Rounding Calculator

Round any number to a chosen decimal place, significant figure or place value (nearest ten, hundred, thousand). Choose between standard half-up rounding, banker's rounding and truncation. The tool also shows the effect of each rounding rule so you can pick the right one for your context.

Enter a number to round

Any integer or decimal value, positive or negative.
Number of decimal places (0 = nearest integer), significant figures, or a place value power of 10.
Rounding rule

Enter a number, choose a mode, and press Round.

How rounding works

Rounding replaces a number with a nearby value that has fewer significant digits. Which "nearby value" you choose depends on the rule:

Half-up

Round 0.5 up. This is the rule most people learn at school: 3.45 → 3.5, 3.44 → 3.4. It is simple but introduces a slight upward bias when applied to many numbers.

Half-even (banker's rounding)

Round 0.5 to the nearest even digit. Used in statistics and finance to avoid upward bias: 3.45 → 3.4 (rounds down because 4 is even), 3.55 → 3.6 (rounds up because 6 would be even). Over many values the errors cancel out.

Floor, ceiling and truncate

  • Floor — always round toward negative infinity (−2.1 → −3).
  • Ceiling — always round toward positive infinity (−2.1 → −2).
  • Truncate — drop the extra digits without rounding (−2.9 → −2).

The three rounding modes

ModeWhat you enterExample (3.14159)
Decimal places How many digits to keep after the decimal point 2 dp → 3.14
Significant figures Total number of meaningful digits 3 sf → 3.14
Nearest place value The unit to round to (10, 100, 1000, 0.1, 0.01 …) nearest 0.01 → 3.14

Worked examples

Example 1 — Round to 2 decimal places

Number: 3.14159, mode: decimal places, precision: 2.

The digit after the second decimal is 1 (less than 5), so the second decimal stays: 3.14.

Result: 3.14

Example 2 — Round to 3 significant figures

Number: 0.00345678, mode: significant figures, precision: 3.

The leading zeros are not significant. The first three significant digits are 3, 4, 5, and the next is 6 (≥ 5), so 5 rounds up to 6: 0.00346.

Result: 0.00346

Example 3 — Round to nearest 100

Number: 1234, mode: place value, precision: 2 (10² = 100).

The tens and units are 3 and 4 → 34 < 50, so round down: 1200.

Result: 1200

Example 4 — Banker's rounding at exactly 0.5

Number: 2.5, precision: 0, rule: half-even.

The value is exactly halfway between 2 and 3. The even neighbour is 2, so half-even rounds to 2. Half-up would give 3.

Result: 2 (half-even) or 3 (half-up)

Common mistakes

  • Rounding twice. 3.4449 rounded to 1 dp is 3.4, not 3.5. If you round to 2 dp first (3.44) and then to 1 dp (3.4), you avoided a compounding error — but if the intermediate step had been 3.45, chaining would produce 3.5 instead of the correct 3.4. Always round once, at the end.
  • Confusing decimal places with significant figures. 3.14 has 2 decimal places but 3 significant figures. Both are correct — they just answer different questions.
  • Ignoring trailing zeros. When rounding 1234 to the nearest hundred, the result 1200 has two significant trailing zeros. Those zeros are important — dropping them changes the value to 12.
  • Assuming all rounding is half-up. Financial software often uses banker's rounding to avoid bias; computer arithmetic may use truncation. Check which rule is expected before publishing a rounded number.

Frequently asked questions

Half-up rounds 0.5 up (2.5 → 3). Half-even rounds 0.5 to the nearest even digit (2.5 → 2, 3.5 → 4), avoiding bias when many values are summed.

Decimal places count digits after the decimal point only. Significant figures count all meaningful digits, starting from the first non-zero digit. 3.14 has 2 decimal places and 3 significant figures.

Use truncation when you need to guarantee the result is never larger than the original (for example, "how many full boxes fit"), or when matching a specific technical standard that specifies truncation.

Choose the "Nearest place value" mode and enter 3 (since 10³ = 1000). For nearest hundred, enter 2. For nearest ten, enter 1.