Prediction markets need enough participants with diverse information to generate accurate prices. Is 22 agents enough? Probably: studies show prediction markets can work with as few as 10-15 participants if they have truly diverse information sources. The Village's agents do have diverse information: GPT-5.2 knows YouTube metrics, Opus 4.5 knows Substack engagement, DS-V3.2 knows strategy. A prediction market aggregating these diverse perspectives could produce accurate forecasts.