Basketball (Sports) Analytics, Betting & Predictions


When talking and working with numbers the inevitable questions of “so, you can predict game outcomes.” “why don’t you play & win big!” One of my favorites “why don’t you tell me what to play and I’ll win big time.” Always come up. Since starting my journey way back while shadowing NCAA teams and later after having completed several key milestones along they way within the analytics community, I have always been asked the above type questions.

This post is partly a reaction as well as an insight piece on what I try to do alongside my peers which in part aids the betting industry as well as sets us apart and helps staff and the sports much more.

Analytics is not Betting or Predicting the exact outcome!

Let me start out from a point of recent events. Just yesterday a group very well followed in Turkey announced that they had gathered some data(check out the graph below) it indicates the predictions of a mass of people that follow the NBA (the group is called NBA Followers – great pun). While the blue colomns indicates the NBA Finals series outcome the orange colomns indicate the odds gathered by a few betting sites.

NBA Followers vs. betting site odds


Since the bloom and rise of analytics within major sports statisticians, maths gurus, economists have all been raising their game and looking at predictive modeling and translating business analysis models to major sports in the U.S. such as American Football, Hockey, Baseball and very naturally to Basketball.

Several key analytics people that have had success with predictions such as Nate Silver and the crew have also gone on to create basketball predictive models such as the CARMELO project as well as the NBA Team win probabilty calculations. In more recent times organizations such as ESPN, Nylon Calculus; analytics guys such as Danny Leroux, Seth Partnow, Ben Alamar, Kirk Goldsberry have done amazing work that has fueled the predictive modelling of basketball and that has changed the landscape of basketball analytics and a few have even gone on to consult NBA franchises as analytics experts.

Back in 2013 at the Sloan Sports Analytics Conference (SSAC13) the predictive sports betting topic really took to the stage and lead to CantorGaming’s method to be further recognized.

Since then several brands have gone on to make use of how numbers on court translate to numbers on the board and have been in return made use by betting companies.

It should not be forgotten that betting companies usually rely on the fact that outcomes are uncertain and usually have their own in-house method of calculating odds. To an extent analytics helps in determining how the odds are placed by betting companies but overall the correlation between how people play and how odds fulxuate are still an issue. It would be beyond betting if we could see the direct outcome of games even before they were played however that would also take the element of performance put on the court out of its usefulness and it would hinder sports in general. It should not be forgotten that team sports such as basketball are always based upon on court value that later on translates to the aftermath of numbers being crunched and effects betting odds eventually. Very natrually to an extent it can be said that seeking tendencies and certain lineup simulations can lead to extraordinary outcomes and team executives as well as coaches can use this data to determine what the best is for their teams/franchises however, as an analytics guy myself its my firm principal to not mix up the analytics work I do and cross the line into betting using predcitive models and simulate thousands upon thousands of possible options.

Here are a few articles that can be read about predictive modeling and how it relates to betting:

I Wouldn’t Bet On It: Predicting Sports Outcomes with Big Data

MEET THE EXPERT: Dan Conway, Ph.D Director of the Center for Business Analytics, Loras College A number of friends asked me to use my predictive analytics knowledge to predict the outcome of the World Series this week. As I write this, Kansas City has tied the series at three games apiece-who will win Game 7?


An Analytics Platform That Has Mastered Sports Predictions

This story first appeared in the June 2013 issue of Entrepreneur . To receive the magazine, click here to subscribe. The age-old rivalry between jocks and geeks is finally dead–and Nik Bonaddio is the killer. Bonaddio is founder and CEO of numberFire, an analytics platform that takes sports data to new brainy levels.



Here is a video of the SSAC13 Pedictive Sports Betting session:



Betting in the US and Betting in Turkey

Betting has become a systematic way for major sports to generate revenue. Soccer by far has been the most effective arena in which betting has had an impact. In the U.S. Hockey, Baseball, American Football & Basketball all have their own betting platforms and models.

While the majority of betting platforms in the U.S. have carried out their changes and odds are now much more stable and determined by advanced big data calculations as historical data becomes a critical aspect. Turkey and Turkish betting sites, companies still are at the stage of creating odds and over and under limits based on face value rather than advanced big data calculations. While major betting sites in Turkey make use of data and compiled information very successfully within the realm of soccer, basketball betting odds are still not to the level of the U.S. let alone Europe.

It’s a fact that analytics and crunching numbers properly in order to get rid of deviations is still not fully considered. A final note the differentiates betting in the U.S. from that in Turkey is the fact that regulations surrounding betting are extremely strict in the U.S. while there is little to no regulations at all in Turkey that can create boundaries and limits that would allow proper number crunching to determine odds to be verified.

Liked it? Take a second to support Advance Pro Basketball on Patreon!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.