Backtest portfolios on historical data with Sharpe, Sortino, drawdown, win rate, and Monte Carlo stats.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.