I’m a long-time Bitcoiner who kept running into the same question:
How much Bitcoin can I hold if I actually plan to live off my portfolio someday?
Most retirement simulators either ignore Bitcoin entirely or behave badly when you introduce high-volatility assets. I wanted a tool focused on sequence-of-returns risk, withdrawal sustainability, and how Bitcoin changes the distribution of outcomes — not price prediction.
So I built HODLPath, a Monte Carlo retirement simulator designed to stress-test portfolios that include Bitcoin.
URL:
https://hodlpath.streamlit.app/Questions and constructive feedback are welcome.
What HODLPath Does
- Simulates thousands of portfolio paths over multi-decade horizons
- Models correlated multi-asset returns
- Explicitly focuses on withdrawal risk, not terminal wealth
- Supports:
- start-of-year vs end-of-year withdrawals
- inflation-indexed or nominal spending
- optional spending guardrails after drawdowns
- annual rebalance or natural drift
Outputs include:
- success / ruin probabilities
- full wealth distributions (fan charts)
- spending path distributions
- downloadable PDF reports (including comparison exports)
Return Model (Important)
HODLPath uses a lognormal (log-return) model:
- Users enter expected return and volatility as annual arithmetic values (the intuitive inputs most people think in).
- Internally, these are converted into log-return parameters and sampled with a correlation matrix.
- This ensures:
- portfolio values never go below zero
- realistic long-horizon compounding
- more appropriate behavior for volatile assets like Bitcoin
This is not a forecasting tool — it’s meant for decision clarity under uncertainty and sensitivity analysis.
Garbage in → garbage out still applies, and assumptions are intentionally explicit.
Who This Is For
HODLPath is aimed at:
- long-horizon investors
- early retirees / FIRE planners
- anyone modeling withdrawals with Bitcoin in the portfolio
It’s not investment advice — just a way to explore how fragile or robust different allocations are when you actually start spending.
Side note: I built this as a learning exercise using modern AI coding assistants, which turned out to be surprisingly effective for this kind of analytical tool — but all modeling choices and assumptions are explicit and user-controlled.
