The Future of Statistics: Player Tracking Data in the NBA

One of the most promi­nent experts in bas­ket­ball ana­lyt­ics shares the pow­er of play­er track­ing data to extract insights for mea­sur­ing ath­letes, for­mu­lat­ing game strat­e­gy, and eval­u­at­ing team per­for­mance at an elite level.

Tracking data is a pow­er­ful ana­lyt­i­cal tool in the world of bas­ket­ball, and Luke Bornn is one of the fore­most fig­ures in this inno­v­a­tive area of sports data.

As Vice President of Strategy and Analytics for the NBA’s Sacramento Kings, Bornn’s role is to essen­tial­ly help the Kings make more objec­tive deci­sions through advanced track­ing data and sta­tis­ti­cal modelling.

Until very recent­ly, Basketball coach­es were dri­ven by box score data, which is sim­ply a struc­tured sum­ma­ry of the results of the game. In com­par­i­son, track­ing data can mea­sure things that are hap­pen­ing off the ball, while also mea­sur­ing spac­ing, move­ment, and team formations.

Tracking data real­ly catch­es things with­in the sport that coach­es talk about and the way coach­es think,” said Bornn.

Coaches don’t always talk about the out­comes that are mea­sured at that lev­el of data, but they do talk about move­ment and spac­ing and for­ma­tions. So track­ing data real­ly allows us to mea­sure fea­tures of the game which are much more close­ly aligned to the way coach­es think.”

Tracking data real­ly catch­es things with­in the sport that coach­es talk about and the way coach­es think”

In the past teams would only mea­sure data for in-game per­for­mance, where­as with track­ing tech­nolo­gies NBA fran­chis­es can obtain far more data points.

Bornn explains how mea­sur­ing train­ing per­for­mance data leads to much more effec­tive insights and a high­er accu­ra­cy and under­stand­ing of shoot­ing skills.

If you think about the data that you’re get­ting with­in a bas­ket­ball match, you’re get­ting maybe around 200 shots in total, and if you’re think­ing of look­ing at mea­sur­ing skill in a par­tic­u­lar area of the court, you need a lot of more shots than that to know if a play­er is good or bad,” said Bornn.

A typ­i­cal play­er shoot­ing a cor­ner three point shot, you might only see one or two shots from there in a game, so you would need to see sev­er­al games to know if a play­er is good or bad at shoot­ing cor­ner threes. However, if you look with­in prac­tice, many of these guys are tak­ing thou­sands and thou­sands of shots every day, so using track­ing data from train­ing allows you to under­stand a play­ers skill lev­el much more rapid­ly than if you just look at a match.”

There are a lot of clubs who are spend­ing in this area, whether its pur­chas­ing data or hir­ing ana­lyt­i­cal tal­ent, that don’t have the cul­ture or the process in place to make deci­sions from this data,” said Bornn.

You could have a sig­nif­i­cant invest­ment that is not impact­ing your club which is not ideal.”

Secondly, is hav­ing ana­lyt­i­cal tal­ent in your staff to be able to han­dle com­plex data.

Five or ten years ago clubs would hire peo­ple with very basic tech­ni­cal skills, the abil­i­ty to manip­u­late excel spread­sheets for exam­ple,” said Bornn.

But in the past few years, espe­cial­ly with the advent of track­ing data, this new data has such a vol­ume and com­plex­i­ty that it requires a whole dif­fer­ent tech­ni­cal skill set. It requires the abil­i­ty to han­dle mas­sive data sets and also spa­tial tem­po­ral modelling.

The tal­ent that is need­ed don’t just have sim­ple analy­sis back­grounds, but full machine learn­ing and sta­tis­tics back­grounds so that they are able to mod­el these big com­plex systems.”

To learn more about how Hudl uses ana­lyt­ics to fuel the mod­ern game, take a look at how we devel­oped the future of track­ing sys­tems with Danish Superliga team Brondby IF, and how that feeds into Hudl Pro Suite.