How to Round Numbers: Decimals, Tens, Hundreds and Significant Figures

By HSP · Published October 22, 2026 · 7 min read · Math & Dates

Rounding is one of those skills you use every day without noticing. A price of $19.97 becomes "about twenty dollars", a 2.6-hour drive becomes "around two and a half hours", and a school report turns 87.46% into 87.5%. Rounding makes numbers easier to read, compare and remember. The rule is short, but money, negative numbers, significant figures and spreadsheets catch people out, so here are worked examples for each.

The basic rule in three steps

1. Find the rounding place. Decide where you want to stop: the nearest whole number, one decimal place (tenths), two decimal places (hundredths), the nearest ten, hundred, and so on. Underline that digit.
2. Look one digit to the right. This is the "decider" digit. Ignore everything after it.
3. Round up or keep. If the decider is 5, 6, 7, 8 or 9, add one to the underlined digit. If it is 0, 1, 2, 3 or 4, leave the underlined digit as it is. Then drop every digit after the rounding place (after the decimal point) or replace them with zeros (before it).

Example: round 6.4827 to two decimal places. The hundredths digit is 8, the decider is 2, so the 8 stays and the answer is 6.48. Round the same number to one decimal place and the decider is 8, so 4 becomes 5: the answer is 6.5.

Rounding to tens, hundreds and thousands

The same rule works on the left of the decimal point, but the dropped digits become zeros so the number keeps its size. Take 1,847:

One trap: always look at the original number, not at an answer you have already rounded. Rounding 1,847 to the nearest hundred gives 1,800. If you first rounded to the nearest ten (1,850) and then rounded that to the nearest hundred, you would wrongly get 1,900.

What about a carry, like 9.97?

When the digit you round up is a 9, it becomes 10 and carries over, just like in addition. 9.97 to one decimal place: the tenths digit is 9, the decider is 7, so 9 tenths becomes 10 tenths, which is one whole. The answer is 10.0. Keep the trailing zero when asked for one decimal place: "10.0" shows the value is accurate to a tenth.

Significant figures, explained simply

Scientists and engineers often round to a number of significant figures (sig figs) instead of decimal places. Significant figures count from the first non-zero digit, so the size of the number does not matter. Leading zeros never count.

• 48,263 to 2 significant figures is 48,000 (the decider is 2, so 48 stays and the rest become zeros).
• 0.004567 to 2 significant figures is 0.0046 (the first two significant digits are 4 and 5, the decider is 6, so 5 rounds up to 6).
• 3.14159 to 3 significant figures is 3.14.

Sig figs are handy when a measurement is only so precise: if your scale reads to the nearest gram, writing 312.4867 g pretends to know more than you do. Students meet them in physics and chemistry, and they turn up in exams such as the SAT, GCSE and A-levels; see our study hub and exams hub for free prep pages.

Why does 2.5 sometimes round to 2?

The 5-or-more rule is called "round half up", and it is what most schools teach. But when a number sits exactly halfway, some tools use a different rule called "round half to even", also known as banker's rounding: the halfway value goes to whichever neighbour is even. Under that rule 0.5 becomes 0, 1.5 becomes 2, 2.5 becomes 2 and 3.5 becomes 4.

Why bother? Because always rounding halves up nudges totals upwards. Add 0.5 + 1.5 + 2.5 + 3.5 and the true sum is 8. Round each one half up first (1 + 2 + 3 + 4) and you get 10. Round half to even (0 + 2 + 2 + 4) and you get 8 again. For large sets of data, that fairness matters.

Where you will see the difference: the Python programming language's built-in round() uses round half to even, so round(2.5) gives 2. Excel's ROUND function rounds halves away from zero, so =ROUND(2.5, 0) gives 3 and =ROUND(-1.475, 2) gives -1.48. Python's documentation also warns that round(2.675, 2) gives 2.67, not 2.68, because most decimal fractions cannot be stored exactly in binary floating point. Neither is wrong; they simply follow different rules.

Negative numbers

For negative numbers, think about size first. Under the usual school rule, -2.4 rounds to -2 and -2.6 rounds to -3. For an exact half, our Rounding Calculator rounds away from zero, like Excel, so -2.5 becomes -3 and 2.5 becomes 3.

Rounding money

Most currencies are rounded to two decimal places (cents, pence or paise). Two practical tips:

• Round at the end, not in the middle. Split $10 three ways and each share is $3.3333... Round each share to $3.33 and the three shares add up to $9.99, a cent short. Keep full precision while you calculate and round only the final figures, then decide who covers the leftover cent. Our Percentage Calculator and our guide on how to calculate a tip show this in action.
• Sometimes you must round up, whatever the decider says. If 120 students go on a trip and each bus holds 50, 120 ÷ 50 = 2.4, which normally rounds to 2, but two buses leave 20 students behind. You need 3. The same goes for tins of paint or boxes of tiles. Planning a bigger trip? Our flights hub compares fares.

Rounding vs truncating

Truncating simply chops digits off without looking at the decider. 7.89 truncated to one decimal place is 7.8; rounded, it is 7.9. Some systems truncate, which is why results can look one digit "off".

Quick practice

Try these, then check them in the Rounding Calculator: 12.348 to one decimal place (12.3), 0.0719 to 2 significant figures (0.072), 4,550 to the nearest hundred (4,600; pick 2 significant figures in the calculator) and 99.95 to one decimal place (100.0, which the calculator shows as 100). Working with fractions first? Convert them with our guide on how to convert a fraction to a decimal or the Fraction Calculator, and for averages that come out with long decimals, use the Average Calculator.

VSH take: rounding feels like the easiest part of maths, yet most real mistakes we see come from rounding too early, rounding twice, or not knowing that a spreadsheet and a programming language handle halves differently. Write down the full answer, round once at the end, and say how you rounded (for example "to 2 decimal places"). That small habit makes your numbers trustworthy, whether it is homework, a bill split or a lab report.

Rounding rules are standard arithmetic; behaviour of Python's round() and Excel's ROUND function is taken from their official documentation as of October 2026. Worked examples are our own. ViralScopeHub is not affiliated with Microsoft or the Python Software Foundation.