How we calculate
The calculators on this site aim to give numbers whose method is clear, not numbers that merely look plausible. This page sets out how we calculate and what we could not confirm. Results are for reference and can differ from the calculation of a real product.
1. Calculation happens only in this browser
What you enter is not sent to a server; it is calculated inside this browser. The calculators do not load outside services or outside code to work. The page itself may carry an analytics or advertising script, and this site is not set up to pass what you enter to it. See Privacy for details.
2. Amounts are calculated exactly
Ordinary computer decimal (floating-point) arithmetic can introduce tiny errors, for example 0.1 + 0.2 comes out as 0.30000000000000004. So the calculators here handle amounts and rates as integers (a smallest unit such as one won or one cent) or as fractions (integer numerator and denominator), and by default round only once, at the digit that is shown. If a calculator rounds at every period, it says so on screen.
A few calculations, however, cannot be written as exact integers or fractions, so they use ordinary floating-point approximations: the Rule of 72 (the exact value uses logarithms), CAGR (roots), the effective-to-nominal rate conversion (roots), the IRR in the NPV/IRR calculator (iterative search), and the annualised return in the investment return calculator (fractional exponent). The retirement simulator uses exact fractions for the saving phase and 34-decimal fixed-point arithmetic (truncated) for the withdrawal phase. These results carry an "approximate value" note, and tiny differences below the displayed digits can occur.
3. The rounding rule is shown on screen
Results change depending on where you round and by which rule. The calculators here use the default rule (round half up) and, where rounding affects the result, either offer a choice of rule or state on screen which rule was used. A real financial institution may round down or use another rule.
There are four rules to choose from: round half up, round down, round up and round half to even (banker's rounding: at exactly 0.5 go to the nearest even digit). Some calculators offer only some of them or fix one.
4. The day-count basis is stated
When interest is calculated by the day, the result depends on the basis: counting actual calendar days, or treating a month as 30 days and a year as 360 days. When a calculator uses dates, it states on screen how it counts days. Institutions can differ, so the actual amount can differ.
For example, the simple interest calculator lets you choose "actual days/365" or "actual days/360" when the term is entered in days; when it is entered in years or months, one month counts as 1/12 of a year. "30/360" can be used only with years and months.
5. Inputs have limits
Very large numbers or very long inputs can freeze the page, so each calculator sets an input limit and explains why when it is exceeded. These limits are this site's own choice and are not limits set by law or by a product.
6. Tests use hand calculation and official examples
Each calculator has tests, and the expected values are hand-checked values or examples printed in official documents. We do not copy the calculator's own output as the answer. Edge cases and error paths such as division by zero, very large values and empty input are tested too.
7. Tax rates, limits and rate benchmarks only from official sources
Values that can change, such as tax rates, deductions, limits and benchmark rates, are used only when we opened the official source ourselves, and we list the source and the date on the Sources page and in each calculator. A value we could not confirm is not used; you enter it yourself.
8. What we could not confirm, and limits
- The terms, preferential conditions, fees and tax treatment of real products differ between institutions, and this site cannot reflect all of them.
- Rates, returns and periods are assumed to stay as you entered them.
- For the limits of our responsibility, see the Disclaimer.