<
>

Predicting breakouts has its challenges

Finding college basketball breakout candidates used to be an art. Fans and media types alike would scour the country and gather info from coaches, and we'd read old recruiting scouting reports, and we'd re-watch players in their formative campaigns, and we'd formulate conclusions -- which were really more like gut feelings -- about little-known players ready to "step up" and "move into the spotlight" and "make the leap" and "lead their teams." These are all fuzzy feelings. Or at least they used to be.

In 2012-13, this process more closely resembles a science. That's thanks to the work of tempo-free statistics pioneers like Dean Oliver and Ken Pomeroy and the folks at Basketball Prospectus, whose per-possession quantification of our sport has allowed us to refocus on the statistics that matter and group those statistics into delineated trends.

In short, we know more, and we're better at predicting, and there are few better examples of this than Luke Winn's Breakout Sophomore Formula work for SI.com. Luke dropped his latest edition of the players likely to make large leaps in production and profile as sophomores, predictions he is able to make based on a gap between players' freshman-year obscurity and actual per-possession usage. This year, Luke identifies a handful of intriguing players: VCU's Treveon Graham, Minnesota's Austin Hollins, Iowa's Aaron White, just to name a few. It's an interesting read and well worth your time.

Or -- bum bum bum -- is it? This year, Luke provides a quick rundown of the predictions his formula got wrong in 2011-12, and the reasons that happened. To wit:

UCLA's super-sized power forward, Josh Smith, never got in shape and thus never broke out, and Ole Miss' Dundrecous Nelson was derailed by the munchies. Seriously: Nelson was apparently as prolific a smoker as he was a shooter, and local police -- who were initially ignored when they knocked on his apartment door after smelling marijuana -- used a Domino's Pizza delivery man as part of a 1 a.m. sting operation in January. They barged in once Nelson opened the door to receive his pizza, found drug paraphernalia, and the resulting drug ticket led to his dismissal from the Rebels.

He joined Nebraska's Christian Standhardinger (a 2010 selection who was projected as a 20-point scorer) in Breakout Formula infamy: After disagreements with coach Doc Sadler, Standhardinger left the Huskers midway through his sophomore season and then was cited for public indecency when a Lincoln police officer found him in a compromising position with a shirtless woman in a local park at 3:40 a.m. I stand by this formula, but feel the need to warn you about its flaws.

Sorry, but if a formula can't predict a player's future Domino's Pizza-related marijuana bust, what good can it really be?

I kid, I kid. If anything, this is a nice summation of why tempo-free statistics are so helpful, and also why they can't reach baseball-like levels of machine-like quantification, at least not anytime soon. For one, baseball is finite and transactional. Basketball doesn't work that way; it's a fluid game that requires tape study as much as statistical underpinnings to truly analyze.

But most importantly, sometimes the most accurate usage rate projections in the world can't prepare you for the news of a late-night Domino's delivery driver allowing police officers to detain a potential star, or a 3:40 a.m. indecency incident with a shirtless lady friend in tow. Ha! Take that, math!