MROUND in Excel: Round to the Nearest Multiple and Recognize Midpoint Limits

• Introduced: Original introduction release not established by the official history reviewed.
• Availability: official help lists Excel for Microsoft 365, Excel for Microsoft 365 for Mac, Excel 2024, Excel 2024 for Mac, Excel 2021, Excel 2021 for Mac, Excel 2019, Excel 2016.
• Checked: October 14, 2026; official MROUND reference. Examples are independently calculated expectations, not native Excel test results.

=MROUND(number, multiple)

Both arguments are required. The second is a step, not a count of decimal places. Microsoft documents #NUM! for opposite nonzero signs and an away-from-zero midpoint rule, alongside a specific warning that midpoint direction is undefined when the multiple is decimal. Official MROUND reference and known limitations

This guide uses clear integer steps for midpoint demonstrations and treats decimal midpoints separately. Do not assume a nearest-multiple function is suitable for a policy requiring a minimum quantity or a guaranteed upper bound.

Enter 31, 32, 33, 37, 38, 40 in A2:A7. In B2 enter and fill down:

=MROUND(A2,5)

Expected B2:B7: 30, 30, 35, 35, 40, 40. For input 32, distances to 30 and 35 are two and three. For 33, those distances are three and two. An exact multiple such as 40 remains unchanged.

Because the function may round down, 31 becomes 30. If a stock order must cover all 31 required units, nearest rounding is the wrong business rule even though the calculation is correct.

=MROUND(37.5,5)
=MROUND(-37.5,-5)

Expected results: 40 and −40. Each input is halfway between adjacent five-unit multiples. The documented midpoint rule moves away from zero in these integer-step examples.

Changing the input sign without changing the step sign is not equivalent:

=MROUND(-37.5,5)

Expected result: #NUM!, because the two nonzero arguments have opposite signs. A positive-only test set would not expose that requirement.

For numeric inputs in A2 and a fixed five-unit step, a readable approach is:

=IF(A2=0,0,IF(A2>0,MROUND(A2,5),MROUND(A2,-5)))

Expected outputs for −38, −32, 0, 32, 38 are −40, −30, 0, 30, 40. The zero branch is an explicit workflow choice; no inference about every zero-sign edge case is required.

IF makes the step sign visible. Keep the source column numeric and separately validate imported text rather than treating this branch as a complete input-cleaning routine.

Microsoft specifically documents inconsistent direction at decimal midpoints: its examples show MROUND(6.05,0.1) producing 6.0 and MROUND(7.05,0.1) producing 7.1. Those are vendor-documented examples, not native results observed for this guide.

Do not promise that every midpoint with a 0.1 step rounds upward. The issue matters when a formal rounding policy depends on exact decimal ties. Simply displaying two decimals does not resolve it, and this guide does not offer a universal floating-point workaround.

If an application stores an exact integer unit, such as a fictional count of 73 or 78 ticks, integer-step rounding is easier to audit: =MROUND(73,5) gives 75, while =MROUND(78,5) gives 80. Choose a justified unit before importing and calculating; converting already-imprecise values later is not guaranteed to recover lost accuracy.

For 12, 12, 12 in D2:D4, apply =MROUND(D2,5) to each row. The three results are 10, 10, 10, totaling 30.

Now compare:

=MROUND(SUM(D2:D4),5)

Expected result: 35, because the unrounded total is 36. Rounding each component and rounding the aggregate are different operations. Write the required stage of rounding into the workbook specification rather than assuming these methods agree.

=MROUND(10.6,0.25)
=MROUND(10.7,0.25)

Expected results: 10.5 and 10.75. These examples avoid an exact midpoint. For 10.6, the distances to 10.5 and 10.75 are 0.1 and 0.15; for 10.7 they are 0.2 and 0.05.

Retain the original values so the rounded report can be checked against actual measurements. ROUND is appropriate when the policy concerns decimal places rather than multiples; functions that always choose a lower or upper multiple serve different purposes.

Test an exact multiple, values on both sides of a midpoint, a signed pair and the actual step used by the workbook. Check a sum both before and after rounding if aggregation matters. Document any domain restriction, especially for decimal midpoint inputs. These checks establish the intended calculation without implying native coverage of every floating-point boundary.

Verification: Fictional inputs and independently calculated outputs; no native Excel execution. See the editorial testing standards.