ELIO LIBERATORE

I build Monte Carlo simulators that compute outcomes instead of scraping them off a web page. Public feeds go in, the model replays what is left thousands of times, and fair probabilities come out as JSON, CSV or an API call.

What I maintain

• Soccer Predictions API - title, top-four, playoff and relegation odds across 16 leagues, Premier League to Liga MX and MLS.
• Soccer Dixon-Coles Match Predictor - one probability grid per fixture: 1X2, over/under 2.5, both teams to score and the likeliest scorelines, fitted from each team's goal history.
• MLB Playoff Odds API - division, wild card, postseason and top-seed probabilities for all 30 teams, with projected records, market edge and quarter-Kelly stake sizing.
• NFL Playoff Odds API - the same for all 32 teams, No. 1 seed included.
• NHL Playoff Odds API - playoff, division and Presidents' Trophy probabilities for all 32 teams.
• NBA Playoff Odds API - playoff, play-in, division, conference and championship probabilities for all 30 teams.
• Sports Probabilities MCP - the models above as tools, so Claude, ChatGPT or Cursor return a computed probability, not an invented one.

How the models work

Season models blend current form with a multi-season baseline, regressed towards average by how much season is left, and redraw each team's true talent every simulated season. The match model fits attack and defence ratings per team with a Dixon-Coles bivariate Poisson, weighted towards recent games.

What people use them for

Comparing fair probabilities against market prices. Feeding a dashboard on a schedule. Backtesting a past season before trusting the model on the current one. Giving an AI agent a tool that computes instead of guessing.

No Actor needs an API key of your own. Every task on the Tasks tab is a ready-made configuration you can run in one click. Bugs and feature requests welcome in the Issues tab of any Actor.