Looking back at 2015 Colorado Rockies PECOTA projections

With less than a month until pitchers and catchers report to spring training, we've almost made it through another long, cold winter. The 2016 season being right around the corner also means that there's a brand new set of player projections right around the corner. We already covered what Marcel the Monkey thinks about the 2016 Rockies.

One of my favorite projection systems is Baseball prospectus' PECOTA. Unlike most systems, PECOTA provides a range of potential outcomes for each player. The 10th percentile projection is a player's realistic worst-case scenario, while 90th percentile is the realistic best-case scenario. In most cases, we only get the median projection. A range like this not only offers a wide spectrum of things to expect from a certain player, it acknowledges that projections are not predictions.

As we did prior to the 2015 season, we'll be taking a look at these projections when they are released, but something equally fun is looking back at the projections from the previous season. Who outperformed their projections? On the other hand, who underperformed?

Before we get started, let's talk about a few of the stats listed here that many people aren't familiar with.

TAv (True average) is Baseball Prospectus' measure of offensive value relative to average. TAv is both park and league adjusted, with .260 being league average. For TAv, a higher number is better.

VORP (Value Over Replacement Player) is an estimate of the number of runs above average a player provides over a replacement level player at the same position with the same number of plate appearances.

Now that we all know what we're looking at here, let's get into it.

The Overachievers

Nolan Arenado

Actual: .287/.323/.575, .299 TAv, 47.1 VORP, 7.5 WARP

90th percentile: .318/.357/.509, .292 TAv, 32.6 VORP, 5 WARP

Charlie Blackmon

Actual: .287/.347/.450, .272 TAv, 31.1 VORP, 2.7 WARP

90th percentile: .311/.362/.475, .285 TAv, 32.4 VORP, 2.7 WARP

DJ LeMahieu

Actual: .301/.358/.388, .253 TAv, 16.9 VORP, 1.7 WARP

70th percentile: .295/.330/.397, .249 TAv, 11.4 VORP, 1.7 WARP

Ben Paulsen

Actual: ..277/.326/.462, 269 TAv, 8.5 VORP, 0.8 WARP

70th percentile: .274/.329/.512, .272 TAv, 5.1 VORP, 0.8 WARP

Nick Hundley

Actual: .301/.339/.467, .268 TAv, 20.3 VORP, 0.8 WARP

80th percentile: .281/.335/.457, .270 TAv, 16.4 VORP, 1.9 WARP

Several players exceeded expectations here, but Arenado is the one who really jumps off the page. He not only did better than the average projection, he even blew the 90th percentile "best-case scenario" projections out of the water. In a perfect world in which almost everything went Arenado's way, PECOTA thought he'd have a 5 WARP season. In reality, Arenado's 90th percentile WARP projection was closer to his 50th percentile projection of 2.7 WARP than it was to his actual, real-life performance of 7.5!

This could mean a few different things. Arenado may have just had an outlier season that won't ever be replicated, PECOTA may have underestimated Arenado, or, most likely, it was a combination of the two. When the 2016 PECOTA projections come out, it will have accounted for what he did in 2015.

Charlie Blackmon also somewhat strangely hit his 90th percentile projection. It isn't so strange to see him put up the numbers he did,  but it's strange to see his mean projections be so low after having very similar offensive seasons in 2013 and 2014. This may be a case of PECOTA overreacting to Blackmon's poor debut in 2012, or perhaps it is a belief that for some reason Blackmon's last two seasons weren't indicative of the player he would be going forward. In any case, it may be time for PECOTA to adjust Blackmon's projection.

Beyond Arenado and Blackmon, we have three more overachievers in LeMahieu, Paulsen, and Hundley, who all hope to repeat their strong seasons in 2016.

Not everyone did better than expected, though. As happens just about every season, while some players exceeded expectations, others fell short. Most of these guys are players one might expect, with one obvious and notable exception.

The Underachievers

Carlos Gonzalez

Actual: .271/.325/.540, .284 TAv, 24.4 VORP, 2.2 WARP

30th percentile: .276/.333/.491, .280 TAv, 22.5 VORP, 1.8 WARP

Corey Dickerson

Actual: .304/.333/.536, .288 TAv, 10.8 VORP, 1 WARP

30th percentile: .277/.321/.482, .272 TAv, 14 VORP, 1.2 WARP

Rafael Ynoa

Actual: .260/.277/.339, .210 TAv, -2.2 VORP, -0.3 WARP

20th percentile: .237/.281/.338, .212 TAv, -2.8 VORP, -0.1 WARP

Brandon Barnes

Actual: .251/.314/.341, .229 TAv, -1 VORP, 0 WARP

20th percentile: .233/.273/.371, .220 TAv, -0.9 VORP, 0 WARP

Daniel Descalso

Actual: .205/.283/.324, .225 TAv, -1.3 VORP, -0.2 WARP

20th percentile: .231/.289/.337, .219 TAv, 0 VORP, 0 WARP

If you're like me, you didn't expect to see Carlos Gonzalez on the underperformers list after hitting a career high 40 home runs, playing in a career high 153 games, and posting his highest slugging percentage since 2010. A slow first month and a half and a low on-base percentage due to a low BABIP ended up hurting his offensive and, in turn, overall value, which led to Gonzalez being stuck between his 30th and 40th percentile projections. It will be interesting to see what his 2016 projection looks like after having a healthy, productive, and yet somehow disappointing, year in 2015.

One other notable player on this list is Corey Dickerson. Injuries that limited him to just 65 games in 2015 hindered his overall production and kept him close to his 30th percentile projection, but when healthy he was right in between his 50th and 60th percentile. I'd expect similar projections for him in 2016, with perhaps a slight reduction in games played.

After these two, it wasn't very difficult to peg Ynoa, Barnes, and Descalso as underachievers. Notably, however, they underachieved their already low bars. They'll all look to rebound in 2016 with what will likely be more pessimistic projections than they received prior to 2015.

★ ★ ★

In any given season, a team will have its share of players who do both better and worse than expected. Some of these are more predictable than others -- Arenado breaking out in his age-24 season was intuitively more likely than Hundley having a better than expected season at age-31 -- but it can happen to anyone, as making projections is an inexact science. The range of possibilities they account for acknowledge imprecision. So when PECOTA's 2016 projections will come out, it's important to remember the limitations of projections, and to look above and below the median forecast.