Chance of success

The gauge on the Simulations tab is the headline number, so it is worth being precise about what it counts.

The definition

It is the share of 500 simulated futures in which your plan never ran out of money.

Each of those 500 runs is a complete life of your plan against a different possible market: different returns, different inflation, different order of good and bad years. A run counts as a success only if every single year of it was paid for. If the caption under the gauge says 437 of 500, then in 63 of those futures the money did not last.

Why “never ran out” and not “ended with money left”

These come to the same thing here, because in this model running out is permanent.

If a year cannot be paid for, the plan stops there. Later years stay at zero, including income that would have arrived afterward. A household that runs dry at 68 does not quietly recover when Social Security starts at 70.

That rule exists because the alternative scores plans dishonestly. Judged only on the final balance, a plan could fail to cover its spending for three straight years, pick up a pension, and still finish with a healthy number, reporting a clean 100%. Any plan with a late income event could hide years of shortfall behind its ending balance.

The three zones

Reading	Label	What it suggests
Below 70%	High risk of depletion	More futures than not are uncomfortable. Something structural is worth revisiting.
70% to 90%	Moderate chance of lasting	The plan works in most futures and fails in a meaningful minority.
Above 90%	Strong chance of lasting	The gauge turns green. The plan holds up in all but the worst outcomes.

Why 90% and not 100%

A plan that survives all 500 futures, including the very worst sequence in nearly a century of market history, is usually a plan that is spending considerably less than it could afford to. Insisting on 100% is not free. It is paid for in years of working longer or spending less, to insure against a scenario more extreme than the Depression.

90% is where the gauge turns green. It is a way to read the number, not a target we are telling you to hit. What the right number is for you depends on how much flexibility you have if things go badly, and that is a conversation for you and, if you have one, a professional.

What the percentage does not claim

This is the part most worth internalizing.

It is not the probability that your plan will work. It is the probability that your plan works in this model, given your inputs, if the future resembles the past. Three different things there, each of which could be wrong.

It is not a statement about anything outside the historical record. The simulations are built from market history from 1928 through 2025. The future is free to land outside that range, and occasionally does.

It is not comparable to another tool’s success rate. Every planner defines success differently and simulates markets differently. A 91% here and an 84% somewhere else may describe identical plans.

It moves. Change your spending by a few hundred dollars a month and watch it swing. That sensitivity is real information: it tells you how much slack the plan has. A plan at 92% that drops to 71% when spending rises 10% is a different animal from one that only drops to 88%.

The fan chart underneath

Below the gauge is a chart with a median line and two shaded bands: the 25th to 75th percentile, and the 10th to 90th.

Those bands are a cross-section, not a path. For each year separately, all 500 simulations are sorted and the middle and edges are plotted. The 10th percentile in year 5 and the 10th percentile in year 30 usually come from completely different simulations. No single future traces the edge of a band, and no real market produces that curve.

This is exactly why the Projections tab shows one real simulated path instead: the year-by-year detail there describes a future that actually happened in the model, with its taxes and withdrawals all consistent with each other. The two tabs answer different questions. Projections asks what one plausible future looks like in detail. Simulations asks how often things work out.

If the number is lower than you hoped

The model responds to a handful of inputs far more than the rest: how much you spend, when you stop working, when you claim Social Security, and how much of the portfolio is in stocks. The most useful thing you can do is clone the plan and change exactly one of them, then compare. Because every plan is simulated against the same 500 market futures, the difference you see is your edit and not a different roll of the dice.