How the simulation works
Each of the 1,000 runs draws a random return for every year from a normal distribution with your average and volatility. The withdrawal comes out at the start of each year and rises with inflation; whatever is left grows or shrinks by that year's return. A run fails if the balance hits zero before the final year. The shaded band shows where the middle 80% of runs sit each year, in today's dollars.
Reading the result
- Above 90%: the plan survives almost every market path. You may have room to spend more.
- 75–90%: a common target, as long as you'd trim spending after a bad run.
- Below 75%: lower the withdrawal, work a little longer, or plan guaranteed income (delay Social Security or an annuity).
Limits of the model
Real returns have fatter tails than a normal distribution and tend to cluster, and inflation isn't constant. The model also ignores taxes and fees; subtract your fund costs from the average return. Treat the success rate as a comparison tool, not a promise.
Questions people ask
What is a Monte Carlo retirement calculator?
Instead of assuming the same return every year, it runs your plan hundreds or thousands of times with random yearly returns drawn around an average and a volatility you choose. The share of runs where the money lasts is your probability of success.
What success rate should I aim for?
Many planners target 75% to 90%. 100% usually means you are underspending. A lower rate is acceptable if you are willing to cut spending in bad markets, because the simulation assumes you never adjust.
What volatility should I use?
US stocks have had an annual standard deviation of roughly 15% to 20%; a 60/40 stock/bond portfolio around 10% to 12%. Higher volatility with the same average return lowers the success rate.
Why do results differ from the simple withdrawal calculator?
A fixed-return calculator ignores the order of returns. Bad years early in retirement do far more damage than the same bad years later, so a plan that works at a steady 5% can fail in a meaningful share of random runs.
Is the simulation random each time?
It uses a fixed seed so the same inputs always give the same answer, which makes it easier to compare changes. Each run still draws a fresh random sequence of returns.