← @bgSponge

Kelly Criterion Simulator

Summary

Kim Birch's Kelly Criterion Simulator uses Monte Carlo methods to model bankroll evolution as a multiplicative process, showing how fixed edge assumptions drive long-run geometric growth and volatility.

  1. Kelly maximizes asymptotic growth.

    The formula targets maximum long-run geometric growth by maximizing expected logarithmic utility, though this objective does not guarantee smooth or low-stress capital paths.

    Growth optimality does not mean smooth growth, safe growth, or low-stress growth.

  2. Full Kelly increases instability.

    Applying the complete mathematical fraction often results in higher expected growth alongside deeper interim drawdowns and wider outcome dispersion.

    Higher expected growth under full Kelly usually means greater volatility, deeper interim drawdowns

  3. Fractional Kelly improves robustness.

    Reducing the stake to half or quarter-Kelly sacrifices some long-run growth to lower drawdown pressure and reduce sensitivity to noisy probability estimates.

    It reduces expected long-run growth relative to full Kelly, but it also reduces volatility

  4. Probability errors cause overbetting.

    The formula is highly sensitive to input quality; optimistic edge estimates lead to excessive stake sizes that turn sound logic into fragile processes.

    If the edge estimate is too optimistic, the Kelly fraction can become too large.

  5. Simulations reveal structural dispersion.

    Monte Carlo paths demonstrate that identical assumptions produce divergent trajectories, proving that drawdown and volatility are inherent features of the model, not noise.

    Some paths may grow quickly, others may stagnate, and others may suffer deep drawdowns

Highlights

  • In that sense, the real challenge is not understanding the Kelly Criterion. The real challenge is knowing whether the assumed edge deserves to be trusted

  • If the edge estimate is too optimistic, the Kelly fraction can become too large. That leads to overbetting, which is one of the fastest ways

  • Higher expected growth under full Kelly usually means greater volatility, deeper interim drawdowns, and wider outcome dispersion.

  • Some paths may grow quickly, others may stagnate, and others may suffer deep drawdowns before recovering or failing to recover at all.

  • This simulator excludes estimation error, market adaptation, transaction costs, execution friction, liquidity constraints, and structural breaks.

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →