I recently wrote about my attempt to design an indicator that would predict when players were at a higher risk for having a collapse-type year. I named the metric CLIFFORD, referring to the fact that players identified by it were at risk of falling off a cliff offensively. My inspiration was Adam Dunn and his disastrous 2011, in which his wOBA declined by .113.
My initial research showed that 58% of collapse candidates identified by Marcel actually experience a wOBA decline of at least .03 (or 30 points)–2.43 times the likelihood of non-collapse candidates. Collapse candidates identified by CLIFFORD actually decreased by at least 30 points of wOBA 53% of the time–2.14 times the likelihood of non-collapse candidates.
Marcel initially appeared to do a better job identifying these candidates. If we knew nothing else outside of just the Marcel projection, our chances were better at identifying collapse candidates than if we used CLIFFORD (and, yes, the difference between the relative risk for both measures is statistically significant).
However, and here’s the bright spot, there was not much overlap between the two metrics.
Out of the 34 collapse candidates identified by CLIFFORD, only three were also identified by Marcel. My thinking on the difference between the two is simply that while both are trying to capture aging, Marcel relies on a known, more gradual aging curve while CLIFFORD is trying to find players that–for one reason or another–are poised to age faster.
As a quick refresher, players are given one point if their performance in any of the following categories declines between Year 1 and Year 2 by an amount worse than the 25th percentile for change across the league:
|Metric||25th Percentile||Relatve Risk for Collapse|
Players with three or more points are then classified as CLIFFORD candidates. Calculating CLIFFORD values for players entering 2013 produced five collapse candidates:
|Name||2012 Age||2012 wOBA||2012 Z-Contact% (pfx)||2012 Spd||2012 FA% (pfx)||2012 UBR||2012-2011 Z-Contact% (pfx)||2012-2011 Spd||2012-2011 FA% (pfx)||2012-2011 UBR|
Now, 2013 Marcel projections are not out yet so we don’t know how much overlap there will be between the two. I could see Cano’s wOBA declining simply due to age (he’s entering his age-30 season) and basic regression, for example. A drop of 30 points would put his 2013 wOBA at .364. Even adjusting for park, that would still make him a top-5 offensive second basemen. Still, 30 points of wOBA is nothing to sneeze at.
Curtis Granderson was the second most interesting name on the list. Granderson will be entering his age-32 season, and is already coming off a huge offensive decline in 2012 (116 wRC+ in 2012 vs. 146 in 2011). Granderson’s 2011 was easily his best as a professional, so we shouldn’t have been too surprised that he came back down to earth. However, CLIFFORD is not identifying him as a collapse candidate simply because his overall production declined. Mostly, it’s predicting bad things for Granderson because of his 6.6% decline in Z-Contact%. Since 2007, Granderson has never whiffed at more than 86.5% of the pitches he has offered at in the zone. Decline in Z-Contact% has the highest individual impact on collapse risk. Add in big declines in both Spd and UBR, and Granderson could be in for an even worse 2013.
The other player that jumps out is Delmon Young, who was recently signed by the Phillies and appears poised to take the majority of plate appearances in right field. (Still having trouble believing I just wrote that.) Young does have an above average split against left-handed pitchers, but over 608 plate appearances last year he posted a meager .305 wOBA (89 wRC+). Young will only be in his age-27 season in 2013, but he hasn’t posted an above average offensive season since 2010–his only such season with any significant playing time. Additionally, Young’s Z-Contact% dropped by over 6% last year, one of the largest drops last year. For a player already flirting with a .300 wOBA, it’s one more sign that the Phillies would be taking a big risk if they are truly giving Young so many opportunities in 2013.
There’s still more work to be done with CLIFFORD, but the initial results should give us something else to think about when we try to project what these five players are likely to do offensively in 2013.
Bill leads Predictive Modeling and Data Science consulting at Gallup. In his free time, he writes for The Hardball Times, speaks about baseball research and analytics, has consulted for a Major League Baseball team, and has appeared on MLB Network's Clubhouse Confidential as well as several MLB-produced documentaries. He is also the creator of the baseballr package for the R programming language. Along with Jeff Zimmerman, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @BillPetti.