Flipping through the factor playbook to find systematic edges in sports betting.

Ballin with Beta is written by a quant-nerd for quant-nerds. I’m assuming that you, the reader, are familiar with sports betting and factor investing terminology, specifically long/short factor portfolio construction via cross-sectional sorts. Maybe you have a sports betting strategy that could benefit from a new perspective. Or maybe you have a multifactor investment strategy that could benefit from an expanded opportunity set.

This blog is a collection of studies that attempt to replicate well-known constituents of the academic finance factor-zoo in sports betting markets, primarily NBA spread markets. By doing so, I hope to:

1. Build a framework for sports bettors to dynamically over- or under-weight inputs used in their unique decision-making processes.
2. Contribute to the replication study literature by either supporting or challenging the robustness of academic risk factors to data mining.

Traditional approaches to sports betting are matchup-centric – they attempt to independently predict the outcome of a game and compare their prediction to the market’s prediction. I am more interested in a different question: which box-score stats are being mispriced by the market in their importance to the outcome of all of today’s games?

Each blog post will focus on replicating a different anomaly documented in the asset pricing literature, by making use of long/short factor portfolios popularized by Fama and French’s original size and value anomalies. Just like Fama and French ranked stocks cross-sectionally, I will create long/short portfolios from common box-score stats that are known drivers of a team’s performance, for example: offensive and defensive field goal percentage, turnover rate, rebound rate, free throw rate, and assist rate.

But what does sorting box-score stats into long/short portfolios even mean? It is a way to isolate the market price of a box-score stat on a given day, regardless of which teams are playing, enabling us to quantitatively express the guiding question that I posed above. To translate each factor from the quantitative finance world to the sports betting world, I will provide a step-by-step description of how I sort these box-score stats into long/short factor replicating portfolios.

If I haven’t scared you off yet, I want to thank you for going on this factor exploration journey with me. Head over to the Factor Playbook to see a list of quant finance factors that I’ve attempted to translate to the sports betting world. Let’s do to sports betting what quants did to stock-picking – and re-attribute the handicapper’s edge from mystical alpha into alternative betas.