A better way to find assists

Tracking where assists go and what types of baskets they create is becoming more and more important in college basketball analysis. Dave Ryan did a great post the other day about Scott Machado’s assists. Everyone has also seen the great graphics that Luke Winn does in his Power Rankings. This week he even went and checked for erroneous assists (see the North Carolina section). But whenever someone wants to do analysis of assists they have to go over the box score by hand. (At least I think they were.) I’m hoping to eliminate that effort.

While watching Kentucky’s 73-72 victory over North Carolina today, an outstanding contest, I took the time to make a parser in Python that goes through the ESPN play-by-play, pulls out all the assists for a player in two ways: 1) Who they were assisted to. 2) What type of basket it resulted in. It’s a pretty simple piece of parsing code, but it’s also a really valuable tool.

For instance, here’s Kendall Marshall’s assists against Kentucky:

By Player: John Henson (2), Reggie Bullock (2), James Michael McAdoo, P.J. Hairston, Dexter Strickland and Harrison Barnes

By Type: three (4), jumper (2),  layup and dunk

It took me way more time to write this post than it did to find that information, as it takes the program about 10 seconds to do all the calculations. I hope to get a version of the program that I can release to the public by next weekend. As the program moves forward hopefully it can help open up the world of assists to more people and allow for better analysis of those situations. For now if you’ve got a game or player you’d like me to look at drop a note in the comments.

One thought on “A better way to find assists

  1. Pingback: Breaking down the CCSU freshmen guards | Big Apple Buckets

Leave a Reply

Your email address will not be published. Required fields are marked *