All projects

Monopoly odds

Which Monopoly properties actually pay off?

Why I built it

Rents and prices are fixed by the rules, but how often a player lands on each square is not. We generated that data by simulation and checked it against an exact Markov-chain solution. Group project for the Data Storage course at CentraleSupélec with Augustin Vlandas, Eliott Beghin and Saty Viard Laroque.

What it does

Monte Carlo and Markov chain simulations of the board to find the most profitable properties.

Worth a look

  • Jail is the most visited square: 6.3% of turns, 2.5 times an average square.
  • The orange group is the best investment, and the third house is where returns jump.

Under the hood

  1. Simulate

    Monte Carlo of turns with the official rules (doubles, speeding to jail, both card decks, two jail strategies), seeded for reproducibility.

  2. Prove it

    Exact long-run probabilities from the 40-square Markov transition matrix, used to validate the simulation.

  3. Clean the deeds

    OCR errors in the title deeds fixed (“3O €” becomes “30 €”), with checks that rents increase with houses.

  4. Store and benchmark

    The event log goes to CSV, SQLite and Parquet, with a benchmark of size and query times.

  5. Expected rent and payback

    Expected rent, payback time and ROI per property and colour group, then the charts.