Tool · from historical returns
Retirement withdrawal calculator: would this rate have survived, historically?
The "4% rule" and its cousins usually get tested against an assumed return distribution. This resamples real monthly returns for the exact funds you hold — using contiguous historical blocks, not single shuffled months, so real multi-year drawdowns and recoveries stay intact — and reports how often a fixed withdrawal would have run the portfolio to zero.
How to read this
Enter your fund mix, a withdrawal rate, and a horizon. Each of several thousand simulated trials stitches together random ~2-year blocks of the funds' actual calendar history until the horizon is filled, applying the same historical months to every fund at once so real cross-fund correlation (everything falling together in 2008 or 2020) survives. Success rate is the share of trials where the balance never hit zero. This is a historical-scenario statistic, not a forecast, guarantee, or personalised recommendation.
Your plan
Sources and limits
Monthly total return (dividends and splits included) from Yahoo Finance, a different data source and provenance from the SEC N-PORT holdings used everywhere else on this site. History depth varies a lot by fund — a multi-fund mix can only be bootstrapped from the window where every selected fund has data at once, which the result discloses; it is never silently padded with a longer-lived fund's data past where a shorter-lived one begins.
Nominal, not real: there is no inflation series in this project. The withdrawal amount is fixed in dollar terms for the whole horizon and is not adjusted upward for inflation each year — over 20–30 years its real purchasing power will erode even in scenarios marked a "success." Method is block bootstrap with non-overlapping-history, ~24-month blocks (long enough to carry a real drawdown-and-recovery pair), never wrapped past the end of the data. Free to use and cite with attribution to this page. Found an error? Say so on the contact page.
Built with AI assistance (Claude); the figures are not AI-generated — every trial resamples real historical returns, nothing here is invented. See how this site is built.