Monte Carlo Simulation
Project season outcomes by simulating thousands of remaining schedules. Each simulation plays out every unplayed game using team strength estimates, then aggregates results into probability distributions for standings, tournament bids, and championship odds.
Definition
Monte Carlo simulation is a computational technique that runs a model thousands of times with randomized inputs to estimate the probability distribution of outcomes. In sports: simulate the remaining schedule 10,000 times, vary game results based on team quality, and count how often each team finishes in each position.
Inputs
Power rating derived from current record, strength of schedule, run differential (baseball) or point differential (football/basketball)
Every unplayed game with opponent, home/away, and date
Sport-specific home-field win rate adjustment applied per game
Tiebreaker logic, division structure, and tournament seeding criteria by conference
RPI thresholds, at-large bid criteria, and automatic qualifier rules for postseason modeling
Assumptions
- •Team strength is treated as a fixed value that doesn't change across the remaining schedule. In reality, injuries, transfers, and development alter team quality mid-season.
- •Each game outcome is independent. Series momentum, fatigue effects from back-to-back games, and travel impact are not modeled.
- •10,000 simulations is the default run count. This produces stable probability estimates (standard error <1% for most outcomes) without excessive computation time.
Validation
Backtest approach: run the simulation at multiple points during past seasons and compare projected standings to actual final standings.
Failure Modes
Early-season instability — with fewer than 15 games played, team strength estimates are noisy. Projections before mid-March (baseball) should be treated as rough sketches, not forecasts.
Transfer portal churn — college baseball rosters can change materially between fall and opening weekend. Pre-season strength ratings may not reflect actual roster composition.
Conference tournament chaos — single-elimination formats amplify variance. The simulation captures this probabilistically, but individual bracket outcomes are inherently unpredictable.
Schedule incompleteness — if remaining schedule data has gaps (postponements not yet rescheduled), the simulation underestimates remaining games and skews projections.
Version History
Initial methodology documentation. Simulation engine scaffolded. Backtest validation pending 2025 season completion.
Austin Humphrey. (2026, February 17). BSI Monte Carlo Simulation Model. Blaze Sports Intel. https://blazesportsintel.com/models/monte-carlo