Monte Carlo Retirement Simulation: How to Read Probability of Success

Direct answer: What does probability of success mean in retirement planning?

Monte Carlo retirement simulation runs a plan through many market paths and reports success probability, depletion risk, shortfall severity, and the assumptions driving the result.

How Celestice helps

Celestice connects goals, accounts, cash-flow assumptions, tax context, and scenarios in one reviewed workflow. In practice, Celestice helps compare alternatives, preserve assumptions, surface next actions, and route high-impact recommendations through approval instead of leaving the answer in a static planning note.

Why one projection is the wrong tool

Plug your savings, contributions, and an assumed return into a calculator and you get a clean, confident line ending at a single number. The problem is that the single number almost certainly will not happen. Real returns are not a smooth average — they arrive in a random sequence of good and bad years, and that sequence matters enormously, especially near retirement. A projection that assumes 7% every year tells you what happens in a world that does not exist.

Monte Carlo simulation fixes this by running your plan through hundreds or thousands of randomized return paths and reporting the distribution of outcomes. Instead of "you will have $X," it answers the more useful question: "across many possible futures, how often does this plan succeed, and how badly does it fail when it fails?"

The metrics that actually matter

A good simulation produces a structured packet of evidence, not just a single score. The ones worth understanding:

• Target-hit probability. The share of simulated paths in which you meet your goal. This is the headline "probability of success." A plan at 85% succeeds in most futures but not all.
• Depletion probability. The share of paths where the money runs out before the horizon ends. For retirement, this is often more important than the upside.
• Percentile bands. The range of ending outcomes — the pessimistic (e.g. 10th percentile), median, and optimistic (90th percentile) paths. The width of the band tells you how uncertain the plan is.
• Average target shortfall and expected surplus. When the plan misses, by how much on average? When it succeeds, how much cushion is there? These quantify the severity on both sides, not just the odds.
• Convergence and warnings. Whether enough paths were run for the numbers to be stable, and any flags about the assumptions. A result that has not converged is not yet trustworthy.

Probability is not certainty — and that's the point

A common mistake is treating an 85% success probability as either "basically guaranteed" or "a coin-flip." It is neither. It means that in roughly one in seven simulated futures, the plan as currently designed does not reach the goal. Whether that is acceptable depends on the consequence of the 15% — running short in retirement is far more painful than missing a stretch savings target. The probability is an input to judgment, not a substitute for it.

Drivers: knowing why, not just what

The most actionable part of a simulation is the driver analysis — which assumptions move the outcome most. If success probability is most sensitive to your savings rate, the lever is contributions. If it is dominated by your withdrawal rate or claiming age, the lever is the spending plan. If it is driven by return assumptions, the honest move is to test more conservative ones. Knowing the driver turns a scary or reassuring number into a specific next step.

Comparing scenarios is where decisions get made

Simulation does not decide the action by itself; it supplies a structured uncertainty packet that other planning workflows can compare and cite. The real value emerges when you run two scenarios side by side: claim Social Security at 67 versus 70, save more versus retire later, a 60/40 portfolio versus 70/30. Each produces its own probability, depletion risk, and bands, and the comparison makes the trade-off concrete. The question shifts from "is my plan good?" to "which version of my plan is better, and why?"

Where it fits in a continuous plan

Because assumptions, balances, and law tables drift, a simulation is a snapshot that goes stale. In a continuous, agentic platform, simulation becomes the probabilistic evidence layer feeding the rest of planning: a goal, a retirement plan, or a what-if scenario is re-tested when inputs change, the result packet is attached as evidence, and any consequential decision it informs is routed through review. The number is never the decision — it is the support for one.

The takeaway

Monte Carlo simulation trades a falsely precise single number for an honest range of outcomes. Read it by looking past the headline probability to depletion risk, the width of the percentile bands, the severity of shortfalls, and — above all — the drivers you can actually act on. Used that way, it is the most clear-eyed tool available for planning under uncertainty, which is the only kind of planning there is.