Rounding Calculator
Round any number to a chosen place value (from millions to millionths) in seven rounding modes, with the rule applied shown.
Details
Rounded value
3.14
To the nearest hundredths
This rounds a number to whatever place you need: a decimal place, a whole number, or the nearest ten, hundred or thousand.
It offers several rounding modes, including the banker's rounding that spreadsheets and financial systems use, and shows how each one treats your number differently.
How rounding works
Rounding replaces a number with a shorter one that is close enough for the job. 3.14159 becomes 3.14, and 4,832 becomes 5,000.
The standard rule is the one taught in school: look at the digit to the right of where you are rounding. Five or more rounds up, four or less rounds down.
That rule is not the only one, and the differences matter. Money, negative numbers and large datasets all have their own conventions, which is why this calculator offers several modes rather than assuming one.
Only the single digit immediately to the right decides. The 59 that follows it is irrelevant, which is why 3.1449 rounds to 3.14 rather than 3.15.
What to enter
- Number to round
- Any number, positive or negative.
- Round to the nearest
- A decimal place, a whole number, or the nearest 10, 100 or 1,000.
- Rounding mode
- Half-up is the school rule. The others handle negatives and ties differently, and are explained below.
The rounding modes, and where each is used
- Half up
- The school rule. Exactly 0.5 goes up, away from zero. 2.5 → 3, and −2.5 → −3.
- Round down (toward zero)
- Chops off the extra digits. 2.9 → 2 and −2.9 → −2. Also called truncation.
- Floor (toward −∞)
- Always goes down the number line. 2.9 → 2 but −2.1 → −3. Not the same as rounding down once negatives are involved.
- Ceiling (toward +∞)
- Always goes up the number line. 2.1 → 3 and −2.9 → −2. Used where you need enough of something, like buying materials.
- Banker's rounding
- Ties go to the nearest even number: 2.5 → 2 but 3.5 → 4. The default in spreadsheets and most financial systems.
What this assumes
Rounding is applied once, to the original number. Rounding repeatedly compounds the error.
How to calculate a rounded number
Find the place you are rounding to, then look at the single digit immediately to its right.
- the next digit
- Only the one immediately to the right. Everything after it is ignored
Find your place. Rounding to two decimals means keeping two digits after the point.
Look one digit to the right. That single digit decides everything.
Round and drop the rest. Five or more takes the kept digit up by one. Four or less leaves it alone.
Check the mode for ties and negatives. An exact 0.5, or a negative number, behaves differently depending on which mode you are using.
See a worked example: why 3.1449 rounds to 3.14, not 3.15
- Number
- 3.1449
- Round to
- 2 decimal places
Keeping two decimals means keeping 3.14.
The digit immediately to the right is 4, so it rounds down.
The 49 after it does not matter, even though 3.1449 is very close to 3.145.
Rounding twice would give the wrong answer: 3.1449 → 3.145 → 3.15. Always round once, from the original.
3.14
Frequently asked questions
Under the school rule, yes. But financial systems and spreadsheets commonly use banker's rounding, where an exact tie goes to the nearest even number instead: 2.5 → 2, while 3.5 → 4.
The reason is bias. Always rounding ties up nudges long columns of figures slightly upward. Sending half the ties down cancels that out, which matters when you are adding thousands of rows.
Nothing at all for positive numbers. They differ entirely for negatives.
Rounding down means toward zero, so −2.9 becomes −2. Floor means down the number line, so −2.9 becomes −3. Spreadsheet functions distinguish these, and mixing them up on negative values is a common source of quiet errors.
Whenever a partial unit is no use to you. Needing 4.2 tins of paint means buying 5, not 4.
It is the right mode for materials, containers, and anything you cannot buy a fraction of.
Because each figure is rounded on its own, and the small errors accumulate. Three values of 3.33 round to 3.3 each, totalling 9.9 rather than 10.
Financial reports handle this by rounding the total separately and noting that components may not sum, which is why you see that footnote so often.
Same rule, different place. For the nearest 10, look at the ones digit: 47 → 50, 42 → 40. For the nearest 100, look at the tens digit: 1,480 → 1,500.
Rounding to the nearest thousand or higher is how large figures get reported, and it is worth remembering that '5,000 people' may be anything from 4,500 to 5,499.
Problems people actually run into
Rounding twice and shifting the answer
Rounding 3.1449 to three decimals gives 3.145, and rounding that to two gives 3.15. Rounding straight from the original gives 3.14, which is correct.
Always round once, from the full-precision figure. This bites hardest in spreadsheets where a rounded cell feeds another calculation, so the error travels quietly through the workbook.
Rounding money before the final step
Rounding each line of an invoice to the cent before adding them up gives a different total than adding first and rounding once, and on long invoices the difference is visible.
Keep full precision through the calculation and round only the figure you present. It is also what accounting standards generally expect.
Results are estimates for general information only and are not professional financial, medical, or legal advice. Read our full disclaimer.
Last updated: September 4, 2026