Almanac

Prediction markets rely on high-quality trading signals, but there is no standardized way to identify, rank, or reward the traders generating the most accurate predictions. Traditional centralized systems cannot verify signal quality across independent traders, and traders have no transparent path to monetize their edge or compete fairly against others.

A decentralized network of traders competing under a shared incentive mechanism solves this by creating transparency and fairness. Each trade becomes a scored prediction evaluated against real-world outcomes. Traders who consistently generate profitable signals earn rewards proportional to their accuracy, ROI, and trading volume—eliminating gatekeepers and allowing any competent trader to compete.

Sportstensor implements this as a subnet on Bittensor where miners trade on Almanac (a Polymarket interface) and validators score their performance hourly using a two-phase optimization system. The mechanism evaluates ROI, qualified trading volume, and historical consistency over a rolling 30-day window, then allocates Alpha Token rewards fairly and transparently. Validators sync on-chain metadata, ingest trading history from Almanac's backend, compute scores using the dual-pool system, and publish weights to determine next-epoch emissions.

Sportstensor is a decentralized competition network for sports prediction where miners generate trading signals by trading on Almanac (a Polymarket interface), and validators score their performance using a two-phase optimization system that rewards miners based on ROI, trading volume, and historical consistency. Miners earn Alpha Token emissions proportional to their profitability and signal quality, while validators sync metadata, ingest trading data, and set on-chain weights to determine rewards.