How the numbers are worked out

1. Thousands of futures, not one forecast

Real markets never repeat the same return twice. So a single projected line, however carefully drawn, is always wrong in a way that’s hard to see. Instead the tool runs your plan through thousands of simulated market futures (typically 2,000), a technique called Monte-Carlo simulation. Each future draws a different sequence of yearly returns, and plays your plan through it, year by year.

That’s what makes the outputs honest about uncertainty:

• The success rate is the share of those simulated futures in which your money lasted the whole plan. 85% means that in 85 of every 100 simulated futures, you didn’t run out. It’s a probability under the stated assumptions, not a guarantee and not a prediction.
• The range chart shows the spread across the futures: a central path (half did better, half worse), with strong-market and weak-market bands around it.
• Where a single line is shown, it’s labelled a central projection: the median-return path, shown for readability, with the probability figure alongside it.

Repeating a simulation gives the same answer every time (the random sequences are seeded), so results are stable and testable.

2. Everything is in today’s money

All inputs and outputs are in real terms: £30,000 always means what £30,000 buys today. The simulation uses real (after-inflation) returns, so inflation is inside the numbers rather than a separate guess you have to make. Where something is naturally quoted in cash terms, such as a mortgage rate or a level (non-indexed) pension, the tool converts or erodes it explicitly and says so.

3. Return assumptions: below long-run averages, always editable

Each year of each simulated future, invested money grows by a randomly drawn real return with a stated average and volatility, and charges are deducted. The default return assumptions sit deliberately below long-run historical averages. Every one is an editable slider: a starting point, not a claim about the future.

4. Why the figure you see is lower than the assumption

Two different numbers are in play, and it is worth separating them. The assumption above is an average of the returns drawn each year: what goes in. The figure the tool reports is the median outcome: the middle of all the simulated futures, where half come out better and half worse, which is what comes out. The outcome is always lower than the assumption that produced it, for two reasons.

• Charges come off every year, whatever the market did.
• Ups and downs drag on growth. A year of −20% followed by a year of +20% is not a flat two years, it leaves you down about 4%. The more a return bounces around, the further the typical outcome falls below the average, which is why volatility is an assumption in its own right rather than a detail.

On the defaults above, an equity assumption of 4.5% , before charges a year works out at roughly 2.7% a year as a typical outcome once charges and that drag are taken off, and a 60/40 mix with cash lands near 1.5%. Those are the same figures, not different ones: one is what goes in, the other is what tends to come out. We report the median because reporting the average would flatter the picture, and more than half of the simulated futures would fall short of it.

5. UK tax is modelled properly, from one source of truth

Each simulated year applies the actual UK rules to that year’s income and withdrawals:

• Income tax: the personal allowance, the basic, higher and additional bands, and the personal-allowance taper above £100,000 (the “60% band”).
• Pensions: the 25% tax-free lump sum (capped by the Lump Sum Allowance), taken upfront, spread or staged; the money-purchase annual allowance where a pension has been flexibly accessed; the State Pension deferral uplift.
• Savings & investments: the savings and dividend allowances, and capital gains tax on investments held outside wrappers.
• Inheritance tax: nil-rate and residence nil-rate bands, spousal transfers, the 7-year gift taper, the 36% charitable rate, and the April 2027 change that brings unused pensions into the estate.
• Fiscal drag: frozen thresholds are held frozen for the legislated period, then indexed, so allowances quietly shrinking in real terms are modelled rather than ignored.

All rates and thresholds live in one configuration object in the engine. The in-app Model Assumptions page renders directly from it, and an automated test suite (currently 300+ tests) checks the arithmetic against independently worked HMRC / gov.uk examples, so a tax figure can’t quietly drift from the rules, or from what the page says.

6. How a year actually plays out

Each simulated year runs in the same order. Guaranteed income arrives first (the State Pension, final-salary or DB pensions, and annuities), net of tax. Any employment or other income is added. The remaining spending need is drawn from your pots in your chosen order, with tax applied to pension withdrawals. One-off costs, gifts and life events land in their year. Then what’s left grows or falls with that year’s drawn return, less charges. If a year’s spending can’t be met, that future has failed, and that’s what the success rate counts.

7. What the tool deliberately does not do

• It doesn’t recommend. No product, fund, provider, allocation or course of action is ever named, ranked or suggested. Comparisons are neutral side-by-sides worked out from your own figures.
• It doesn’t predict. The assumptions are illustrative and editable, and actual returns, tax law and your circumstances will differ.
• It doesn’t know everything about you. A regulated adviser weighs up your full circumstances; this tool works only from what you enter.