How Doughsense Projections Work

Good financial planning should not be a black box. If a tool tells you that you can retire at 58, you deserve to know how it reached that number, what it assumed, and where it stops being certain. This page explains how Doughsense projects your finances, how the Monte Carlo simulation works, and what the UK tax modelling covers. It is also honest about the edges: the places where the model simplifies, and the things it deliberately does not do.

The projection engine

A projection is a month-by-month model of your financial future. Doughsense takes your accounts, assets, liabilities, income, and expenses, then rolls them forward through time using the assumptions you set:

• Growth rates on each asset, so investments and savings compound at rates you choose.
• Inflation, applied so that future pounds are shown in terms you can reason about.
• Contributions and withdrawals, including one-off events and recurring cash flow.
• Interest and repayment on liabilities, including promotional credit-card segments.

Because the assumptions are yours, the projection is not a fixed forecast handed down from on high. It is a model you can interrogate and change. Adjust a growth rate or a monthly contribution and the whole timeline recomputes, so you can see the effect of a decision before you make it.

The same engine walk produces every figure in the app: this month's cash flow and safe-to-spend numbers come out of the same computation as the value on Future's projection chart. There is no separate short-term model to fall out of step with the long-term one.

Doughsense projects at monthly granularity for the long term, with finer cash-flow detail for the near term. The point is not to predict the future to the penny. It is to give you a defensible, adjustable baseline for the decisions in front of you.

Monte Carlo: modelling uncertainty honestly

A single projection line assumes markets deliver the same return every year. Real markets do not. Some years are strong, some are brutal, and the order they arrive in matters: a bad run just as you start drawing on your savings costs more than the same run later. Monte Carlo simulation shows how wide the range runs for your plan, instead of hiding it behind one line.

Here is the method, stated plainly:

• Each simulation draws its market years from history in five-year blocks (a block bootstrap). A block keeps each year's return beside that year's inflation, so a year like 1974, when US shares fell 26% and US prices rose 12%, stays one year. A milestone test at a return and volatility you set draws each year's return from those instead, and holds inflation at history's average.
• The historical series is the S&P 500 total return and US CPI, covering 1928 to 2024.
• Each simulation then lives through its years one at a time, in order, each year's return beside that year's inflation. So a bad year just as you start drawing on your savings costs more than the same year later on.
• Future's range runs 1,000 simulations, centred on your own growth and inflation assumptions. It shows a best case and a worst case year for financial independence, and in how many simulations your money lasts past your life expectancy age.
• Test each milestone against market history runs 250 simulations at history's own returns, or at a return and volatility you set, and reports how many reach each milestone. It answers a different question from Future's range, so one milestone can read differently in each.

The result is a range, not a promise: not "you will retire at 58" but "in this many of the simulations, your plan held up." A risk event such as running out of money counts the simulations where it happens. It is a projection built on your assumptions, not a prediction, and a more honest picture than a single reassuring line.

An honest limitation

The Monte Carlo engine currently samples from US market history (the S&P 500 and US CPI). That history is long, high quality, and widely used for this kind of modelling, which is why it is a sensible starting point. But Doughsense is a UK-native planner, and US equity and inflation history is not a perfect stand-in for a UK investor's experience. We would rather tell you that plainly than pretend the distinction does not exist. Use the confidence figures as a guide to the shape and spread of risk, not as a UK-specific guarantee.

UK tax modelling

Projections that ignore tax drift away from reality quickly. Doughsense models the UK tax wrappers that matter most for long-term planning:

• ISA and LISA contributions and their allowances.
• Pensions, with marginal-rate relief applied to contributions.
• Tax-aware planning in the solver, so the figures it solves for account for the tax treatment of the money involved.

Tracking wrapper allowances keeps your projections realistic: contributions cap where they should, and tax treatment is applied rather than assumed away.

What the tax model does not cover

Depth matters, so it is worth being clear about the ceiling. Doughsense is not a full financial-adviser tax engine. It does not model defined-benefit pension mechanics, flexi-access drawdown rules such as MPAA or the fine detail of UFPLS and PCLS withdrawals, inheritance tax and estate planning, offshore bonds, EIS or VCT investments, or trusts. A partner in your plan is taxed as an individual, on their own tax schedule and allowances. Nothing that links two people's tax is modelled, such as the Marriage Allowance or the High Income Child Benefit Charge.

If your plan hinges on any of those, treat Doughsense as one input rather than the last word, and take professional advice for the specialist parts.

Why we publish this

Transparency is the point. The people who plan carefully tend to distrust tools that hand over a number with no working shown, and they are right to. Publishing the method, the data behind it, and its limits is how a projection earns trust. If something here is unclear or you think we have a detail wrong, tell us: hello@doughsense.com.