How to Explain Monte Carlo Probability of Success to Clients
A retirement plan that ends with "you have an 87% probability of success" is only useful if the client understands what the 87% means — and what it does not. For advisors, the hardest part of Monte Carlo analysis is rarely the simulation. It is the translation: turning a distribution of thousands of outcomes into a sentence a client can act on without either panicking or becoming complacent.
This guide walks through how to frame probability of success, ruin probability, and the confidence around that number in client conversations and written reports.
What does "probability of success" actually measure?
Run your own numbers — FREE
10,000 Monte Carlo simulations. Forward-looking forecasts from BlackRock, JPMorgan, Vanguard, GMO, Schwab, Invesco. No account needed.
Try QuantCalc Free →Probability of success is the share of simulated scenarios in which the client's portfolio funds every modeled expense through the end of the plan without running out. If 10,000 scenarios are run and the money lasts in 8,700 of them, the plan shows an 87% probability of success.
Two clarifications are worth making explicitly to clients:
- It is conditional on the assumptions. The number reflects the return assumptions, inflation model, spending path, and time horizon you fed in. Change any input and the number moves. It is a measure of plan robustness under a stated set of assumptions, not a forecast of the future.
- "Success" is binary per scenario, but the failures are not all equal. A scenario that falls short by one year at age 94 is very different from one that fails at 78. This is why ruin probability and the timing of shortfalls matter as much as the headline percentage.
Run your own numbers in the free calculator →
How do you reframe the number so a client can use it?
The instinct is to chase 100%. A plan engineered for 100% probability of success usually means the client is underspending — leaving lifestyle (or legacy) on the table to insure against scenarios that are themselves unlikely. The more useful framing is a target band:
| Probability of success | How to frame it |
|---|---|
| Below ~70% | The plan is fragile under these assumptions; revisit spending, timing, or savings. |
| ~75%–90% | A reasonable planning zone for most clients; monitor and adjust over time. |
| Above ~95% | Robust — but check whether the client is overinsuring against low-probability outcomes. |
The point of the band is to move the conversation from "is my number high enough?" to "what trade-offs am I making, and am I comfortable with them?"
What the number costs: the same plan under two withdrawal rules
There is a deeper problem than explaining the percentage. A probability of success answers a question nobody actually asked: what happens if this client never changes their spending, whatever the portfolio does, for the next 35 years? No retiree behaves that way, and no advisor would let them. The number prices a commitment to do nothing.
To put a figure on that, we ran four withdrawal rules across the same 10,000 Monte Carlo paths — 35 years, a 60/40 portfolio, common random numbers so every rule faces an identical set of futures. Two of them make the point:
| Static 4% real | Guyton-Klinger guardrails, starting at 5% | |
|---|---|---|
| Plan survival | 83.0% | 98.9% |
| Paths where the money runs out | 17% | 1.1% |
| Lifetime spending | Baseline | 15% more |
| Typical worst-case real spending cut | None — income never moves | 27% |
| Median real ending balance | $1.16M | 40% smaller |
The guardrail plan starts at a higher withdrawal rate, spends more over the retirement, and still fails far less often. Read as forecasts, those two numbers look like a free lunch. They are not, and the reason is the entire point.
Guardrails do not find extra money. They move risk from a place clients cannot tolerate — running out — to a place they can: a temporary pay cut. The static rule buys a flat income and a large legacy, and pays for them with a 17% chance of depletion. The guardrail rule buys higher lifetime spending and a 1.1% chance of depletion, and pays for them with variable income and a smaller legacy.
Those are two different products for two different clients. Neither is a prediction, and a single probability cannot express either one, because it collapses the whole decision onto the axis the client cares about least. Full parameters and per-rule results are in the underlying study: guardrails vs. the fixed 4% rule, quantified.
The two numbers to put on the screen instead
The reframe that changes a meeting is to stop showing a percentage and start showing the two portfolio values that will actually govern the client’s behaviour: the level at which spending should come down, and the level at which it can go up.
Both are arithmetic, not simulation. Guyton-Klinger sets its guardrails at ±20% of the initial withdrawal rate, so for a client retiring on $1,000,000 and spending $40,000 a year:
- the portfolio falls to $833,000 → trim spending to $36,000
- the portfolio rises to $1,250,000 → raise spending to $44,000
That is a rule a client can hold in their head, repeat to a spouse, and recognise when it arrives. It converts an unanswerable question — is 83% good? — into a decision they have already agreed to, and it gives you something to say in the next bad quarter other than re-running the projection.
The upside guardrail matters as much as the downside one. A great deal of advisory energy goes into protecting clients from depletion and comparatively little into telling them when they have permission to spend more — yet in our run that upside trigger is where much of the extra 15% of lifetime spending came from. Under a static rule, a client whose portfolio doubles keeps spending as though it had not.
These are simulated outcomes under one return model, one asset mix and one parameterisation of the rule; changing the bands or the starting rate moves every figure above. Guyton-Klinger also suppresses its spending cut inside the final 15 years of a plan, which flatters late-retirement results relative to a naive reading.
What do ruin probability and survival curves add?
Probability of success answers "do I make it?" Ruin probability and survival curves answer "if not, when?" A survival curve plots the share of scenarios in which the portfolio is still solvent at each age. It turns a single percentage into a picture: clients can see the age range where shortfalls begin to appear and how steeply risk rises late in the plan.
This is often more persuasive than the headline number. A client who is uneasy about an 85% plan frequently relaxes once they see that the failures cluster past age 92 — and that adjusting spending modestly in a bad market pushes those curves out.
How precise is the percentage itself?
Because Monte Carlo results come from sampling, the reported probability of success is itself an estimate. Running more scenarios narrows the uncertainty around it. QuantCalc reports a Wilson 95% confidence interval on the success rate so the precision of the estimate is visible — an 87% with a tight interval is a different conversation than an 87% computed from a handful of paths. For client-facing work, running at the higher path counts (QuantCalc PRO computes 10,000-path runs) keeps that interval tight enough that the headline number is stable from meeting to meeting.
What belongs in the client report?
When you hand a client a written plan, the probability-of-success number should never travel alone. A defensible client report pairs it with:
- The return and inflation assumptions behind it, stated plainly and sourced.
- A survival curve so the timing of risk is visible, not just the headline.
- The confidence interval on the success rate.
- The levers — spending, retirement date, allocation — and how the number responds to each.
QuantCalc's Advisor PRO tier ($129/month) produces a white-label PDF report and a separate four-page methodology supplement built for exactly this: the report goes to the client, and the supplement documents the engine, sampling regime, and assumptions for the planning file. Advisors who want to see the format before committing can download the live sample report at quantcalc.app/advisors.
Frequently asked questions
Is a higher probability of success always better?
No. Beyond roughly 90%, a higher number usually signals underspending rather than a better plan. The goal is a probability the client is comfortable with given the trade-offs, not the maximum achievable.
How many scenarios should I run for client work?
Enough that the confidence interval around the success rate is narrow and stable between runs. QuantCalc's free tier runs 3 simulations per day at 100 paths each for exploration; PRO runs 10,000 paths, which is the appropriate setting for a number you put in front of a client.
How do I explain a plan that fails?
Show the survival curve. Clients respond to when shortfalls appear, not just whether they might, and seeing the failures cluster late in life reframes the risk in a way a single percentage cannot.
QuantCalc is an independent educational tool, not affiliated with any referenced firm. Not financial advice.