ADVERTISEMENT

Semi-OT Using Analytics in Recruiting

Queztastic

Junior
Nov 1, 2013
831
499
63
Interesting read on a company that uses analytics to identify what recruits a school has a chance with. Working with Northwestern, the company claims that their algorithms have predicated which recruits would not sign with Northwestern with 94% accuracy and also predicted which recruits would sign with the Wildcats with 80% accuracy. They claim its basically a way for teams to identify which recruits to concentrate their time on. However, isn't it a self-fulfilling prophecy? Once the company tells you not to waste your time on a recruit, wouldn't you then not recruit as hard, making it less of chance that the recruit picks you?http://www.usatoday.com/story/sport...approach-inexact-science-recruiting/96963326/
 
Good Post. Good model should contain some concept of time spent recruiting as a control. Lets think about how a model for RU would look? I'd expect state (NJ) and public HS to have positive relationships with our probability of getting the kid. I'd expect private school and power 5 offers to have negative relationship with our probability of getting the kid. We need a model that can suss out "kid wants big time".
 
Good Post. Good model should contain some concept of time spent recruiting as a control. Lets think about how a model for RU would look? I'd expect state (NJ) and public HS to have positive relationships with our probability of getting the kid. I'd expect private school and power 5 offers to have negative relationship with our probability of getting the kid. We need a model that can suss out "kid wants big time".


But you can't lock into assumptions that can be changed. After all, RU got Nova. Hamilton and Carroo (along with Huggins and Lumpkin) in a 3 year window.
 
I'm more interested in utilizing analytics to identify under the radar recruits. Doesn't mean squat if that 80% of kids suck and do not become contributors.
 
ADVERTISEMENT
ADVERTISEMENT