How to average while excluding outliers (TRIMMEAN)
✓ Verified in LibreOffice 25.8.7.3A mean that one crazy value can't wreck — response times, judge scores, sensor data.
The formula
| App | Formula | Notes |
|---|---|---|
| Excel | =TRIMMEAN(A2:A11,0.2) | 0.2 = trim 20% of the COUNT total, split between both ends (10% top, 10% bottom). |
| Google Sheets | =TRIMMEAN(A2:A11,0.2) | Identical. |
| LibreOffice Calc | =TRIMMEAN(A2:A11,0.2) | Identical. |
How it works
TRIMMEAN sorts the data, discards the specified fraction from BOTH ends, and averages what's left: trimming 40% of five values drops one from each end (the 2 and the 100), leaving a clean mean of 4 — while the plain AVERAGE of the same data is a useless 22.8. It's Olympic-judging math. The fraction counts total values trimmed, always split evenly; rounding means small datasets may trim fewer than you expect. For asymmetric or rule-based exclusions, AVERAGEIFS with explicit bounds is the transparent alternative — and MEDIAN remains the maximally outlier-proof option.
Verified, not just documented
We ran =TRIMMEAN(A2:A6,0.4) in LibreOffice 25.8.7.3 (headless, with forced recalculation) and it returned 4 — exactly the expected result. Every formula here is confirmed by actually executing it.