Differential Labs Analysis Shows AI Advantage Play Eroding Baccarat Side Bet Revenue Across Asian Casinos

A Differential Labs analysis finds that the spread of baccarat side bets across Asia has created fertile ground for AI-assisted advantage play, with teams now using the technology to identify positive expected value opportunities that erode house edges. The consultancy estimates this activity costs operators at both land-based and online properties between US$500 million and US$700 million annually. Side bets represent only 5% of betting volume yet generate 40% to 60% of baccarat revenue at many venues, according to the report.

Scale and Mechanics of AI-Driven Exploitation

What began primarily in the online channel has shifted to bricks-and-mortar casinos, especially in Asia. Computer-aided counting paired with AI supports “portfolio counting,” where a player tracks multiple side bets simultaneously. A single bet in Macau may yield a positive EV situation one to three times per shoe; monitoring all bets raises that figure to around seven. For the “Super Lucky Seven” bet, flat betting seven units returns 0.27 units per shoe. Allocating those units across bets via portfolio counting lifts the return to 1.04 units, moving the chance of profiting on a shoe from roughly 8% to nearly one in three.

The number of players deploying AI-assisted counting has grown since 2021. Teams operate with floor monitors, bankroll leaders and offsite coordination, while AI-powered apps are sold openly online. Electronic table games offer a soft target because one player can observe several at once. The highest-risk bets include “High Tie-0,” “High Tie-3,” “High Tie-9,” “High Tie-8,” “Dragon Seven” and “Four Card Tie,” which are gaining traction in Manila and Singapore because they count easily and produce consistent profitable shoes. “Monkey/No Monkey” is the simplest to count, improving selection accuracy by 46% even without technology, though volatility limits the edge. By contrast, “Big/Small Tiger” accounts for around 20% of Macau revenue but resists a durable player edge.

Using AI to Counter AI

Differential Labs notes that some bets remain risky even if difficult to count alone, because they integrate into broader schemes that obscure detection. The firm cautions against heavier reliance on deeper cut cards or added friction. Instead, casinos can deploy AI to detect suspicious play proactively and at scale. This threat has intensified alongside smart table technology, which has accelerated side bet innovation through improved chip security and flexible layouts.

The report, published by Inside Asian Gaming, makes clear that operators must treat AI as both the vector of leakage and a core component of surveillance if side bet economics are to remain viable.

Reporting: Inside Asian Gaming

Generated by SCCG’s automated editorial system from published source reporting. SCCG Management holds editorial responsibility.