SI Vault
 
A GAME OF RISK
WILL CARROLL
August 04, 2011
A computer science nerd has an innovative way to help you assess risk and build your team, Wall Street--style
Decrease font Decrease font
Enlarge font Enlarge font
August 04, 2011

A Game Of Risk

A computer science nerd has an innovative way to help you assess risk and build your team, Wall Street--style

1

TOP-TIER RB Adrian PETERSON VIKINGS

Few—if any—choices are safer than Minnesota's talented, workhorse runner

2

SAFE RB Steven JACKSON RAMS

Consistent year-to-year performance makes Jackson a worthy RB2

3

SAFE WR Greg JENNINGS PACKERS

A prolific offense means guaranteed production from the top Pack receiver

4

RISKY RB/WR Marques COLSTON SAINTS

Off-season knee surgery and a QB who spreads the ball raise a red flag

5

SAFE QB/RB/WR/TE Antonio GATES CHARGERS

Foot issues aside, Gates is as much a sure thing as you'll get in round 5

YOU'RE PLOTTING YOUR DRAFT STRATEGY. You've got your list of must-haves, sleepers, personal favorites and potential busts. What you might not have, however, is some truly innovative analysis around player fantasy value—analysis built around the kind of number crunching usually found on Wall Street. Until now.

Nik Bonaddio is a 28-year-old honors graduate of Carnegie Mellon, where he was a self-described "computer science nerd." After graduating with a master's degree in information systems management in 2005, Bonaddio worked for Yahoo! and a series of start-ups doing creative and visual directing. Then in 2010 he founded numberFire, a sports-analytics framework that predicts and delivers advice on player performance.

Analyzing millions of rows of football data, numberFire projects a player's performance based on a "similarity-scoring" algorithm, which attempts to find similar players, games and situations to the ones he will be facing in the upcoming year. Once numberFire gets an idea of how the player will score, it can determine the difference in "expected wins" based against an average fantasy football team using league-average fantasy football players.

Through his number crunching, Bonaddio has found that one of the biggest misconceptions about fantasy football is that success is predicated entirely on maximizing your returns from individual players. In the abstract sense, that is true; you need to outscore your opponent. But as Bonaddio has found, winning in fantasy football is sometimes less about starting the player who will break out and score 20 points and more about not starting the player who will put up zero. On any given week, there are more players who are crashing out than breaking out, so more focus needs to be paid toward minimizing the risk in your lineup.

Risk is defined by numberFire in three ways: injury, consistency of performance on a week-by-week basis and yearlong performance in relation to draft position. Injuries, of course, are very hard to quantify, but the latter two are not, so numberFire's analysis leans more heavily on them.

From a draft perspective, the best way to do this is to carefully calibrate the risk profiles of your team. Some players are inherently less risky than others, which is a factor of, among other things, the teams they play for and their injury history. Falcons receiver Roddy White is as consistent as receivers come week-to-week, whereas Eagles wideout DeSean Jackson may be explosive on some Sundays but just won't show up for you on others.

The fantasy teams with the best chance of success are the ones with the right mix of risky and safe choices. Bonaddio says to "think of your team like a stock portfolio." You need to balance high-risk investments with tried and true options. Imagine starting an entire team of Matt Forte--esque players; you could score 15 or 150 points. However, a portfolio that is too safe may leave you with a squad that loses regularly to teams with a single breakout player.

On the next two pages we put some names around this concept, assuming a 10-team league.

numberFire's MOST RISKY

Continue Story
1 2 3 4 5 6