Why Monte Carlo retirement simulations beat the spreadsheet

Most retirement calculators give you one number: "you'll have $1.2M in 30 years." That number is false. The future has a distribution. A static calculator hides the distribution behind a confident-looking dollar figure, and people make irreversible life decisions based on it.

What deterministic models miss

A static "X% return for Y years" model assumes returns happen smoothly. Real markets don't. They cluster: a sequence of bad early years can ruin a portfolio that the same returns in reverse would have saved. This is sequence-of-returns risk, and deterministic models can't see it.

What Monte Carlo adds

Run 10,000 simulations of plausible market paths. For each, model the actual sequence of returns, withdrawals, inflation, and guardrail adjustments. Aggregate the outcomes:

• 5th percentile (worst-case): "If the market punishes me, this is what's left."
• 50th percentile (median): The single most likely outcome.
• 95th percentile (best-case): "If markets are kind, this is the upside."

The gap between p5 and p95 is often 5-10× the median. Two retirees with identical starting positions can end up with $400K vs $4M. That's the variance the deterministic spreadsheet hides.

Why this matters for decisions

Suppose you're trying to decide between retiring this year or working two more. A static model says "you'll have $1.5M either way." Monte Carlo says "85% success rate now, 95% in two years." Now you can decide whether 10 percentage points of success rate is worth two more years of work — that's a real tradeoff, not a false binary.

How k25x implements it

Box-Muller-driven random returns with stock/bond correlation. 10K simulations per run. Guyton-Klinger spending guardrails. Optionally inject a forced first-3-years crash to stress-test sequence-of-returns risk. Results visualize the distribution as a fan chart, not a single line.

It runs entirely on your device. The simulation is insrc/app/(app)/tools/retirement-simulator/actions.ts; nothing leaves your browser.