Quantifying luck: Determining the probability of baccarat wins given player bet mix

Session Title

Gambling Mathematics: Quantitative Finance

Presentation Type

Paper Presentation

Start Date

26-5-2026 12:00 AM

Abstract

Recent advances in smart table technology now allow table game patron wagering behavior to be tracked at the individual bet level, enabling analyses of player behavior that were previously infeasible. This study proposes a framework that treats a player’s evolving bet mix as a sequence of heterogeneous probabilistic trials, where each wager carries a distinct win probability determined by the underlying game and bet type. Using a Poisson binomial formulation, the distribution of wins and losses over a sequence of bets can be characterized conditional on the player’s specific wagering pattern. The proposed approach aims to assess whether observed outcomes are statistically unusual relative to expected variance, offering a risk-normalized alternative to loss-based thresholds. Potential applications span integrity monitoring and customer analytics, particularly in contexts where monetary loss measures are confounded by promotional credits or non-cash wagering instruments.

Author Bios

Dr. Mana Azizsoltani earned a bachelor’s degree in mathematics from the University of Nevada, Las Vegas and a master’s degree in statistics from North Carolina State University. He completed a PhD in Hospitality Administration at UNLV, applying quantitative methods to hospitality and gaming. He is currently Head of Data Science at Differential Labs, where he is focused on applications of data science for casino marketing, surveillance, table game optimization, operations, responsible gaming, and revenue management.

Share

COinS
 
May 26th, 12:00 AM

Quantifying luck: Determining the probability of baccarat wins given player bet mix

Recent advances in smart table technology now allow table game patron wagering behavior to be tracked at the individual bet level, enabling analyses of player behavior that were previously infeasible. This study proposes a framework that treats a player’s evolving bet mix as a sequence of heterogeneous probabilistic trials, where each wager carries a distinct win probability determined by the underlying game and bet type. Using a Poisson binomial formulation, the distribution of wins and losses over a sequence of bets can be characterized conditional on the player’s specific wagering pattern. The proposed approach aims to assess whether observed outcomes are statistically unusual relative to expected variance, offering a risk-normalized alternative to loss-based thresholds. Potential applications span integrity monitoring and customer analytics, particularly in contexts where monetary loss measures are confounded by promotional credits or non-cash wagering instruments.