013: 3.1

The damage that 2013 did to PECOTA’s affections is moderate on offense—nine points of True Average, with small dips to his BABIP, walk rate and doubles rate—but scorched earth on defense, where his forecast dips 10 runs from last year.

Pujols’ drop makes him still the ninth-best player in baseball, according to PECOTA. More: This is the second year in a row he’s been near the top of this list. Here are the same leaderboards as above, except showing the cumulative difference from 2012 to 2014:

Player 2012 projection 2014 projection Difference
Mike Trout 0.6 6.8 6.2
Jean Segura 0.4 4.2 3.8
Brad Miller -1.0 2.8 3.8
Matt Carpenter 1.6 5.3 3.7
Andrelton Simmons 0.6 4.3 3.7
Billy Hamilton -0.3 2.9 3.2
Manny Machado 0.4 3.5 3.1
Bryce Harper -0.2 2.7 2.9
George Springer -0.9 1.9 2.8
Wil Myers 0.5 3.2 2.7

Player 2012 projection 2014 projection Difference
Albert Pujols 8.4 4.7 -3.7
Alex Rodriguez 4.3 1.2 -3.1
Kevin Youkilis 4.1 1.3 -2.8
Lance Berkman 3.8 1.0 -2.8
Jesus Montero 2.8 0.0 -2.8
Carlos Pena 2.6 0.1 -2.5
B.J. Upton 3.2 0.7 -2.5
Clint Barmes 2.5 0.3 -2.2
Rickie Weeks 3.1 0.9 -2.2
Danny Espinosa 2.8 0.7 -2.1

Put another way: Matt Kemp is on the 2014 losers list after dropping from projected superstar to projected All-Star, and his high-water WARP projection was lower than Pujols’ is now, after two huge drops in a row.

Most of the PECOTA gainers were young players, guys who hadn’t previously given PECOTA a ton of information to work with—nine of the 10 either debuted or played their first full seasons in the majors in 2013. Most of the PECOTA losers, meanwhile, are older guys for whom age can act as something of a multiplier on declines. Further, the young guys are generally adding playing time into their projections, while the olds are losing. So as much as Victorino was an outlier on the first table, so is Montero on the second.

Most of that drop comes from the switch from catcher to 1B/DH, though his offensive projection did take a hit after a season split between Seattle (where he was terrible) and Tacoma (where he was bad). Montero, who will be just 24 this year, manages to rank high on the three-year-PECOTA-projection-drop leaderboard, too:

Player 2011 projection 2014 projection Difference
Mike Trout -0.7 6.8 7.5
Paul Goldschmidt -1.1 4.0 5.1
Anthony Rizzo -1.0 2.9 3.9
Freddie Freeman -1.2 2.3 3.5
Jean Segura 1.0 4.2 3.2
Avisail Garcia -2.8 0.4 3.2
Brett Lawrie 0.6 3.7 3.1
Mark Trumbo -0.6 2.5 3.1
Andrelton Simmons 1.2 4.3 3.1
Billy Hamilton -0.1 2.9 3.0

Player 2011 projection 2014 projection Difference
Alex Rodriguez 4.6 1.2 -3.4
Albert Pujols 7.7 4.7 -3.0
Kevin Youkilis 3.9 1.3 -2.6
Lance Berkman 3.4 1.0 -2.4
Ryan Howard 3.4 1.1 -2.3
Mark Reynolds 2.4 0.1 -2.3
Justin Morneau 2.4 0.3 -2.1
Jesus Montero 2.1 0.0 -2.1
Jason Bay 2.0 0.9 -1.1
Mike Hessman 2.1 0.1 -2.0

Six of the 10 player on this year’s loser list, and eight of the 10 on the three-year drop list, are first basemen or DHs. Maybe that’s just because old guys are more likely to have fallen to the end of the spectrum, or maybe it’s that playing a skill position provides a bit of ballast that makes capsizing less likely, in PECOTA’s eyes.

You need to be logged in to comment. Login or Subscribe
Doesn't projecting guys like Youkilis, Berkman and A-Rod, who won't play at all in 2014, kind of undercut PECOTA's practical credibility. And when I see Justin Morneau, who has a job in a hitter's paradise for 2014 drop 2.1, I wonder how sensitive PECOTA is to real world data.
Just because Youkilis and Berkman won't play, doesn't mean that they couldn't play. In regards to A-Rod, any performance model that can predict suspensions would be nearing singularity-level advancement. As for Morneau, WARP is context-sensitive, so his job in a hitter's paradise is already factored into WARP itself.
Right, retired/suspended players aren't included in our depth charts and don't affect team projections. We just leave them in the spreadsheet in case people are curious about what they would do if active.
So, just to be clear, they won't show up in PFM? (I like the projections for them, just want to adjust my draft prep if necessary.)
Right, not in PFM. PFM is based on the depth charts.
Interesting stuff. PECOTA is still too bullish on Pujols, methinks. Thanks!
Maybe PECOTA should be using a decaying AROD as a better PED comp for Pujols.
Enjoyed this piece...
"Six of the 10 player on this year’s loser list, and eight of the 10 on the three-year drop list, are first basemen or DHs. Maybe that’s just because old guys are more likely to have fallen to the end of the spectrum, or maybe it’s that playing a skill position provides a bit of ballast that makes capsizing less likely, in PECOTA’s eyes." If it is because old guys fall to the end of the spectrum, does that indicate that PECOTA is having trouble accounting for ageing, or alternatively, that there is more uncertainty in the rate of decline than in the rate of player development? I am curious where these players fell in their PECOTA percentiles.
That is an awesome question - is it more difficult to project accurately on development in players or on decline?
Does any of the increase in Rendon's WARP have to do with a positional change? I assume that 0.4 at the beginning of 2013 was comparing him to other third basemen, while the 2.7 this year is vs. second basemen...
I like how Carpenter isn't discussed. He's just listed in the tables, because it happened and there isn't much else to say about it. :)
Like the question above, I wonder whether PECOTA values older players too much in general, or underestimates the decline phase? I know this is anecdotal, but it seems that there are far more older players who underestimate their projections (see above lists) than exceed them. Or to put it another way, a poor season from an older player seems to be taken as an outlier rather than a "new normal".
My theory is that the data is skewed by the steroid era. In the past there were numerous players in their mid 30's benefiting from PED's that are now acting as com parables.
I wonder whether Bryce Harper is a meaningful example of PECOTA foreseeing a leap. I imagine it will always predict a leap between an age 19 year and an age 21 year, no matter who the player is. I'm also curious about PECOTA's projection of 2.7 WARP for him, just for the sake of discussion. I'm not one to dissect every projection, nor expect every one to be completely fathomable in February. But it does seem odd on its face that he would decline in absolute and rate terms from an age 20 year in which he was injured most of the year. Maybe age 20 injuries have generally signified trouble developing from there.
I have Alfredo Griffin's autograph. That's my substantantive contribution to this thread.
Good stuff. Would you please do this same exercise for pitchers?