You might have noticed we rely a lot on quantitative modeling for our betting picks so we’re big fans of using math to predict sports outcomes.
NBA is interesting because you have a lot of data to work with and many different stats to look that simplifies the process of building a statistical model. In short, you can build very reliable predictive models only looking at numbers but would I rely solely on the model to make winning betting picks? Actually, you can. In fact, our model was profitable on its own without adding the human element, meaning we relied completely on the picks the model spit out ignoring our own biases and judgment.
Is it optimal? Not really. Sometimes it’s hard to quantify factors and that’s where statistical models fail.
Basically, you need to combine math with your “human” eyes in order to make the best predictions possible.
It’s too much to cover in one article how to build a statistical model for NBA without having a background in statistics so my suggestion to someone who is interested in taking on a mathematical approach to sports betting, in general, is to learn python or R(both powerful programming languages for statistics) and maybe take at least a basic course in statistics.
You could self-learn everything these days but my formal stat classes in university definitely helped me get through the basics.