How Whatify Money calculates scenarios

What these tools do

Whatify Money helps you compare choices. You enter a few facts about your money. We apply a set of rules and show what could change.

Most tools do budget math with fixed inputs. When You Stop also tests many possible futures. None of the tools can tell you what will happen or choose the right plan for you.

How the budget tools work

These tools do not draw random market returns. If you keep the inputs and the rules the same, the result stays the same.

In the debt tools, we use the entered yearly rate divided by 12 for the monthly rate. Real accounts may use daily interest, fees, or other rules. Check the terms for your account.

A simple runway example

Suppose you have $12,000 in cash, spend $3,000 a month, and still bring in $1,000 a month. The monthly gap is $2,000. Your cash would cover six months at that pace.

This is a made-up example, not a forecast. If income covers spending, this simple model has no monthly draw on reserves. That does not mean your cash will last forever in real life.

Use the tools to test an income drop, a surprise bill, or debt versus savings. You can also check housing costs and monthly cash flow.

How When You Stop tests retirement

In an engine-backed run, we test 1,000 possible futures for your household. Each future has yearly market returns, price changes, and a lifespan. This is called a Monte Carlo simulation: repeat a model with different random draws to see a range of results.

We compare stopping work sooner, at your planned date, and later. The paths share the same draws. That helps keep the comparison fair: one stop date does not get lucky just because it drew a different set of markets.

A path passes if the plan can pay the full spending target in every year the household is alive. A displayed success rate counts paths that pass. It is a result inside the model, not a measured chance that your real plan will work.

Each stop date's own rate uses futures where someone lives to that date. A direct comparison uses the same group of futures that reach the latest stop date. Because those groups can differ, subtracting the headline rates may not match the tool's comparison.

The main rules in an engine-backed run

The yearly spread above is called standard deviation. It sets how widely the random draws vary around the average. It is not a best-case or worst-case limit. Stock and bond averages are before the model's investment fee.

These are simple yearly averages after inflation, not compound growth rates. The new assumptions record marks them as provisional. It records the existing settings, not a proven forecast or a new change to expected returns.

The engine now supports linked stock and bond draws within a year. It can also draw blocks of past years to keep some links across years. The current app has not switched to those choices. Its market draws still use the separate, fresh yearly draws shown above. Engine features and app settings are not the same thing.

The model also includes rules for couples, survivor benefits, and required withdrawals from tax-deferred accounts. For those required withdrawals, it uses the older partner's age for the household's combined pretax balance. That can differ from the rules for separate real accounts.

The lifespan data come from the Social Security Administration's 2022 period life table. A period table uses death rates from one period; it does not forecast future gains in lifespan. Tax settings use stored rules, including IRS guidance for 2026. We do not fetch new tax rules during each run.

A demo or preview can use sample results. Those samples are not evidence about your own plan. Check that you are using a live, engine-backed result before relying on a comparison.

What the dollar amounts mean

Today's dollars show buying power at today's prices. Future dollars show the number of dollars in that later year. The same buying power can cost more dollars as prices rise. BLS explains the difference.

In the retirement year table, balances use today's dollars. Cash flows such as spending, benefits, and taxes use that year's dollars. Check the labels before comparing amounts. We round numbers for display, so totals may differ slightly from the sum of the shown values.

Why the market-history guide is a separate test

Our guide to three ways of modeling markets uses a simpler test plan. It compares past years kept in order, past years picked at random, and the bell-shaped market model used by When You Stop. All three use the same savings, spending, investment mix, and time span. All leave out taxes, fees, benefits, health costs, and changes in lifespan.

The third test takes its market settings from the app's current code. It removes the product's 0.4% yearly fee to match the two other fee-free tests. This is a test of the market model, not a full product run. Do not read these results as the rates for your own stop dates. The guide includes inputs, exact counts, a chart, sources, and downloads so you can check the example.

The current model, like the random-year test, does not keep links between one year and the next. Rolling windows keep the actual order, but cover only past events. Comparing the three shows how much the model choice changes this example; a higher success rate does not prove a better model.

The current stock average is 5%, versus 8.61% in the guide's stored past sample. The bond average is 2%, slightly above that sample's 1.83%. So it is not correct to say both return averages are lower. Spreads differ too.

In a fresh extra check, we changed history to match the model's averages and spreads, then drew blocks averaging five years. That case passed 78.09% of 10,000 paths, near the current model's 77.50%. Changing only average returns gave 74.21%. Other path methods still differed. The guide reports all the checks, including rolling windows. The changed history is made-up test data, not actual returns or proof that the models are equal.

Past returns need not describe the future. That is a reason to test other return assumptions, but it does not prove our current settings are the right forecast. The benchmark checks results under those settings; it does not establish their basis as forecasts. Their use in a future-facing model needs clear source evidence and tests of other plausible values.

How we check and update the methods

We test budget math with known inputs and expected answers. Engine tests also check repeat runs with the same seed, spending shortfalls, and tax and benefit rules. These checks can catch code errors. They do not prove that a model predicts the future.

The updated engine adds separate checks of financial math against reference results. We keep those checks apart from tests of market assumptions. A correct tax or cash-flow calculation does not show that a chosen future return is right.

Read how we check taxes, health costs, and retirement income for the benchmark sources, fresh results, and yearly examples. That separate guide also distinguishes tested engine features from those enabled in the current app.

The linked market-history guide is one benchmark: a fixed test plan that we can rerun. Its downloads show the measured counts, engine version, data version, and inputs. The underlying code is private, so this is not an independent outside audit.

Our recorded cFIREsim comparison checks the engine's rolling-history path against a past result from that tool, with a set tolerance. The two data sets cover different years. This check does not test the random-past-year mode or the market settings used by When You Stop. Passing it is not proof that all engine modes are accurate forecasts.

Budget tools use monthly steps; the retirement model uses yearly steps and whole-year ages. A monthly cash-flow average is not a day-by-day cash forecast. Dates on these pages name the day of a content update, not a claim that source data are from that day.

When a method, source, or important claim changes, we update the page date and content version. We rerun affected examples, check the numbers, and ask for human review before release. Small spelling or layout fixes do not imply new research. Sources and data years are named beside the rules they support.

Version 0.1 was a page placeholder. Version 0.2 added this explanation and the two-method pilot. Version 0.3 added the app's current market model. Version 0.4 used engine 0.11.0 and added fresh checks of matched return assumptions. Version 1.0 is the reviewed publication edition, with links to both guides. No independent professional review is claimed.

How to use the results

• Start with a budget that matches your life. Check amounts and time units.
• Change one input at a time to see what drives the result.
• Test higher costs or lower income, not just your best case.
• Treat a small difference between options with care. It may depend on the inputs or the draws.
• Review major choices with a qualified adviser who can see your full finances.

No model includes every risk. The SEC's guide to performance claims explains why a back-test or hypothetical result can leave out costs and hide limits. These tools offer estimates for learning, not financial, tax, or legal advice.