A: Yes. After you've signed up, you can create a plan for either a single person or a couple. The program models couples using the tax rules that apply to married couples and civil partners. Unmarried couples can still use the program, but the modelled projections will reflect those tax assumptions.
A: Some online tools ask for very detailed information and then apply equally detailed rules when projecting the future. However, these rules often rely on simplified assumptions. For example, assuming identical investment growth every year or stopping projections at a fixed time horizon. Real-world returns vary from year to year, and no one can predict with certainty how long they will live.
EvolveMyRetirement® takes a different approach. It captures the key elements of your finances and attitudes and then uses simulation to explore a wide range of possible future scenarios. This allows the projections to reflect uncertainties such as investment returns, property prices, inflation, and longevity.
EvolveMyRetirement® is not a substitute for independent financial advice, and we recommend consulting a qualified adviser before acting on any modelled strategy. You can share your plan directly with your adviser to support that discussion.
EvolveMyRetirement® isn’t designed to give specific investment advice or to recommend particular asset classes. Instead, it focuses on modelling strategy options at a high level, based on the information and preferences you enter, including your attitudes to risk and leaving a legacy. The program uses assumptions about investment returns, including both average returns and volatility, and you can adjust these if you wish.
Decisions about how to achieve any particular mix of investments fall outside the scope of this website and should be discussed with an independent financial adviser.
When you set a target cash buffer, the program treats that amount as being held in the lowest-risk return category defined in your plan assumptions. Because lower-risk assets typically have lower expected returns, holding more cash can lead to different projected outcomes compared with investing the same amount in higher-risk assets.
Some people prefer to keep a cash buffer because they find it reassuring not to rely on selling investments at specific times. Others prefer to minimise cash holdings. EvolveMyRetirement® doesn’t take a view on which approach is better; it simply reflects your choice, and can show you how different choices affect the projections.
If you’re unsure how much cash buffer is appropriate for your circumstances, it may help to discuss this with an independent financial adviser.
A: When you enter either an Employment or a Self-Employment for a Member, you specify the retirement date from that work, and also whether it's pension-related (i.e. pension contributions are made or there's a DB pension). Once the Member has retired from all pension-related work and has also passed the earliest age at which pension funds may be accessed (currently 55, rising to 57 in 2028), the program models the Member as being retired for projection purposes.
So it's possible for a Member to be treated as retired in the model while still in work, provided there are no related pension contributions. Also, if a Member stops all work very early, say at 45, then the model only treats the Member as retired once they reach the earliest age at which pension funds may be accessed.
A: EvolveMyRetirement® is designed to cover many common financial situations. There may be cases where the most appropriate choice of inputs is not immediately obvious. If you’re unsure how best to represent your circumstances in the model, please feel free tocontact us and we’ll be happy to help.
A: Yes. If your plan uses a Scottish postcode, the program applies Scottish income tax rules when modelling future projections.
A: If you’re not a Premium User, you can still include potential long-term care costs by adding an amount to the lifetime contingency on the Plan page. This allows the model to incorporate an estimate of care-related spending.
Premium Users can model long-term care more explicitly by selecting the relevant option on the Plan page and entering an annual cost per member. The timing of long-term care is randomised in the simulation. For each member, the number of years before death is drawn from a normal distribution, with the mean and standard deviation set in the Assumptions section. If the randomised value is negative, the model treats that member as never requiring long-term care.
This randomisation approach is a simplification, but the default settings are broadly consistent with published data. The study we reference categorises long-term care needs as Low, Medium, and High; our model and defaults focus on the Medium and High categories.
When long-term care begins for a member, any member-specific spending marked as ceasing in care will stop entirely. Joint spending is reduced in the same way as if the member had died, and discretionary spending ceases for any member in care.
In the simulation, the main home (if any) is normally retained until the death of the last member. However, if all members are in care and all other sources of cash have been exhausted, the model assumes the home may be sold to meet expenses. This is a modelling assumption and not a recommendation about real-world decisions.
A: Once your plan is complete, and you've generated the results, there's a button on the Results page to download a PDF report summarising your inputs, assumptions, strategy settings, projections, and charts. This file is perfect for printing, or you can save it or attach it to emails. This feature is available to Standard and Premium users.
A: You can delete a plan even if it’s your only one. On the Plan page, scroll to the bottom, and you’ll see a “Delete Plan” link next to the button for updating the plan. Using this option will permanently remove all the information in that plan, after which you can create a new one if you wish. If you have more than one plan, deleting one will not affect any of the others.
A: Yes. On the Spending page, you can choose a payment frequency of monthly, yearly, or one-off. If you select one-off, the start date you enter is treated as the date on which the amount is spent in full.
A: If you enter your property-related expenses on the Spending page, the model has no way of knowing which property they relate to. That may not matter if your only property is your main home, and you don’t plan to resize it. But if you own a second property that might be sold during the projection (whether planned or forced), the model needs to know which expenses are linked to that property so they can stop when the property is sold. For this reason, property-specific expenses are best recorded against the relevant property on the Assets page.
Similarly, if you plan to resize your main home, the model will automatically adjust associated expenses at the appropriate time. This avoids the need to create separate sets of spending entries on the Spending page.
A: This is a common question, because discretionary spending behaves differently from other spending categories in the model. A few points help explain how it works:
1) The Spending page captures your current discretionary spending, not the level used in the strategy.
As noted on the Spending page, the model determines discretionary spending later as part of the strategy settings. The amount you enter on the Spending page represents your current spending pattern, which may differ from the level used for future projections.
2) The Strategy does not use the discretionary spending entered on the Spending page.
This doesn’t make the entry pointless – it allows the Results page to highlight whether the strategy implies higher or lower discretionary spending than your current level.
3) The Spending page chart includes your current discretionary spending.
This can be useful for comparison. Once you're comfortable with the discretionary spending level shown in your strategy, you can update the Spending page to match it if you want consistency across the plan.
4) The checkbox on the Plan page relates only to how discretionary spending changes over time.
Unchecking it means that the level of discretionary spending defined in the strategy (not on the Spending page) will simply increase with inflation each year in the projections. It does not prevent how the initial level is determined.
5) Premium users can lock discretionary spending in the strategy.
With a Premium subscription, you can manually set the discretionary spending level in the strategy and lock it before generating a strategy. The model will then only consider strategies that use that starting level.
When this option is enabled, the model adjusts discretionary spending from year to year based on how your projected net worth compares with the model’s expected path. At the start of each projection year, the program compares your plan’s net worth (adjusted for the estimated present value of any guaranteed income) with the value it would expect in an average-case scenario without any uncertainty. The ratio between these two figures is then used to adjust discretionary spending for the following year.
If the model indicates that discretionary spending could be increased, the adjustment is moderated according to the legacy importance setting in your plan. For example, if the plan specifies the maximum legacy importance, discretionary spending will not rise above its initial real level. Decreases are not moderated in the same way.
This mechanism is a modelling heuristic designed to reflect how discretionary spending might vary in response to changing circumstances within the simulation. It is not intended as guidance on how to manage real-world spending.
If you periodically update your plan based on real-world changes, you can regenerate your strategy at any time. When you do so, the model treats the current date as the starting point and uses the most up-to-date information available, without reference to previous projections.
A. Let's take a brief look at each of these approaches.
Different approaches to adjusting spending exist, and each is based on its own assumptions. The method used in EvolveMyRetirement® reflects the structure of the model and the way it represents a household’s overall finances. A brief comparison may help clarify the differences.
Guardrails
Guardrails are typically designed around portfolio withdrawals. In EvolveMyRetirement®, withdrawals are treated as one component of a wider financial picture that also includes property, debts, state pension, other pensions, annuities, and any other assets or income sources. Because the model evaluates spending in the context of the wider financial picture rather than portfolio withdrawals alone, the 'guardrails' framework does not map directly onto how the simulation is structured.
Dynamic spending
Dynamic spending rules share some similarities with the approach used in the model, but they generally focus on investment portfolios and often treat all spending as adjustable. EvolveMyRetirement® distinguishes between essential spending and discretionary spending. The model bases adjustments on a measure of net worth that includes the estimated present value of guaranteed income. This provides a broader indicator than portfolio values alone. The model also moderates increases in discretionary spending according to the legacy-importance setting, rather than using fixed floors or ceilings.
The approach actually used
The method used in EvolveMyRetirement® is designed to work consistently with the way the model represents all assets, liabilities, and income sources. It is also structured to operate within a Monte Carlo simulation, where spending adjustments need to respond to changes across the full financial picture rather than a single component such as an investment portfolio.
A: That depends on what you want the model to illustrate. When the box is not checked, the projections use a constant level of discretionary spending in real terms (other than any fixed taper you may have specified). This produces a projection in which discretionary spending does not change over time.
When the box is checked, the model allows discretionary spending to vary from year to year in response to changes in the simulated financial position. This reflects how spending might adjust within the model when circumstances differ from the average-case path.
Some users choose to leave the box unchecked while generating a strategy and then enable it afterwards to see how variable discretionary spending behaves in the projections. Others leave it checked throughout. The model takes the setting into account either way.
Whichever option you choose, the model always treats the current day as the starting point for its projections. If you update your plan in the future to reflect changes in your circumstances, you can regenerate your strategy at any time. When you do so, the model uses the most up-to-date information available, without reference to previous projections.
A: The average legacy is based on the model’s average scenario. In that scenario, net worth follows the expected path without any deviations. When the option to vary discretionary spending is enabled, adjustments would be expected to occur in those years in which net worth differs from the expected value for that year. But because the average scenario never deviates from its expected path, no adjustments are triggered. As a result, the checkbox setting has no effect on the average scenario, and therefore no effect on the average legacy.
However, the checkbox setting does affect the distribution of possible legacy outcomes across all simulated scenarios. When discretionary spending is allowed to vary, the model adjusts spending in response to differences between simulated net worth and the expected path. This influences how widely net-worth outcomes can diverge from the average path within the simulation.
You can see this by running a Survival Analysis for each setting (Premium users may run unlimited Survival Analyses) and comparing the chart titled “Real Net Worth Confidence Intervals By Year”. With variable discretionary spending enabled, the confidence intervals will typically be narrower, reflecting reduced divergence from the expected path within the model. Since legacy depends on net worth at death, this affects the spread of possible legacy outcomes, even though the average legacy remains unchanged.
A: There is a dedicated page that explains the projection rules used by the model. You can access it using this link which describes how each year of the projection is calculated.
A: EvolveMyRetirement® uses mortality tables published by the Office for National Statistics. Based on gender, date of birth, and region, life expectancy is derived from the probabilities in those tables and is randomised during each Monte Carlo simulation.
You can manually adjust the average life expectancy up or down if you wish to reflect personal considerations such as health or demographics. Changing this setting will influence the projection period used in the model, so it is worth ensuring that any adjustment reflects information that is relevant to your circumstances.
A: When you create your plan, it is given a set of default assumptions. You can view and modify these in the Assumptions section on the Plan page. The defaults are intended to provide a starting point, but you are free to adjust them if you prefer different values.
If you are working with a financial adviser, and they have specific views about assumptions such as inflation or asset growth, they (or you) can update the figures directly in your plan.
Premium users can also create multiple plans, each with its own customised set of default assumptions.
A: Under current UK rules, pension tax relief is available only before age 75. In the model, any employer contributions made after that age are treated as additional taxable salary. Employee and self-employed contributions are not included in the projections after age 75, and any target percentage of earnings specified on the Strategy page is not applied beyond that point.
A: During projection, the model releases equity from the main home only when other sources of funds are not available within the simulation and only in the amounts needed to meet projected outgoings for that year. The amount of equity that can be released is limited by age, based on up to 25% of the primary-home value at age 55, increasing on a sliding scale to 50% at age 85 or above. Equity release is not modelled for anyone under age 55.
The model assumes that equity release takes place through a reverse-mortgage-style mechanism unless all members are in full-time care, in which case the main home is assumed to be sold. The interest rate applied to reverse-mortgage borrowing can be adjusted in the Assumptions section (secured borrowing).
Premium users can also choose to disable equity release entirely within their plan.
A: Yes. EvolveMyRetirement® incorporates current rates and bands for income tax, National Insurance, capital gains tax, and inheritance tax, and applies these within its projections. Tax rules for ISAs and pension plans are also included. You can view the tax rates and rules used by selecting the Projection Assumptions button at the bottom of the Results page.
The model calculates tax based solely on the information entered into your plan, so the resulting figures are an approximation of how tax might apply within the projection.
A: Risk aversion influences how the strategy is constructed, but the projection results depend on how the simulated scenarios unfold. In some scenarios where the model shows all funds being exhausted, this may occur close to the end of the final member’s life, which affects how the outcome appears in the statistics. In other scenarios, the model may show funds being exhausted even though significant assets remain, reflecting situations where projected outgoings exceed available liquid resources at that point in the simulation.
The percentage shown also depends on whether discretionary spending is held constant or allowed to vary. If discretionary spending is fixed in real terms, the model does not adjust it in response to changes in the simulated financial position. If discretionary spending is allowed to vary, the model adjusts it year by year within the simulation, which can change the distribution of outcomes. This setting is available on the Plan page.
In all cases, the model treats the current date as the starting point for projections. If your circumstances change in the future, you can update your plan and regenerate your strategy so that the projections reflect the latest information.
A: The pension contributions you enter for employments and self-employments represent your current contribution levels. The strategy, however, sets out a target percentage of earnings for future projections. This target may be higher or lower than your current contributions. If it is higher, the model uses the higher amount in the projections, subject to sufficient funds being available.
In some cases, the model may project contributions above the target percentage if surplus funds are available and the strategy settings allow for additional pension contributions. This is limited by the maximum percentage of earnings specified in the strategy. For some members, this cap can influence how contributions are modelled before retirement.
A: The target taxable retirement income is a strategy setting that guides how the model calculates pension drawdown during retirement. Earlier versions of the model used a fixed percentage drawdown from pension funds, but this approach did not reflect situations where taxable income from other sources varies over time.
The current approach sets a target level of taxable income for each member in retirement, increasing annually with inflation. For members who have not yet retired, this setting applies only from the point of retirement onwards. Each year, the model calculates the pension drawdown needed to reach the target taxable income after taking into account all other taxable income sources. This results in drawdown amounts that vary from year to year within the simulation.
If the target taxable income isn't enough to meet projected outgoings, the model uses non-pension assets where available. If a shortfall remains, additional pension drawdown may be modelled, so the target is not treated as a strict maximum.
If the target taxable income exceeds the amount needed to meet outgoings, any surplus is modelled as being reinvested, into ISAs where possible, otherwise into taxable investments.
When the model generates a strategy, it aims for a target taxable income that reflects several factors within the simulation, including:
• Projected outgoings.
• How taxable income interacts with tax allowances and bands over time.
• The role of pension funds in the projected legacy.
• The stated importance of leaving a legacy.
A: Manual strategy editing is available with a Standard or Premium subscription. If you do not currently have one of these subscriptions, the editing controls remain inactive. Subscription options can be found on the Plan page.
EvolveMyRetirement® represents investment risk using a single sliding-scale measure rather than explicit allocations to multiple asset classes. When designing the model, we considered using separate categories such as equities, bonds, and cash, but this would have required detailed assumptions for each asset class and additional strategy settings. To keep the model manageable, we use a simplified approach with two reference points: lowest-risk asset and highest-risk asset.
The lowest-risk end of the scale (0) is treated as an income-oriented asset with minimum volatility. The highest-risk end of the scale (100) is treated as a growth-oriented asset with maximum volatility. The assumptions for the highest-risk asset, such as mean growth, standard deviation, and bid/offer spread, can be adjusted in the Assumptions section if you wish to reflect different views or research.
Intermediate risk levels represent a proportional blend of these two reference points. For example, a risk level of 60 corresponds to mixing 60% of the highest-risk asset with 40% of the lowest-risk asset within the model’s framework. The model applies the corresponding growth and income characteristics based on this blend.
This approach provides a consistent way to model investment risk without requiring detailed configuration of multiple asset classes.
A: EvolveMyRetirement® uses Monte Carlo simulation. The variability of market returns is generated using probability distributions that are informed by historical market data, but the model does not reproduce actual historical sequences of returns.
There are two reasons for this. First, past sequences do not repeat in the same order. Second, historical return data covers only a limited period, which would restrict the number of scenarios available. Monte Carlo simulation allows the model to generate a much larger number (10,000) of independent scenarios, based on the statistical characteristics of past data rather than the specific historical paths.
A: Different probability distributions are used in different contexts:
• A uniform distribution is used for the employment bonus, between the specified minimum and maximum.
• A normal distribution is used for the rate of inflation, using the mean and standard deviation specified in the plan's assumptions. Although inflation may be randomised as negative, there's a lower limit of -1%. Any randomised inflation rate below this will be treated as -1%.
• A log normal distribution is used for the growth rate of investments and properties. The mean and standard deviation are specified in the plan's assumptions. For investments, the mean and standard deviation are only specified for the lowest and highest risk investments (corresponding to investment risks of 0 and 100 respectively in the strategy). For risks in between, the mean and standard deviation are calculated based on a pro rata mix of the lowest and highest risk investments, assuming zero correlation. Note that investment growth rates include any reinvested income.
• For life expectancy, the model does not use a standard probability distribution. Instead, it uses mortality tables from the Office for National Statistics. Based on each member’s biological age at the start of the projection, along with gender and UK region, the year of death is randomised by stepping through the relevant mortality table.
A: Straight-line projections assume that investments grow at the same rate every year. This can give a simplified picture of how a portfolio might develop, but it does not reflect the variability of real-world returns. For example, an average annual return of 7% implies that an investment might double over ten years on average, but actual outcomes can be higher or lower depending on how returns vary from year to year.
Because investment returns fluctuate, two sequences with the same average return can produce very different outcomes. This is particularly relevant during retirement, when withdrawals are being made: the timing of market rises and falls can influence how long a portfolio lasts within a projection.
Monte Carlo simulation incorporates this variability by modelling many different possible return paths rather than a single straight-line path. This allows the model to reflect a range of potential outcomes rather than relying on a single average figure.
A: EvolveMyRetirement® compares strategies by assigning each one a numerical score called utility. In this context, utility is a model-generated measure of how closely a strategy aligns with the preferences and assumptions set out in your plan when tested across many simulated scenarios.
To calculate this score, the model runs a Monte Carlo simulation. Each simulation generates a large number of possible future scenarios based on the plan’s inputs and assumptions, and the probability distributions used for investment returns, inflation, life expectancy, and other uncertainties. The outcomes of these scenarios are then analysed using several factors, including:
• The level of spending achieved in each scenario.
• The estimated probability of running out of money.
• The estimated probability of ending up with negative net worth.
• The size of any legacy.
• The stated importance of leaving a legacy.
• The plan members' stated risk aversion.
A: EvolveMyRetirement® treats retirement dates as personal choices rather than values that can be automated. Although it would be possible in theory to include retirement dates in the strategy comparison process, doing so would require the model to quantify trade-offs that are highly subjective. For example, it would need to assign a numerical value to an extra year of free time compared with having a different level of discretionary spending throughout the rest of your life.
Because these judgements vary widely between individuals and cannot be expressed reliably through general assumptions, retirement dates are left for users to set manually.
A: Not in the way that can occur with AI systems based on Large Language Models (LLMs). EvolveMyRetirement® does not use an LLM for its strategy-comparison process, and it does not rely on external training data that might contain hidden biases. Instead, it uses a genetic algorithm. The data it works from is generated dynamically during the process and is based entirely on the specific contents of your plan.
The comparison of strategies is driven solely by simulated outcomes based on the plan’s inputs and assumptions. The only potential source of bias lies in the built-in projection rules that form part of the model. These rules are kept as minimal as practical, with the intention that as many aspects as possible are determined by the strategy settings rather than by fixed assumptions.
A: There's no fixed schedule for revisiting your strategy. EvolveMyRetirement® bases its calculations on the information in your plan at the time you run it, so any change in your circumstances or assumptions may lead to different results if you choose to run the process again.
Examples of changes that could affect your plan inputs include:
• A significant rise or fall in the value of your assets.
• Markets might rise or fall significantly resulting in a revaluation of your assets.
• Changes in annuity rates.
• Redundancy or an unexpected career change.
• A decision to retire earlier or later than originally planned.
• An unexpected windfall or increase in expenses.
• Changes in health for you or your partner.
• A change in your views about leaving a legacy.
• Bereavement.
Even if nothing else changes, the passage of time can affect the results because life expectancy changes with age. Also, the model may be updated periodically with revised mortality tables from the Office for National Statistics.
If any of your circumstances, assumptions, or preferences change, or simply as time moves on, you may wish to update your plan and run the process again so the model reflects your latest information.
A: EvolveMyRetirement® uses a genetic algorithm to explore different strategy settings. This type of algorithm includes random variation at each step, so the exact path it follows can differ from one run to the next. Because the model also uses Monte Carlo simulation, which introduces its own randomness, two runs based on the same inputs will not necessarily produce identical results.
In some cases, the model may find several strategies with very similar utility scores, even if their settings differ slightly. This can happen for a number of reasons, such as:
• Increasing a particular setting beyond a certain point may have no material effect on the simulated outcomes. For example, it may make little practical difference whether a person aged 60 intends to begin rebalancing investments in 35 years or in 40 years.
• Different combinations of settings can produce similar overall results. For instance, increasing discretionary spending may raise risk, but another setting may offset that effect in a way that leads to a similar utility score, depending on the risk preferences specified in the plan.
Because of these factors, small variations between runs are normal and reflect the way the algorithm explores the space of possible strategies rather than changes in your underlying plan.
A: Adjusting one setting at a time can be useful for fine-tuning, but it is unlikely to reproduce the outcomes produced by a genetic algorithm. The strategy settings interact with each other in complex ways, so changing one setting often affects the impact of others. A change in one area may require compensating adjustments elsewhere, which in turn may influence additional settings. Because of these interdependencies, exploring the settings manually on a one-by-one basis is generally impractical.
The genetic algorithm used by EvolveMyRetirement® evaluates many combinations of settings simultaneously and explores how they work together across simulated scenarios. If you believe your plan inputs have changed, you can run the process again to reflect your latest information and then make any minor manual adjustments you prefer.
A: The idea of gradually reducing investment risk as retirement approaches is often referred to as a “glide path”. Many traditional approaches assume that this reduction should be completed before retirement. EvolveMyRetirement® does not make that assumption. The model does not start with any view about whether investment risk should decrease, increase or remain constant, nor does it assume any particular timing for changes in risk levels.
Instead, the model evaluates many possible combinations of settings using Monte Carlo simulation and compares their utility scores based on the preferences and assumptions in your plan. Depending on those inputs, the simulated outcomes may indicate that a glide path completing before retirement, at retirement, or after retirement all produce similar or differing utility scores.
Because the model does not impose a predefined glide-path structure, the timing of rebalancing in a generated strategy reflects the simulated outcomes under your specific plan inputs rather than any general rule about when risk should change.
A: The lifetime contingency represents essential but unpredictable future expenditure. These are costs that cannot be forecast accurately in advance and fall outside the specific items listed in your plan. Because these amounts are inherently uncertain, they depend on personal judgement rather than on information the model can infer from the rest of your plan.
If you are unsure how to estimate this figure, you may wish to discuss it with an independent financial adviser. If you share your plan with an adviser, they can enter an amount that reflects their professional assessment of your circumstances. In some cases, insurance products may also help reduce exposure to certain types of unforeseen costs, which may influence the amount you choose to enter.
If the program attempted to determine the lifetime contingency automatically, it would tend to favour very low values. This is because reducing essential expenditure leaves more room for discretionary spending within the model, which increases the utility score. However, a very low contingency may not reflect the reality that unexpected essential costs can and do arise. For this reason, the lifetime contingency is left for users, or their advisers, to set manually.
A: Lifetime annuity rates are influenced mainly by life expectancy and gilt yields. Although they do not fluctuate as sharply as some financial markets, they do tend to vary gradually over time. EvolveMyRetirement® does not attempt to predict future movements in annuity rates. Instead, it uses current annuity rates in its calculations, and these are updated periodically
If annuity rates change significantly in the future, the results shown by your plan may differ when you next run the strategy generation process because the model will be working with updated information.
A: On the Members page you can enter your State Pension amount and, if it hasn’t started yet, the age at which you expect to receive it. If you’re in the UK, the government provides an online service that shows your State Pension forecast in today’s money, based on your National Insurance record. This can help you decide what figure to enter in your plan.
A: Deferring the State Pension is an option in the UK. If you choose to defer, the amount you eventually receive is increased under rules set by the government. Whether deferral is appropriate depends on personal circumstances, including health, finances, and preferences, and is not something EvolveMyRetirement® can assess.
The UK government provides information explaining how State Pension deferral works and how the increased amount is calculated. This can help you understand the implications and decide what figures to enter in your plan if you want to explore how deferral would affect your results.
If you are considering deferring your State Pension, you may wish to discuss the decision with an independent financial adviser.
A: Sensitivity analysis varies one setting at a time while holding all others constant. This can sometimes reveal a setting that produces a lower simulated risk of insolvency for that single change. However, the strategy generation process evaluates many combinations of settings together, and the utility score reflects all of your plan’s preferences and assumptions, not just insolvency risk.
There are several reasons why a sensitivity analysis might show a lower insolvency risk for an individual setting:
• Discretionary spending: Lowering discretionary spending can affect the utility score in both directions: it reduces lifetime spending, which lowers the utility score, but it also generally reduces the simulated probability of insolvency, which raises the utility score depending on your stated attitude to risk. The strategy generation process evaluates overall utility across several criteria, not insolvency risk in isolation.
• Legacy preferences: If you've indicated a preference for leaving a legacy, some settings may support that preference more effectively, and they may indicate a higher probability of insolvency. For example, settings that involve purchasing a lifetime annuity early on might reduce the estimated probability of insolvency but also reduce the potential for a legacy.
• Monte Carlo variation: Differences in the estimated probability of insolvency may fall within the normal margin of error of Monte Carlo simulation, especially when the differences are small.
Because sensitivity analysis isolates one setting at a time, while the strategy generation process evaluates many interacting settings together, it is normal for the two tools to highlight different aspects of the model’s behaviour.
A: Sensitivity analysis runs a fresh Monte Carlo simulation for each alternative setting it tests. Because Monte Carlo simulation has a natural margin of error, the utility score for any single run is an estimate rather than an exact value. When two settings produce utility scores that are close to each other, small differences in the simulation can cause one to appear higher than the other.
This means that a setting highlighted by sensitivity analysis may differ from the one selected by the strategy generation process simply because of normal simulation variation. In such cases, the underlying utility scores are effectively very similar within the model’s margin of error.
A: The probability of insolvency and the median legacy are only two of the factors used to calculate the utility score. The strategy generation process evaluates utility across several criteria simultaneously. Changes that improve some metrics may reduce others. Factors that can influence the utility score include:
• The likelihood of ending with negative net worth: This is distinct from insolvency and can affect the utility score even when the insolvency probability seems low.
• The distribution of legacy outcomes. The median legacy is only one point in the distribution. Changes in the spread or shape of the distribution can also affect the utility score.
• Legacy preferences: If you have indicated a preference for leaving a legacy, the utility score reflects not only the size of legacies in simulated scenarios but also how consistently they appear.
• Risk aversion: Your stated attitude to risk affects how the utility function weighs uncertainty. Two settings with similar median outcomes may differ in how widely results vary, which can influence the utility score.
Because utility incorporates several dimensions of uncertainty, it's possible for a setting to seem favourable on one or two individual metrics while still producing a lower overall utility score within the model.
A: The assumptions (shown at the bottom of the Plan page) represent factors you cannot control but can estimate. EvolveMyRetirement® provides a set of default assumptions that are intended to be reasonable and cautiously framed. Different economists or financial advisers may hold different views, which is why the assumptions are configurable.
Whatever assumptions you choose, the strategy generation process takes them into account when evaluating utility. It is normal for a resulting strategy to be highly sensitive to some assumptions but not others, because certain assumptions have a stronger influence on long-term outcomes within the model.
If you are unsure about the assumption settings, you may wish to share your plan with an independent financial adviser, who can review them and help you understand how they relate to your circumstances.
A: Although it may seem counter-intuitive, this can happen with certain strategy settings related to buying new lifetime annuities (annuitisation). For example, your strategy may include a rule to purchase annuities if, at some point in the future, they cover at least a specified percentage of your spending. If that condition is met, part of your net worth is converted into annuities, which reduces the potential legacy in the simulation.
If you increase discretionary spending from the start, the annuitisation condition may be less likely to be met. In that case, the simulation may show fewer annuity purchases, meaning more of your net worth remains unconverted, which can lead to a higher median legacy.
However, avoiding annuity purchases by increasing discretionary spending also tends to increase the simulated probability of running out of money. The model reflects both effects: the change in legacy outcomes and the change in insolvency risk.
A: This can happen because of the strategy settings related to buying new lifetime annuities. In any given year, if the conditions specified in your strategy are met, the model may convert some pension funds or other assets into new annuities. This conversion reduces net worth at the point of purchase because annuities are treated as income-producing rather than as part of net worth.
Although net worth falls at that moment, the resulting annuity income can reduce the rate at which net worth erodes in later years. The sudden drop you see in the survival analysis reflects this conversion rather than any unplanned spending.
A: For a plan with two members, the Survival Analysis always models net worth on the assumption that both partners survive throughout the projection, regardless of their individual life expectancies. Because the life insurance payout only occurs on death, it never appears in the Survival Analysis view.
You should, however, see life insurance payouts in:
The Average Scenario, where the model includes the payout in years in which death occurs in the underlying simulations.
Random Scenarios, in any scenario where you die before your spouse.
Monte Carlo simulation results, whenever the simulated timing of death triggers the payout.
The absence of the payout in the Survival Analysis is therefore expected behaviour: that view is specifically designed to show outcomes assuming both partners remain alive.
A: You can view bar charts on the Results page once you’ve entered your information and generated a strategy. Many other tools place bar charts at the centre of their analysis and use them to suggest whether you are “on track”. This can be misleading because a single projected path cannot capture the uncertainty inherent in long-term financial planning.
EvolveMyRetirement® uses Monte Carlo simulation, running thousands of scenarios to estimate the likelihood of different outcomes. The Results page shows the overall simulation results as pie charts, which summarise the probabilities of key outcomes.
You can also view individual scenarios as bar charts, including randomly generated scenarios and a scenario based on average values. These charts are intended as illustrations of how the future might unfold in specific simulations, not as predictions to rely on in isolation.
A: This can happen because of the strategy settings related to buying new lifetime annuities. In any given year, if the conditions specified in your strategy are met, the model may convert some pension funds or other assets into new annuities. This conversion reduces net worth at the point of purchase because annuities are treated as income-producing rather than as part of net worth.
Although net worth falls at that moment, the resulting annuity income can reduce the rate at which net worth erodes in later years. The sudden drop you see in the scenario reflects this conversion rather than any unplanned one-off spending.
A: The value you enter for your ISAs represents their total value today. Entering zero simply means you currently hold no ISA assets. It does not prevent the model from allocating future investments into ISAs.
If your strategy and liquidity allow it, the model will make use of ISA allowances in future years, because ISAs are treated as tax-advantaged compared with taxable investments. When sufficient funds are available, the simulation will allocate as much as possible to ISAs up to the annual allowance, before allocating to taxable accounts.
As a result, even if your starting ISA value is zero, your ISA balance may grow over time in simulated scenarios.
A: Deferred Withdrawal appears when the required spending for the current year exceeds the income available in that same year. The model records the shortfall as a liability, which is then settled at the start of the following year by withdrawing from disposable assets, supplemented by borrowing if both required and permitted by the plan.
It's quite normal for scenarios to show Deferred Withdrawal. For example, tax liabilities are typically settled in the following year, which naturally creates a temporary shortfall. Deferred Withdrawal only leads to insolvency in the simulation if the model cannot settle the liability when it becomes due.
A: The charts and tables use two slightly different year conventions, depending on what they are showing:
• Assets & liabilities start at the current year.
The first year represents your position today.
The next year represents your position one year from today, and so on.
• Income & outgoings start with the year after the current year.
The first year represents the year ending one year from today.
The next year represents the year ending two years from today, and so on.
This difference reflects the fact that assets and liabilities are shown at a point in time, whereas income and outgoings relate to flows over the following year.
A: Investment income in the model is defined as a fixed percentage of the mean growth rate for the chosen investment risk level. The growth rates in the assumptions represent the total return with income reinvested. To separate that return into “income” and “growth”, the model uses the following rule:
• Lowest-risk investments: income is assumed to be 100% of the mean growth rate (all income, no capital growth).
• Highest-risk investments: income is assumed to be 0% (all capital growth, no income).
• Intermediate risk levels: the income percentage is linearly interpolated between 0% and 100%.
Because this percentage depends only on the selected risk level, it remains deterministic within a scenario. It will only change if your strategy triggers rebalancing to a different investment risk level. If no rebalancing occurs, the income percentage stays constant throughout the scenario.
The net investment growth percentage, however, is randomised. The “growth” figure shown in a scenario is the net growth percentage minus the income percentage, which is why the growth component varies while the income percentage does not.
A: Internally, the program performs all projections using absolute monetary values, not inflation-adjusted values. Each projection year is calculated in today’s pounds, and then inflation is applied at the end of the year according to your assumptions. Inflation adjustment is used only for reporting, to show values in “today’s money”.
Inflation adjustment in reports reverses the effect of inflation on amounts that grow with inflation and reduces the real value of amounts that do not.
If you run a scenario (in our simple example) without adjusting for inflation, you'll see the discretionary spending grow each year. But if you run it adjusting for inflation, discretionary spending will remain constant.
A simple example:
Assume inflation is fixed at 5% per year. Suppose discretionary spending is £1,000 per year and is set to increase with inflation. Internally, the program will project:
• Year 1: £1,000
• Year 2: £1,050
• Year 3: £1,102.50
• …and so on.
These are nominal values: the actual amounts the model uses in its calculations. But the inflation-adjusted values are:
• Year 1: £1,000
• Year 2: £1,000
• Year 3: £1,000
• …and so on.
How this appears in scenarios
• Without inflation adjustment: You'll see discretionary spending rise each year, because the nominal values are shown.
• With inflation adjustment: You'll see discretionary spending remain constant, because the nominal increases are removed.
The same principle applies to other items
For example, suppose property growth is assumed to be 0% while inflation is 5%. Internally, the property value remains £100,000 every year. In a scenario:
• Without inflation adjustment: The property value appears constant.
• With inflation adjustment: The property value appears to fall each year, because a property with 0% nominal growth loses 5% of its value in real terms each year.
A: When a scenario is shown with inflation adjustment, the values in future years are converted into “today’s money”. This means the program reduces each future year’s nominal value by the cumulative inflation that has occurred since the starting point. If you try to reconcile the change in a fund’s value from one year to the next using only the listed increases and decreases, the numbers will not balance unless you also include this adjustment. The “Inflationary Erosion” row shows the amount removed to express the value in real (inflation-adjusted) terms.
For example, suppose a taxable investment fund has no additions, withdrawals, growth, or reinvested income from one year to the next, and inflation is 5%. If the nominal value is £100,000 in one year, then the inflation-adjusted value for the following year will be:
£100,000 ÷ 1.05 = £95,238
The difference of £4,762 is the reduction required to express the fund’s value in today’s money. That reduction is what the table labels as Inflationary Erosion.
A: The CSV export includes three reconciliation sections. Each one checks that the detailed year-on-year changes add up to the overall movement shown in the scenario:
• Assets Reconciliation. This verifies that the recorded increases and decreases in each type of asset correspond to the actual change in total asset values.
• Liabilities Reconciliation. This checks that the recorded increases and decreases in each type of liability match the overall change in total liability balances.
• Net Worth Reconciliation. This confirms that the recorded income and outgoings for all plan members correspond to the actual change in total net worth.
Each section ends with a reconciliation line that, in theory, would consist entirely of zeroes. However, these calculations are performed in your browser using rounded values rather than the full-precision values used internally on the server. As a result, the reconciliation lines may show small positive or negative numbers. These should typically be very close to zero – usually single-digit amounts.
If you ever see unusually large reconciliation differences, please contact support so we can investigate.
The purpose of these reconciliation sections is to give you confidence that the scenario’s figures are internally consistent and that the projected changes align with the underlying calculations.
A: We prefer to give you a way to explore the tool at your own pace, without any time limits. That’s why the Basic tier acts as a permanent free alternative to a traditional trial. You can build a plan, explore the core functionality, and return to it whenever you like – with no expiry.
Here’s what the Basic tier includes:
• Free and permanent access
• Create your plan and run two full strategy generations at no cost
• Option to buy extra strategy-generation credits without subscribing
• Full flexibility to enter assets, liabilities, income, and expenses
• Suitable for individuals and couples
• Preview Premium features such as Sensitivity Analysis and Survival Analysis
• View sample reports before upgrading
• Edit your plan and regenerate your strategy whenever your circumstances change
If you later decide you want more – such as multiple plans, advanced analysis features, or REST API access – you can upgrade to Standard or Premium. Subscriptions include bonus credits and do not auto-renew. You’ll receive a reminder before expiry, so you stay in control.
A: No. Any user can choose a Premium subscription to access the additional features. Premium is not designed specifically for financial advisers. Most features are intended for anyone who wants deeper analysis or more flexibility.
That said, financial advisers are likely to find many of the Premium features essential, particularly the ability to work with multiple plans and to collaborate with clients who share their plans with them.
A: Premium provides a range of additional capabilities designed for users who want more flexibility, more analysis tools, or the ability to work with multiple plans. Key advantages include:
• Create unlimited plans rather than being limited to a single plan – particularly useful for financial advisers working with many clients.
• Full access to plans shared with you, including editing and generating strategies.
• Collaborate on plans you create for others by transferring ownership while retaining shared access (unless the new owner chooses otherwise).
• Duplicate any plan and modify each copy independently, making it easy to explore alternative settings or what-if scenarios.
• Optionally lock specific strategy settings before generating a strategy, allowing you to hold the values fixed.
• Run unlimited Sensitivity Analyses on any plan you can access.
• Run unlimited Survival Analyses, showing survival probabilities and the statistical spread of projected net worth.
• Customise your default assumptions for new plans and propagate updated assumptions to existing plans based on your previous defaults.
• Receive extra strategy-generation credits as part of your subscription.
• Prioritised access to our fastest servers, helping ensure consistently quick access to results that require Monte Carlo simulation.
• Use the REST API for integration with other systems, programs, or spreadsheets.
A: Once you’ve signed up and created an account, you’ll see an option on the Plan page to upgrade to Premium. Selecting this will take you through the upgrade process and unlock the Premium features for your account.
A: No. Paid subscriptions do not renew automatically. You’ll receive an email reminder roughly one month before your subscription expires, and you can choose to renew at any time, either before expiry or afterwards. Each renewal gives you a full additional year of access.
A: No. All strategy generation and simulation work runs on our servers, not on your computer. Even if your device is slow, the speed of generating strategies is unaffected.
A: No. All strategy generation and simulation work runs on our servers, not on your device. The only part that might be slightly slower on a slow connection is simply receiving the generated strategy back in your browser once the server has finished processing it. But the amount of data transferred is small, so in practice it should feel no different from loading any other webpage.
A: Whenever your plan has changed, the Results page must run a fresh Monte Carlo Simulation with 10,000 trials to reflect the updated inputs. This simulation is performed on our servers, not on your device, and it is the main reason for the delay you see.
If you have a Premium subscription, one of our fastest servers is kept on standby for you. This means Monte Carlo processing typically completes more quickly. These high-performance servers are the same ones used for strategy generation for all users, and they also speed up Sensitivity Analysis, each run of which may require up to four additional Monte Carlo Simulations.
For non-Premium users, only strategy generation is guaranteed to run on our fastest servers. Monte Carlo simulations triggered by visiting the Results page run on the fastest server currently available, which may not always be one of the top-tier machines.
A: It helps to consider what the system is actually doing behind the scenes.
Running a single Monte Carlo Simulation typically takes only a few seconds. But strategy generation doesn’t involve running just one simulation. As the Genetic Algorithm explores different combinations of strategy settings, it must evaluate thousands of candidate strategies: each one requires its own Monte Carlo Simulation.
At first glance, this might sound inefficient. And if we had implemented the Genetic Algorithm in a straightforward, naïve way, it really would take hours to complete. That would make strategy generation impractical.
However, during the development of EvolveMyRetirement®, we incorporated a proprietary technique that dramatically reduces the total processing time, bringing it down from hours to minutes.
If waiting several minutes still feels long, keep in mind that this process is doing something unique among retirement-planning tools: it is exploring a huge space of possible strategy configurations, each tested through full Monte Carlo Simulation, to generate strategies based on your inputs.
A: At the start of strategy generation, the system makes an initial estimate of how many generations the Genetic Algorithm will need. As the process approaches that estimated endpoint, the algorithm checks whether it has reached a utility plateau, meaning that further generations are no longer producing meaningful improvements in the candidate strategies.
If a plateau has not been reached, the system extends the run by adding more generations. This can make it appear as though the process is slowing down, but in reality it's still progressing at the same steady pace. It's simply continuing for longer than originally estimated.
Once a plateau is reached, the strategy generation process draws to a close.
A: For full details on how your information is handled and protected, please refer to our Privacy Policy.
A: No. After a thorough review, we confirmed that EvolveMyRetirement® was not affected by the reported Log4J vulnerability . We continue to monitor our software and infrastructure for this and other potential security issues as part of our ongoing commitment to platform security. We continue to monitor our software for this and other potential vulnerabilities.
A: The 4% Rule is simply a rule of thumb. It is based on the idea of a “safe withdrawal rate” – a level of withdrawals from a retirement fund that is unlikely to deplete the fund during the retiree’s lifetime.
In practice, this is an oversimplification. Different parts of a retirement fund are taxed differently (for example, pension withdrawals are treated as income, ISA withdrawals are tax-free, and sales of other investments may trigger capital gains tax). It also ignores non-financial assets such as property, which may generate rental income, be sold, or be used for equity release. And it assumes that spending always matches withdrawals exactly.
EvolveMyRetirement® uses a different approach based on sustainable discretionary spending. Starting from your essential spending, the system generates a strategy that includes a level of discretionary spending that can be sustained under your stated assumptions. You can choose either a fixed inflation-adjusted level of discretionary spending or a level that varies in response to unfolding financial events.
Actual drawdown decisions depend heavily on tax considerations. Withdrawals from pension plans, ISAs, and taxable investments need to be balanced to help minimise liabilities such as income tax, national insurance, capital gains tax, and inheritance tax, as well as the potential tax liabilities of dependants on inherited pensions. When the program generates a strategy, it takes these tax interactions into account.
A: EvolveMyRetirement® uses two main types of AI: Genetic Algorithms (GAs) and Large Language Models (LLMs).
• A proprietary GA is used for strategy generation.
The GA is a type of evolutionary algorithm that evolves a population of candidate strategies. Each candidate is evaluated using Monte Carlo Simulation, and the interaction between the GA and the simulation engine has been designed to ensure that strategy generation completes within a practical timeframe.
• The AI Assistant is powered by an LLM.
It uses documentation specific to EvolveMyRetirement®, supplemented by the model’s general financial knowledge, to interpret users’ questions and provide clear explanations of how the system works. Its role is to support understanding of the tool, not to give personalised financial advice.
• The AI-generated plan overview also uses the LLM.
Instead of answering individual questions, it produces a narrative report describing your plan, the generated strategy, and the resulting projections. It highlights potential inconsistencies or anomalies in the inputs, helping you understand how the model interprets your data.
A: Not at all. EvolveMyRetirement® does not provide financial advice, and it is not intended to replace regulated advice from a qualified financial adviser. Although the system can generate strategies based on the information you provide, putting any strategy into practice requires expertise, judgement, and discipline. And like any modelling tool, the quality of the output depends on the quality and completeness of the inputs.
An independent financial adviser can help you review your assumptions, understand the implications of your plan, and consider factors that fall outside the scope of the modelling tool. Many users choose to share their plan with an adviser so they can discuss it in the context of their broader financial circumstances.
Learn more about the boundary between the model output and advice.
A:Yes. When you sign up, you can create one financial plan, which may be for either a single person or a couple. If you need to work with multiple clients, you can subscribe as a Premium User, which allows you to create an unlimited number of plans.
You can also collaborate with clients through shared access to their plans, so both you and your clients can view the same information and explore the model’s outputs together.
A: Yes. Premium Users have access to a REST API, along with documentation explaining how to use it.
A: Yes. If you no longer wish to use EvolveMyRetirement® and want all the information you’ve entered to be deleted, you can do this from the Settings page once you’re logged in. If your account includes any unused paid subscription, it will be cancelled automatically when you delete your account. Please note that if you decide to sign up again in the future, you'll need to re-enter your information from scratch.
A: EvolveMyRetirement® is created and owned by Okinawa Zest Limited, an independent company with no ties to financial institutions. The software is developed to model financial plans without favouring any particular types of investment. If you would like further information about the company, you are welcome to contact us.
