Accounting for Bias in Human Swarm to Beat Professional NBA Handicappers

The principle known as swarm intelligence (SI) has been founded in biology and allows group decisions to be made through a system of real-time interactions. Biologists have shown SI to be a powerful mechanism to amplify intelligence across many natural species, such as swarms of bees, flocks of birds, and schools of fish. In 2015, a computer-mediated technique was developed, known as artificial swarm intelligence (ASI). This system allows for humans to form real-time systems that can deliberative and generate predictions and forecasts. Real-time Human Collective Intelligence combined with AI algorithms for an ASI technology also known as “Human Swarming” or “Swarm AI.” This method has been shown in recent studies to amplify the intelligence of novice sports fans to expert-level performance. In this current study, ASI technology was combined with statistical methods accounting for bias in human networks to more accurately forecast the outcomes of NBA regular season games spanning over three seasons (2021–2023). Approximately 3.5 games per night were identified as “predictable” using statistical heuristics after adjusting for bias, and the betting performance on these games was measured against the Vegas betting markets. These heuristics were able to achieve an average accuracy of 57.2%, which was significantly better than both novice sports fans (p = 0.003) and the prior baseline swarm intelligence (p = 0.043). These results explore and confirm, when using ASI technology in tandem with statistical techniques to adjust for human bias, a higher level of performance can be realized.