Online calculators built on JavaScript use 64-bit double-precision floating-point arithmetic, which provides roughly 15 to 17 significant decimal digits. In practice that means you can trust the answer up to about the 15th digit. For science, engineering, and everyday finances, that precision is more than sufficient — but it is not infinite, and a few edge cases reveal the limit.
What "15 to 17 significant digits" means in practice
Significant digits are the meaningful digits in a number, counted from the first non-zero digit. A calculator with 15-digit precision can handle a value like 123,456,789,012.345 accurately — that is 15 digits total. Add a 16th or 17th digit and the result may drift by one unit in the last place.
This is defined by the IEEE 754 double-precision standard, which uses 64 bits: 1 for the sign, 11 for the exponent, and 52 for the fractional part (the mantissa). Those 52 bits give approximately log₁₀(2⁵²) ≈ 15.65 decimal digits of precision. JavaScript's Number.EPSILON, the smallest difference the format can represent, is about 2.22 × 10⁻¹⁶ — a number smaller than one part in a quadrillion.
What this calculator does with that precision
The calculator above stores every intermediate value in full double precision throughout the calculation. When the final result is ready, it rounds to 10 decimal places before displaying it. That display rounding is intentional: it removes the tiny floating-point artefacts (like a trailing ...00000000004) that would otherwise appear in results such as 0.1 + 0.2. The underlying 15-17 digit precision is used for the computation; the 10-decimal display is just for readability.
When 15 digits is more than enough — and when it isn't
- School and university homework — physics, chemistry and engineering problems rarely need more than 6-8 significant figures. Double precision has roughly twice that to spare.
- Financial calculations — currency values are typically 2 decimal places; even a calculation involving trillions of dollars only needs about 15 digits total, which fits comfortably.
- Scientific constants — the speed of light, Avogadro's number and similar values are themselves only known to about 10-12 significant figures from experiment, so the calculator is not the limiting factor.
- Chained calculations with many small decimals — rounding errors can accumulate across long chains. The result is usually still correct to 10+ digits, but for mission-critical computation, a specialist tool with arbitrary precision is better.
- Cryptography and number theory — these fields often work with integers hundreds of digits long. A standard double-precision calculator cannot represent them exactly and is the wrong tool entirely.
Check the limit: type999999999999999 + 1in the calculator above. You get1000000000000000— exact, because both numbers fit within 15 significant digits. Try9999999999999999 + 1(16 nines) and the result may not increment as expected, because you have crossed the boundary of what double precision represents exactly.