Weighted Average Calculator

What a weighted average does

A weighted average gives each number its own say. Instead of adding the values and dividing by how many there are, you multiply each value by a weight, add those products, and divide by the total of the weights. One formula covers grades, prices, survey results, and portfolio returns.

Weighted average = sum of (value × weight) / sum of the weights. When the weights are percentages that add to 100, the denominator is 100 and the answer is just the sum of the products divided by 100. When the weights are counts, such as credit hours or pounds, the denominator is whatever those counts add up to.

The plain average is the special case where every weight is 1. That is why the two numbers often land close together and sometimes do not: a weighted average moves toward whichever values carry the most weight, while a plain average treats a 1 credit course and a 4 credit course as equals.

One property is a free error check. A weighted average can never fall outside the range from the smallest value to the largest. If your result is 92 when your inputs run from 70 to 88, a weight or a decimal point has slipped.

A course grade, step by step

Here is a course with four graded parts. The weights add to 100 percent, so the products add up to a number that can be read directly as the final grade.

The weighted average is 8,530 / 100 = 85.3. The plain average of the same four scores is (85 + 90 + 78 + 88) / 4 = 85.25, so the weighting moved the grade up by 0.05 of a point. The final exam carried the most weight and it was above the plain average, which is why the weighted number came out slightly higher.

Change one weight and the result moves more than most people expect. If the final exam were worth 50 percent and the homework 5 percent, the products become 425 + 1,800 + 1,950 + 4,400 = 8,575, and the grade is 85.75. Same four scores, a different syllabus, a different letter grade in a district that draws the line at 85.5.

When the weights are counts, not percentages

Weights do not have to be percentages. Anything that measures how much each value matters works, as long as the units are consistent. Credit hours are the classic case: a course that meets more often counts more.

Take a 92 in a 4 credit course and an 80 in a 1 credit course. The weighted average is (92 × 4 + 80 × 1) / 5 = 448 / 5 = 89.6. The plain average would be (92 + 80) / 2 = 86, which understates the record because it treats the one credit course as half the degree.

The coffee row is worth reading twice. Three pounds at 2.00 cost 6.00 and one pound at 6.00 costs 6.00, so 12.00 buys 4 pounds at 3.00 a pound. The plain average of the two prices is 4.00, which is not a price anyone paid and not a price that buys the whole 4 pounds. Weighting by pounds is the only version that matches the receipt.

The survey row shows the same idea with people instead of money: 0.6 × 9,000 = 5,400 and 0.4 × 1,000 = 400, so 5,800 of 10,000 people said yes, which is 58 percent. A plain average of the two town percentages would give a big town and a small town equal say.

Weights that do not add up to 100

Weights that do not add to 100 are the most common source of wrong answers. The fix is one line: divide by the sum of the weights you used, not by 100.

With two items scored 90 and 75 at weights of 2 and 1, the sum is 90 × 2 + 75 × 1 = 255, and 255 / 3 = 85. The same pair written as 66.67 percent and 33.33 percent gives 0.6667 × 90 = 60.003 and 0.3333 × 75 = 24.998, which add to 85.0. Both routes agree because both divide by the true total weight.

The wrong version, dividing by 100 while the weights total 3, produces 2.55 and looks absurd. A subtler wrong version appears when one graded item is still missing: homework 85 at 20 percent and a midterm 78 at 25 percent are counted, but the final is not graded yet. The points earned are 1,700 + 1,950 = 3,650, and the average over graded work so far is 3,650 / 45 = 81.11. Dividing by 100 instead would report 36.5, which is why the tool above divides by the total it used.

Keeping the denominator visible is also how you avoid a second trap: a weight of zero. A row with weight 0 contributes nothing to the numerator and nothing to the denominator, so it drops out of the calculation entirely. That is correct behavior, and it is why an ungraded assignment should carry a weight of 0 rather than a score of 0.

How much do you still need on the rest?

The same formula answers the question students ask most: what do I need on the final? Rearranged, the score you need on the remaining weight is (target × 100 - points already earned) / weight remaining.

Suppose you have 82 percent on work worth 70 percent of the course. Points earned are 0.70 × 82 = 57.4, and 30 percent is still ungraded. To finish at 85 you need (85 × 100 - 5,740) / 30 = 2,760 / 30 = 92.00. To finish at 90 you need (90 × 100 - 5,740) / 30 = 3,260 / 30 = 108.67, which no score can deliver. That is the arithmetic behind "the final cannot save this grade", and it is better to know it with six weeks left than on the last day.

The last row shows why the timing matters: the same 82 average needs only 88 on the remaining half, but 92 when only 30 percent is left. The tool above returns this number whenever the weights are entered as percentages and do not yet reach 100.

Weighted means in the real world

Weighted averages are how large numbers get summarized without pretending every observation is equal.

Prices. The Consumer Price Index does not average prices. It averages price changes, and it weights each change by how much of the household budget that category represents. The BLS Handbook of Methods describes the aggregation weights as the expenditure share of each item, so a jump in rent moves the index far more than a jump in the price of salt.

Money. A portfolio with 6,000 earning 8 percent and 2,000 earning 2 percent returned (480 + 40) / 8,000 = 6.5 percent on the whole balance, not the 5 percent a plain average of the two returns suggests. The weights are dollars, because dollars are what earned the return.

Measurements. A lab that averages repeated readings with different uncertainties weights the more precise readings more heavily. The NIST DataPlot reference for the weighted mean states the standard form and notes that frequency counts are a common choice of weight.

The pattern is that the weight should be the thing that makes the values comparable. Dollars for money, people for surveys, precision for measurements. Pick the weight that matches the question, and the average answers it.

Averaging rates needs different weights

Rates are the exception that trips people up. Averaging speeds with equal weights is only correct when each speed ran for the same amount of time, because time is the denominator hidden inside a speed.

Drive 1 mile at 30 miles per hour and 1 mile at 60 miles per hour. The naive average is 45 miles per hour, and it is wrong. The first mile takes 1 / 30 = 0.0333 hours and the second takes 1 / 60 = 0.0167 hours, so 2 miles take 0.05 hours and the true average speed is 2 / 0.05 = 40 miles per hour. The slow mile took twice as long, so it deserves twice the weight, and the correct weighted average of the speeds is (30 × 0.0333 + 60 × 0.0167) / 0.05 = 40.02.

Now drive for 1 hour at 30 and 1 hour at 60. Distances are 30 and 60, total 90 miles in 2 hours, and the average is 45 miles per hour. Same two speeds, different weighting, different answer. Both are right for their own question.

This is the harmonic mean in disguise, and it applies to any rate with the same numerator: miles per gallon, price per unit, words per minute. When the top number is fixed and you average the rate, weight by the bottom number. The average calculator handles the equal weight case, and this page is for the unequal one.

Five habits that prevent wrong answers

Five habits prevent almost every wrong weighted average.

1. Add the weights first. If they are meant to total 100 and they total 95, you are either missing an item or one weight is mistyped. The tool above prints the total for that reason.
2. Do not mix percent weights with count weights. A 20 percent homework weight and a 4 credit course weight cannot sit in the same calculation unless you convert one of them.
3. Weight by the right variable. Weighting course grades by the number of assignments instead of credits gives the wrong answer, and weighting speeds by distance instead of time gives the 45 instead of 40.
4. Keep zero weight rows out, and do not score them as zero. A missing assignment with weight 0 is ignored; the same assignment scored 0 with weight 20 drags the average down as if it had been failed.
5. Check the range. The result must sit between the smallest and largest value in the list. Anything outside that range means an arithmetic slip.

For letter grades with AP or Honors bonus points, the weighted GPA calculator applies the school-specific bonus scale and converts letters to grade points. This page takes plain numbers and any weights you like, which is what you want for costs, survey rates, returns, or a syllabus that uses points instead of letters.