How Weighted Averages Work

If every number in a list counted equally, the plain average would be the only average you ever need. In practice, almost nothing counts equally: exams outweigh quizzes, big portfolios outweigh small ones, large samples outweigh small ones. The weighted average is the fix — the same familiar "add up and divide" rhythm, but with each value multiplied by how much it matters first.

When the plain mean lies

Say a course is graded from a midterm worth 30% and a final worth 70%. You score 90 on the midterm and 60 on the final. The plain mean says 75 — a comfortable pass. The weighted truth: 0.3 × 90 + 0.7 × 60 = 27 + 42 = 69. The plain mean answered "if both exams counted the same," which is not the question the syllabus asked. Whenever importance is uneven and you ignore it, you are averaging the wrong dataset.

The direction of the distortion is worth memorising: the weighted average always slides toward the heavy weights. Score highest on the heavily weighted item and the weighted average beats the plain mean; score worst there and it drops below. If an average looks better than the data deserves, check what was weighted.

The formula, and the shortcut when weights sum to 1

weighted average = Σ(value × weight) / Σ(weight)

Two steps: multiply each value by its weight and add the products up, then divide by the total weight. With values 80, 90 and weights 2, 1, that is (160 + 90) / 3 = 83.3 — not the 85 a plain mean would give, because 80 counts double.

There is a convenient special case. When the weights already sum to 1 — decimal percentages totalling 100%, like 0.3 and 0.7 — the division vanishes and the weighted sum is the answer. That is why grade schemes are stated in percentages that add to 100: one multiplication per item. But the division exists precisely so you don't need that tidy sum. Weights of 2, 1, 1 and 50, 25, 25 and 0.4, 0.2, 0.2 all produce identical results; only the ratios between weights matter.

GPA: the classic weighted average

Your grade point average is a weighted average whose weights are credit hours. Take A (4.0) in a 3-credit course, B (3.0) in a 4-credit course, and A− (3.7) in a 3-credit course:

GPA = (3×4.0 + 4×3.0 + 3×3.7) / (3 + 4 + 3)
   = (12.0 + 12.0 + 11.1) / 10 = 35.1 / 10 = 3.51

The 4-credit B drags harder than either 3-credit course because it carries 4/10 of the total weight. The plain mean of the three grades, 3.57, quietly assumes all courses are the same size. To experiment with letter grades and credit hours directly, the dedicated GPA calculator handles the grade-point tables for you.

Averaging averages: the trap of unequal group sizes

The most common weighted-average mistake in the wild is averaging rates from groups of different sizes. One machine produces 90% good parts on a run of 10, another 60% on a run of 1,000. Averaging the percentages gives 75%; the real combined rate is (0.9×10 + 0.6×1000) / 1010 ≈ 60.3% — the big run dominates. The same trap hides inside "average page load time" and any dashboard that averages percentages without recording the denominators. Rule of thumb: when combining rates, weight by group size.

Portfolio returns behave the same way. A 20% position returning 10% and an 80% position returning 2% did not earn you (10 + 2)/2 = 6%; they earned 0.2 × 10 + 0.8 × 2 = 3.6%. Money weighted by position size, results weighted by sample size, grades weighted by credit hours — one formula covers all of them.

What weights can and cannot be

Weights must be non-negative, and at least one must be positive — a weight of zero removes an item entirely, and negative weights would let the "average" escape the range of the data, which is nonsense. Beyond that, anything goes: decimal fractions, credit hours, headcounts, raw importance scores. One caution from descriptive statistics: a weighted average is still a mean, so it inherits the mean's weaknesses — a single extreme value still drags it, and it says nothing about spread. For when the mean is the wrong summary altogether, see the mean, median and mode guide.

Try it yourself

Paste your values and matching weights into the Weighted Average Calculator — it returns the weighted average, the total weight, and a per-item breakdown showing each value's share of the result, with the plain mean alongside so you can see how much the weighting changed the answer. If your values are rates or percentages from different-sized groups, make each weight that group's size and the tool does the rest.