Using predictive analytics for fantasy football

You’d think that a master of predictive analytics would steer clear of emotional entanglements with the lowly Detroit Lions. But hometown sympathies resist the power of reason and statistics. As a Philadelphian, this is something I know.

You’d think that a master of predictive analytics would steer clear of emotional entanglements with the lowly Detroit Lions. But hometown sympathies resist the power of reason and statistics. As a Philadelphian, this is something I know.

In any case, an IBM analyst named Hetal Thaker reports great success using predictive analytics for her fantasy football team. For 11 years, she chose her roster the old-fashioned way, by the gut, and won her league only once. For the last five years, she’s been building predictive models and using them to pick her players. The results: Three wins in five years.

She says she has developed new metrics over the past years. One, for example, is called the …quot;team factor….quot; She explains:

…quot;[It] takes into account not only the player, but his supporting cast. This is very important because you can haave the best wide receiver in the league, say Larry Fitzgerald of Arizona, but if he doesn’t have a good quarterback passing to him, he’s unlikely to have the fantasy value you would anticipate….quot;
 
As a gut-driven fan, I have to quibble here. It seems to me that some of the statistical superstars often play for middling teams, because the great teams have a more balanced attack. If you think of the greatest running back, by stats, from OJ Simpson to Detroit’s own Barry Sanders, they were often one-man shows on teams going nowhere. Walter Payton’s Bears finally won a Superbowl, but long after his prime.

Still, she’s right about quarterbacks. With a good one, receivers are lost. That said, this discussion inspired me to look at Barry Sanders highlights on YouTube. It’s almost enough to make me cheer for the Lions.