Fat-tail-aware return distribution with a left-tail skewness factor, bracket-based tax modeling per country across 26 supported countries, and a probability of success — not a single deterministic line.
Run my Monte CarloThree modeling choices set Monte Carlo retirement simulators apart from each other. Here is how Retirement Lab handles each:
Lognormal returns without a skewness adjustment. A clean lognormal draw produces too few of the steep, fast drawdowns that real markets deliver. Retirement Lab applies a left-tail skewness factor to its return draws, stretching the worst shocks further into the negative so simulated drawdowns reflect fat-tail risk rather than a textbook bell curve.
Flat-rate or no tax modeling. Subtracting one effective tax rate from gross withdrawals hides bracket effects, capital-gains treatment, and country-specific wealth taxes. Retirement Lab applies bracket-based income tax, capital-gains tax, and (where applicable) wealth tax inside each simulated year, for the residency country you choose, so the after-tax draw reflects the bracket you would actually retire into.
Hiding sequence-of-returns risk behind percentile aggregates. A summary “85% success” number can come from very different distributions of unlucky early-year sequences. Retirement Lab plots percentile bands (10/25/50/75/90) across the full horizon — computed from all 10,000 paths — alongside a sample of individual simulated timelines, so the impact of a bad first decade shows up as a widening fan in the early years instead of being averaged into one number.
The methodology page documents the return-distribution and tax-modeling parameters in full.
Each scenario runs 10,000 independent simulations. Every path draws its own annual market returns from the configured distribution, so the result is a probability distribution of outcomes rather than a single projection. The path count is fixed at 10,000 on the free tier; the Pro tier exposes a higher-resolution mode for users who want tighter percentile estimates.
Each year’s portfolio return is drawn around your configured mean and volatility from a normal distribution, then a left-tail skewness factor stretches draws below a fixed cutoff further into the negative. The result is a return distribution with materially fatter losses than a clean bell curve, calibrated to long-run global equity behavior. The mean and volatility are scenario inputs you set before each run; the skewness parameters are baked into the simulator with defaults the methodology page documents alongside the rationale.
Taxes are computed inside each simulated year, not as a flat haircut on the final balance. For the residency country you choose, the simulator applies bracket-based income tax to taxable withdrawals, capital-gains tax to realized gains, and wealth tax where the country has one. Dual taxation for U.S. citizens abroad is modeled with a simplified foreign-tax-credit overlay at the income-category level. That means the after-tax cash you can actually spend in a given year reflects the bracket the simulated portfolio puts you in that year — a high-withdrawal year and a low-withdrawal year are taxed differently, the way they would be in real life. The methodology page lists the country-by-country defaults and the explicit assumptions for regimes like NHR, IFICI, non-dom, and PFU.
Probability of success is the share of simulated paths in which the portfolio is not depleted before the end of the planning horizon, given your inputs. It is a simulation output, not a recommendation — an “85% success” result means 85 of every 100 simulated futures finished with a non-zero balance under the assumptions you provided. A different inflation, return, tax, or spending input will produce a different number; the metric is only as honest as the inputs it summarizes.