Sports Analytics

We model the game with the rigor markets get. Player tracking is fed through Bayesian state-space models to extract latent skill from noisy outcomes; lineups are valued on a learned positional-adjustment, marginal-revenue-product basis; transfers are priced as discounted-cash-flow problems with stochastic injury and form curves. We run match Monte-Carlo at a hundred-thousand trials per decision, surface dominated strategies, and quantify the price of the alternatives. The deliverable is plain: this is the call, this is its expected value, this is the standard error, and this is exactly what would beat it.