Rounding Calculator

Written byOleh Soloninupdated 5 October 2026

Number

Round to

Method

Result

Result3.5
How it was rounded
Look at the next digit — 7: 5 or more, so round the tenths digit up and drop the digits to its right: 3.47 ≈ 3.5
Rounding error0.03
To other places
to hundreds 0 to tens 0 to whole numbers 3 to tenths 3.5 to hundredths 3.47

Standard, bankers, up, down and truncate

Rounds to any place by five different rules, including round-half-to-even, and reports the error the rounding introduced.

The school rule is simple: five or more goes up, less goes down. The problem shows up over a long column of figures, where every exact half tipping the same way biases the total upward.

Accounting and statistics therefore use bankers rounding, where halves go to the nearest even digit. 2.5 becomes 2 and 3.5 becomes 4, and where the data are evenly spread the bias disappears.

How to use

Enter the number to round.
Choose the place and the rule.
Compare the rules on a value ending in five: standard rounding and round-half-to-even differ only there, while rounding up, down and truncation differ on other values too.

The rules only disagree at a half

The school rule and round-half-to-even agree until the discarded part is exactly one half: both take 3.47 to 3.5 at one decimal, while rounding down gives 3.4. But 2.5 to a whole number is 3 under the school rule and 2 under banker's rounding, which goes to the nearest even.

Banker's rounding is not pedantry. Over many calculations the school rule drifts upward, because it always sends a half up; going to the nearest even sends halves up or down by the parity of the previous digit, which cancels the bias when the data are evenly spread. It guarantees nothing, though: a hundred values of 2.5 sum to 200 instead of 250.

Rounded to whole numbers, except the last row to one decimal. Worked out with this calculator.
Number	Half up	To even	Down	Up
2.5	3	2	2	3
3.5	4	4	3	4
−2.5	−3	−2	−3	−2
3.47	3.5	3.5	3.4	3.5

What this calculator shows

The result under your chosen rule and the error it introduced
Values to the nearest whole number, ten and hundred for quick comparison
The difference standard and bankers rounding make on your figures

What to keep in mind

A negative number of places rounds to the left of the point: −2 means to the nearest hundred.
"Always up" and "always down" follow the number line rather than magnitude; truncation has its own mode.

FAQs

What is bankers rounding?

Halves go to the nearest even number. It removes the systematic upward bias when totalling many rounded values, which is why it is standard in financial reporting and in IEEE 754.

Why doesn't 0.1 + 0.2 equal 0.3?

Neither has an exact binary representation, so the sum lands on 0.30000000000000004. Rounding to a sensible number of places is the usual fix.

How many decimals should money use?

As many as the currency and the rules you follow require — two for most currencies, none for the yen — and only on the final result unless those rules say otherwise. Rounding intermediate values accumulates error that becomes visible in long calculations.

How does truncation differ from rounding down?

Not at all for positive numbers. For negatives it does: −2.7 rounds down to −3 but truncates to −2.

How does 3.47 round to one decimal place?

To 3.5. The next digit is a seven, which is above five, so the tenths digit goes up by one.

Why is 2.5 sometimes rounded down to 2?

That is banker's rounding, to the nearest even. Accounting and the IEEE 754 standard use it so that a long series of roundings does not drift steadily upward.

Worked example

An exact half

Input: 2.5 to a whole number

Output: Standard 3, bankers 2

Note: The two rules only differ on exact halves. 2.6 rounds to 3 either way, and 2.4 rounds to 2 either way.

Useful links

All tools in Math & Statistics

Rounding Calculator

Result

Related tools

Scientific Notation Converter

Average Calculator (Mean, Median, Mode)

Percentage Calculator

Useful links