How the numbers in this tool are actually calculated

The problem with asking one question

Think about how a client answers “are you comfortable with risk?” The honest answer depends heavily on when you ask it. Ask during a strong bull market and most clients will say yes. Ask the same client a week into a sharp sell-off, or right after reading a worrying headline, and you will often get a different answer from the same person. A single question like that isn't measuring something stable about the client. It is measuring their mood on the day, how the market happened to perform recently, and how the question was worded. None of that is a sound basis for building a portfolio that needs to hold up over years.

That is the reason this tool asks 20 shorter questions instead of one big one, and adds up the scores rather than trusting any single answer. If a client misunderstands one question, or answers inconsistently because they are distracted, that one answer only moves the total score slightly. It cannot flip the whole result the way a single yes-or-no question could. Spreading the measurement across many questions is a simple, well-understood way of making a subjective judgment more reliable.

Three things being measured, not one

The 20 questions are not all measuring the same thing. “Risk tolerance” is actually three separate ideas that people often blur together, so the questions are spread across all three:

• Willingness. How does the client feel, in general, about uncertainty and the possibility of loss? This is the emotional side. Some people are naturally more anxious about volatility than others, independent of their actual financial situation.
• Capacity. Regardless of how the client feels, can their actual circumstances absorb a bad run without real damage? A client with a long time horizon, stable income, and other assets to fall back on has a much higher capacity for risk than someone nearing retirement with no other savings, even if both say they feel the same way about risk.
• Track record. How has this client actually behaved the last time markets fell? What someone expects they would do in a downturn, and what they actually did the last time one happened, are frequently two different things. Past behaviour is usually the more honest signal.

How the scoring works

Each answer is worth between 1 and 5 points depending on how it was phrased, all 20 scores are added together, and the total places the client into one of five bands running from Conservative through to Aggressive. Five bands is a deliberate middle ground: enough granularity that a Cautious client and a Balanced client are treated meaningfully differently, but few enough that the result is still something you can explain to a client in one sentence, rather than a confusing scale of fifty.

This general approach, asking a batch of varied questions and scoring the total rather than relying on a single item, is the same underlying principle behind established, internationally used risk-tolerance research, including the widely cited work of Grable and Lytton from the late 1990s (see references below). It is also the same idea behind large commercial risk-profiling questionnaires used by wealth managers internationally. This exact 20-question set has not been through the years of independent academic validation that those benchmark instruments have. What is being borrowed, honestly, is the underlying logic: many small, varied measurements averaged together are far more trustworthy than one confident-sounding question.

Overrides and the portfolio conflict check

The resulting band is a starting point for a conversation, not a verdict handed down by the software. If your professional judgment says the band does not fit the client, you can override it manually in a couple of clicks, and the system records that it was an intentional override rather than the raw questionnaire result, so there is a clear record of the decision later. Once a portfolio has actually been built for the client, the tool also checks whether the proportion held in growth assets (shares and property) is roughly in line with what is typical for that risk band. If a client comes back as Cautious but the proposed portfolio is 80% invested in equities, the tool will quietly flag that mismatch. It will not stop you from proceeding: there may be a perfectly good reason, such as a short-term goal sitting alongside a long-term retirement goal. The point of the flag is simply to make sure a mismatch like that is a deliberate, documented choice rather than something that slipped through unnoticed.

Why six rules instead of one

Ask five different financial planning textbooks how much capital someone needs at retirement and you will get five different rules of thumb, and none of them is simply wrong. Each one was built by a different practitioner or researcher, working from a different simplifying assumption about how retirement drawdowns actually behave. Rather than picking one of those rules and hoping it happens to suit this particular client, the calculator runs six well-established, independent rules side by side and takes the average. This mirrors how a property valuer works: nobody trusts a single valuation method on its own, but when several independent methods land in a similar range, that agreement is genuinely persuasive.

Family one: the replacement-ratio rules

The 80% Rule and the 75% Rule start from the question “how much of the client's current income will they actually need once they stop working?” The 80% Rule assumes they will need 80% of their final salary each year to maintain roughly the same standard of living, once costs like commuting and retirement savings contributions fall away. The 75% Rule uses the same idea with a slightly leaner assumption. Both rules then project that required income forward, every year, for as long as the client is expected to be retired. Rather than assuming the leftover capital just sits still earning nothing while it is drawn down, which would overstate how much is really needed, the calculator assumes it keeps quietly earning a modest 2% a year above inflation. That figure is deliberately conservative rather than optimistic: assume too high a return and you risk leaving a client under-funded if markets disappoint; assume zero and you risk telling a client they need far more than they realistically do. This assumption can be adjusted on the calculator if you hold a different house view.

Family two: the withdrawal-rate rules

The 4% Rule, the 300 Rule, the 15x Rule, and the R1m-for-R5k Rule come at the problem from the opposite direction. Instead of projecting every year of retirement individually, they ask what multiple of annual or monthly income, saved as a single lump sum today, is unlikely to run out over a normal retirement. The 4% Rule is the best known version: it comes from research into historical safe withdrawal rates, and holds that withdrawing 4% of the starting capital in the first year, then adjusting that amount for inflation every year after, has historically had a strong chance of lasting 30 years or more. The 300 Rule expresses that same 4% logic as a quick mental-math multiple of monthly income instead of a percentage. The 15x Rule and the R1m-for-R5k Rule are similar quick multiples already used in South African adviser conversations, each built around a slightly different implied withdrawal rate. The calculator runs all four consistently and shows its working, rather than asking you to remember which multiple goes with which assumption.

What the average is, and is not, for

The average of all six numbers is a fast, defensible opening figure for the retirement conversation. It is not, and is not meant to be, a substitute for a full actuarial retirement plan that accounts for this specific client's tax position, medical costs, and other individual circumstances. Its job is to give you a credible starting number quickly, which then gets properly stress-tested in the Monte Carlo simulation described next.

Why one blended pot is the wrong starting point

It is tempting to treat a client's entire investable wealth as a single undifferentiated portfolio: one allocation, one number, one plan. The problem is that real clients rarely have just one goal, and different goals do not share a time horizon. A deposit for a car needed in three years and a retirement pot needed in thirty years cannot sensibly sit in the same allocation. Whichever mix you pick will be wrong for at least one of them: too aggressive for the near-term goal, which cannot recover from a bad year right before the money is needed, or too conservative for the long-term goal, which has decades to ride out short-term volatility and should be allowed to.

Each goal is its own portfolio

Waypoint builds a separate portfolio for every named goal a client has: its own asset allocation, matched to that goal's own time horizon and required outcome, and its own schedule of contributions and withdrawals. Each goal is simulated and reported on its own terms, with its own probability of success. In practice this means a goal three years away is deliberately held conservatively, weighted toward cash and bonds, because a sharp downturn shortly before the money is needed cannot be recovered from in time. A retirement goal thirty years away can reasonably hold far more in growth assets, because there is enough time for a bad run of markets to be followed by a recovery long before the money is actually required.

The general-purpose bucket

Not every rand a client holds is earmarked to a specific goal on day one, and it shouldn't need to be before it can be invested sensibly. Capital that has not yet been assigned to a named goal sits in a general-purpose bucket, typically allocated in line with the client's overall risk profile from the questionnaire, since there is no specific goal horizon yet to anchor it to. New contributions land here by default until they are assigned to a goal, so a client is never forced to name every future goal before the tool can do anything useful with their money.

What the total portfolio view is, and isn't

A client's total wealth, as shown in a scenario report, is simply the sum of every goal bucket plus the general-purpose bucket. Reg 28 compliance and fund matching are still checked at that combined, blended level, because a regulatory limit does not care which bucket a particular rand happens to sit in. Probability of success, on the other hand, is always reported per goal, never as one number for the whole relationship, because “an 80% chance of success” is meaningless until you know which goal that 80% belongs to. A client can be highly likely to fund their retirement and, at the same time, unlikely to comfortably afford the car they want in three years. Collapsing that into a single blended number would hide exactly the tension a goals-based conversation is meant to bring into the open.

Why a single projection line is misleading

The easiest way to project an investment forward is to pick one average annual return, say 9%, and compound it every year until retirement. It is a comfortable number to put in a spreadsheet, and it is also fiction. Markets never actually move in a smooth 9% line. Some years they return 25%, some years they lose 15%, and the order in which good and bad years happen matters enormously, especially for a client who is drawing an income from the portfolio while those swings are happening. A single straight-line projection hides all of that risk from the client. Monte Carlo simulation is the standard tool used across the investment industry to stop hiding it.

Instead of one projection, the simulation runs the client's portfolio forward through several thousand different possible futures. In each one, every month gets its own randomly generated return for every asset class in the portfolio, drawn in a way that is statistically consistent with how that asset class has actually behaved historically. Run that process a few thousand times and the result is not one number. It is a realistic spread of outcomes: some simulated futures are kind to the client, some are unkind, and most sit somewhere in between. That spread is far more useful to show a client than a single hopeful average, because it prepares them for what a genuine bad patch could look like before it happens in real life.

Where the randomness comes from

For each asset class, the tool takes its historic monthly returns and fits a statistical curve describing the pattern: where returns tend to cluster, and how often unusually good or unusually bad months occur. It does not simply assume a standard bell curve and move on. It specifically checks whether the historical data shows more frequent extreme months than a plain bell curve would predict, a pattern investment researchers call “fat tails.” This is not an abstract statistical concern. Events like the 2008 financial crisis or the COVID-19 market crash of 2020 were the kind of moves a simple bell curve would treat as almost impossible, yet they happened, and markets have produced sharp shocks like that repeatedly over the decades. Where the historical data for a given asset class supports it, the simulation uses a curve shaped to allow for those fatter tails, so the range of simulated outcomes is not artificially narrow. Where the data does not support it, it uses the simpler bell curve instead. That choice is made automatically for each asset class using a statistical test that checks how well each type of curve actually fits the real historical data, not by defaulting to one option out of habit. Which curve was used, and why, is shown on the Simulate step.

Why the asset classes move together

Shares, property, bonds and cash do not move independently of each other. When equity markets fall sharply, listed property usually falls with them, because both are sensitive to the same economic conditions, while cash typically holds its value and government bonds often do relatively well. A simulation that generated returns for each asset class completely independently would occasionally produce nonsense, such as equities crashing while property soars, with little basis in how markets actually behave. To avoid that, the tool measures how each asset class has historically moved in relation to the others and carries that relationship into every simulated month, so a bad month for growth assets in the simulation realistically affects the whole growth portion of the portfolio together.

What happens inside each simulated path

The simulation walks forward one month at a time. In each month it applies that month's randomly generated, correlated return to whatever the portfolio's asset allocation is at that point (allowing for the fact that a client's allocation can change partway through the plan), then applies any contribution or withdrawal scheduled for that month, and moves on to the next month. It repeats that walk a few thousand times, each with a different random sequence of returns, and records where the portfolio ended up in every single one.

Reading the results

Once every simulated path has finished, the tool sorts all the ending portfolio values and reports five points along that spread: the median (the outcome sitting exactly in the middle of every simulated future, a reasonable stand-in for “typical”), a below-average and an above-average outcome (the 25th and 75th percentile), and a realistic worst-case and best-case boundary (the 0.1st and 99.99th percentile). As an example, if the median outcome is R3 million but the worst-case boundary is R700,000, that gap is genuinely important information for the client to see now, while there is still time to adjust contributions or the asset allocation, rather than discovering it only after a bad decade has already happened. If a goal amount was set, the tool also reports what percentage of the thousands of simulated futures actually reached that goal, giving a direct, honest answer to “how likely am I to get there?”

None of this is a novelty invented for this tool. Running many randomly generated scenarios to understand a range of financial outcomes, rather than trusting one average projection, is the same general technique described in modern portfolio and stochastic modelling theory, and it is used in some form by actuaries, asset managers, and retirement fund trustees around the world to stress-test whether a plan is likely to hold up. Waypoint runs that process automatically and shows its working at every step, so it functions as a transparent tool you can stand behind in front of a client, rather than a black box that simply hands you a number.