What Is an Average?
In ordinary speech "the average" means one thing: add everything up and divide by how many. In statistics it is a family of measures, and the arithmetic mean is only its most familiar member. The NIST engineering handbook files them together as measures of location — competing answers to the single question of where a data set sits.
Mean, Median and Mode Are All Called Averages
The mean is the total divided by the count. The median is the middle value once the list is sorted. The mode is the value that appears most often. Three different numbers, three different questions, one shared English word — which is why a quoted "average" is ambiguous until somebody says which one it is.
This page computes the mean. It is the one meant in almost every school exercise, the one behind a batting average and a grade average, and the one the AVERAGE function computes in Excel, Google Sheets and LibreOffice Calc.
The Mean Is the Balance Point of the Data
The mean has a property none of the others share: the distances above it exactly cancel the distances below it. Take those five scores again, whose mean is 78.2. The gaps from the mean are −6.2, +6.8, +11.8, −12.2 and −0.2, and they total zero. That is not a coincidence of these particular numbers — it follows from the definition, because dividing the sum by the count produces precisely the value that leaves nothing left over on either side.
It also explains why the mean moves whenever any single value moves. Every number in the list is holding one end of the see-saw, so pushing one of them further out has to be answered by the mean sliding toward it. The median carries no such obligation; it only cares about position in the sorted order, and that difference is the root of everything else on this page.
The Mean Says Nothing About Spread
Five students who each score 70 return Average: 70.00 | Sum: 350.00 | Count: 5. Five students scoring 40, 55, 70, 85 and 100 return exactly the same three lines. One class is uniform and one is scattered across sixty marks, and the mean cannot tell them apart, because summing deliberately discards the arrangement of the values it sums.
A mean reported without any measure of spread beside it is a true number that can still mislead. The companion figure is the standard deviation, which measures how far a typical value falls from the mean, and it has its own tool — the Standard Deviation Calculator.
What Is the Average Formula?
One line, and it has not needed revising since it was first written down: add the values, divide by how many there are.
The Average Formula
Average (mean) = Sum of all values ÷ Number of values
Order is irrelevant — addition does not care — and so is the sign of the values. The only requirement is that you divide by every value you added, including any zeros.
What Each Part Means
Three quantities, all three printed by the calculator. The right-hand column uses the five test scores 72, 85, 90, 66 and 78.
| Term | Meaning | Value for 72, 85, 90, 66, 78 |
|---|---|---|
| Σx — the sum | every value added together | 391 |
| n — the count | how many values were read | 5 |
| x̄ — the mean | Σx ÷ n | 78.2 |
Entered as a list, those five scores return Average: 78.20 | Sum: 391.00 | Count: 5. The mean and the sum are printed to two decimal places; the count is always a whole number.
The Same Formula in Sigma Notation
Textbooks write the addition with a capital sigma, which is shorthand for "add these up as i runs from 1 to n". The bar over the x marks it as a mean rather than a single reading.
x̄ = (x₁ + x₂ + … + xₙ) ÷ n x̄ = (1 ÷ n) × Σ xᵢ for i = 1 to n
Nothing new is being said in the second line. The notation exists so that the formula can be written once for a list of any length, instead of once for every length.
How Do You Calculate an Average?
By hand it is three operations. Below they are set out in the order the calculator performs them, followed by a worked set, the rules the input box actually enforces, and two shortcuts worth knowing.
Calculating an Average Step by Step
- Write the values out. Their order makes no difference to the answer.
- Add them all together. That total is the sum.
- Count how many values you added. Zeros are values and have to be counted.
- Divide the sum by the count. That quotient is the arithmetic mean.
- Round once, at the end. Rounding the sum before you divide moves the error into the answer.
In the calculator all of it happens on one press. There is no starting list on this page, so the result panel stays empty until you press Calculate, and it does not re-run while you type — edit a value and press the button again.
Worked Example: Five Test Scores
A student sits five papers and scores 72, 85, 90, 66 and 78. Nothing is weighted and no paper is dropped, so a plain mean is the right measure.
Sum = 72 + 85 + 90 + 66 + 78 = 391 Count = 5 Mean = 391 ÷ 5 = 78.2
Typed into the box as 72, 85, 90, 66, 78, the panel returns Average: 78.20 | Sum: 391.00 | Count: 5. The trailing zeros are formatting rather than extra precision: the answer is exactly 78.2, a shade above the middle score of 78.
What the Input Box Accepts
The list is split on commas and on nothing else. That single rule accounts for every surprise the field can produce, so it is worth seeing the behavior laid out rather than discovering it mid-homework.
| What you type | What the calculator does | What comes back |
|---|---|---|
| 72, 85, 90, 66, 78 | reads five values | Average: 78.20 | Sum: 391.00 | Count: 5 |
| 1.5, 2.25, 3.75 | decimals are fine | Average: 2.50 | Sum: 7.50 | Count: 3 |
| -3, -1, 0, 2, 7 | negatives and zero are fine | Average: 1.00 | Sum: 5.00 | Count: 5 |
| 10, apple, 20 | drops the word, keeps the numbers | Average: 15.00 | Sum: 30.00 | Count: 2 |
| 1 2 3 4 5 | no commas, so it sees one unreadable item | No valid numbers entered |
| 1,000, 2,000 | the thousands commas split the numbers apart | Average: 0.75 | Sum: 3.00 | Count: 4 |
Two working rules follow from that table. Separate values with commas, never with spaces or semicolons — both come back reading No valid numbers entered, as do currency symbols, so $10, $20, $30 fails outright. And strip thousands separators before pasting: 1,000, 2,000 is read as the four values 1, 0, 2 and 0, which average 0.75 instead of 1,500. A column pasted straight out of a spreadsheet carries no commas either, so the field sees the whole column as a single item: either that same notice or one run-together number with Count: 1.
Averaging in Your Head: The Base-Number Shortcut
Adding five three-digit numbers mentally is hard. Adding five small ones is not. Pick a round base near the middle of the data, total up how far each value sits from it, divide that total by the count, and add the result back onto the base.
Base = 80 Gaps = −8, +5, +10, −14, −2 → total −9 −9 ÷ 5 = −1.8 Mean = 80 − 1.8 = 78.2
Same answer, arithmetic small enough to do while walking. The method is exact rather than approximate: subtracting a constant from every value subtracts exactly that constant from the mean, so adding it back at the end restores the true figure.
Adding One More Value Without Starting Over
If you already know the mean of n values and one more arrives, there is no need to re-add the list. The new mean is the old mean plus the gap between the new value and that old mean, shared over the new count.
New mean = old mean + (new value − old mean) ÷ (n + 1) 78.2 + (95 − 78.2) ÷ 6 = 78.2 + 2.8 = 81.0
Entering 72, 85, 90, 66, 78, 95 confirms it: Average: 81.00 | Sum: 486.00 | Count: 6. A single 95 lifted the mean by 2.8 marks, and it would have lifted it by less again on a longer list — the divisor grows every time.
Rearranged, the same identity answers the question students actually ask: what do I need on the last paper? Multiply the target by the number of papers and subtract the marks already banked. An 80 average across six papers needs 80 × 6 − 391 = 89. A 90 average would need 149, which a paper marked out of 100 cannot deliver — a useful thing to learn before revising for it.
Average Examples From Real Data
The arithmetic is identical whatever the numbers describe. What changes is what the answer is allowed to mean, and each of these three sets carries its own trap.
Six Months of Revenue
Monthly figures of 12,500, 13,800, 11,200, 15,600, 14,100 and 13,000 return Average: 13366.67 | Sum: 80200.00 | Count: 6. Here the mean carries a second, more useful meaning: it is the flat monthly figure that would have produced the same half-year total, so it doubles as a run rate.
That makes every month directly comparable against it. The weakest month, at 11,200, ran 2,166.67 below the mean; the strongest, at 15,600, ran 2,233.33 above it. Three of the six months fall below the mean and three above, which is the reassuring pattern — a mean with five months below it and one far above would be reporting an outlier, not a typical month.
Rainfall, Where the Zeros Have to Be Counted
Seven days of rainfall in millimeters — 0, 0, 12, 4, 0, 31, 8 — return Average: 7.86 | Sum: 55.00 | Count: 7. Delete the three dry days and the identical 55 mm spreads over four days instead, for a mean of 13.75. That figure is nearly double and it is wrong, because a dry day is a measurement, not a missing one.
The same slip is easy to make in a spreadsheet without noticing. Microsoft documents AVERAGE as ignoring empty cells while counting cells that contain zero, so a column with the dry days left blank quietly reports the wet-day mean. Type the zeros.
Temperatures Below Zero
Five morning readings of -3, -1, 0, 2 and 7 °C return Average: 1.00 | Sum: 5.00 | Count: 5. Negative values need no special handling; they simply pull the sum down, and a mean below zero is a perfectly ordinary answer.
The minus sign itself is the thing to watch. Typed as a plain hyphen it is read as a sign. Pasted from a word processor as an en dash or a typographic minus sign it is not a number at all, and those values are dropped in silence — the same five readings pasted that way come back as Average: 3.00 | Sum: 9.00 | Count: 3. The answer looks plausible, which is exactly what makes it dangerous, and the Count line is what gives it away.
Average Chart: The Kinds of Average Compared
Six different "averages" of one small data set — 2, 3, 3, 4, 8 — showing how far apart they can land before anyone has made a mistake.
Six Averages, One Data Set
| Average | How it is found | Value for 2, 3, 3, 4, 8 |
|---|---|---|
| Arithmetic mean | sum ÷ count = 20 ÷ 5 | 4.00 |
| Median | middle value once sorted | 3 |
| Mode | the value appearing most often | 3 |
| Midrange | (smallest + largest) ÷ 2 = (2 + 8) ÷ 2 | 5 |
| Geometric mean | 5th root of the product 576 | 3.5652 |
| Harmonic mean | 5 ÷ (sum of the reciprocals) | 3.2432 |
The arithmetic, geometric and harmonic means are the classical Pythagorean means — midrange is not one of them — and their order is fixed: harmonic ≤ geometric ≤ arithmetic for any set of positive numbers, with all three equal only when every value in the set is identical. This calculator computes the arithmetic one, and entering 2, 3, 3, 4, 8 returns Average: 4.00 | Sum: 20.00 | Count: 5.
Mean vs Median vs Mode
Swap that 8 for an 88 and the three school-level averages scatter. The set 2, 3, 3, 4, 88 returns Average: 20.00 | Sum: 100.00 | Count: 5.
| Measure | Value | What it tells you |
|---|---|---|
| Mean | 20 | sum ÷ count — hauled upward by the 88 |
| Median | 3 | middle value — the 88 counts once, like every other value |
| Mode | 3 | the most frequent value — completely unchanged |
Four of the five numbers now sit below their own mean. That imbalance is the signature of skew, and it is the reason house prices, household incomes and website response times are published as medians rather than means.
Which Average for Which Data
| Your data | Use | Because |
|---|---|---|
| Symmetric, no extremes — test scores, heights | Mean | it uses every value, so nothing is thrown away |
| Skewed or containing outliers — incomes, house prices | Median | it ranks rather than sums, so one huge value moves it by one place |
| Categories rather than quantities — sizes stocked, options chosen | Mode | it is the only one that still works when the values are not numbers |
| Values carrying different weights — course credits, portfolio holdings | Weighted mean | a plain mean treats a one-credit class like a five-credit one |
| Rates across a fixed distance or quantity — speeds, price per unit | Harmonic mean | the mean of the rates is not the rate of the totals |
| Growth across periods — investment returns, compounding rates | Geometric mean | growth multiplies rather than adds |
Worked Sets for Reference
Every row here came out of this calculator, and each one can be checked with a single addition and a single division.
| Data set | Sum | Count | Average |
|---|---|---|---|
| 72, 85, 90, 66, 78 | 391.00 | 5 | 78.20 |
| 88, 92, 79, 95, 84 | 438.00 | 5 | 87.60 |
| 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 | 75.00 | 10 | 7.50 |
| 12500, 13800, 11200, 15600, 14100, 13000 | 80200.00 | 6 | 13366.67 |
| 0, 0, 12, 4, 0, 31, 8 | 55.00 | 7 | 7.86 |
| 4, 3, 3.7, 2.3 | 13.00 | 4 | 3.25 |
| -3, -1, 0, 2, 7 | 5.00 | 5 | 1.00 |
| 10, 20 | 30.00 | 2 | 15.00 |
| 1, 2, 2 | 5.00 | 3 | 1.67 |
The bottom row is the one to remember. Five shared between three is 1.666…, and the panel prints 1.67 — rounded as it is displayed, not before the division, which is why the figure never drifts however long the list gets.
How to Read Your Result
Three lines come back from every successful run, and each one is doing a job.
What the Three Lines Mean
| Line | What it is | Why it is shown |
|---|---|---|
| Average | the arithmetic mean, to two decimal places | the answer |
| Sum | every value added together | the figure a marker or a spreadsheet audit asks for |
| Count | how many values were actually read | proof the calculator read the list you meant |
Taken together they form a closed check: multiply the Average by the Count and you should land back on the Sum. For the five test scores, 78.20 × 5 = 391.00 exactly. Expect the last penny to disagree whenever the mean does not terminate: the six months of revenue further up give 13366.67 × 6 = 80,200.02 against a Sum of 80,200.00, because the printed mean is rounded to two places while the true value, 13,366.666…, is not.
Use the Count as Your Error Check
The Count is the line that catches paste errors, because it reports how many values the calculator read rather than how many you believe you typed. An empty slot between two commas is treated as a zero rather than skipped: 10,,20 returns Average: 10.00 | Sum: 30.00 | Count: 3, and a list ending in a stray comma — 10, 20, 30, — returns Average: 15.00 | Sum: 60.00 | Count: 4 instead of the 20 that those three numbers genuinely average.
So read the bottom line first. If the Count matches the number of values you intended, the Average above it is right. If the Count is one too high, hunt for a trailing comma or a double comma. If it is too low, something in the list was not readable as a number and was dropped — an en dash, a currency symbol, a stray letter.
Is the Average Any Good?
A mean answers one question — what is the typical value — and refuses every other. Two things decide whether it deserves to be quoted at all. The first is how many values sit behind it: a mean of two readings and a mean of two hundred are printed to the same two decimal places and carry wildly different authority.
The second is how far apart those values are. A mean of 70 drawn from a class that scored between 68 and 72 describes every student in the room. The same 70 drawn from a class that scored between 40 and 100 describes nobody in it. Nothing in the number itself distinguishes the two cases, which is why a mean quoted without a range or a standard deviation beside it is an incomplete report rather than a wrong one.
One thing the mean is genuinely good for is comparison over time — this month's mean against last month's. To convert that pair into the percentage change between them, use the Percentage Increase & Decrease Calculator.
Limits: When the Mean Misleads
The arithmetic never fails. The interpretation does, in five recognizable ways — and in every one of them the calculator is returning the correct answer to a question that was not the one being asked.
One Extreme Value Rewrites the Answer
Four graduate salaries of 30,000, 32,000, 35,000 and 38,000 alongside a fifth of 1,200,000 return Average: 267000.00 | Sum: 1335000.00 | Count: 5. Nobody in that group earns anything close to 267,000, and all four of the ordinary earners take home less than a seventh of it. The median, 35,000, describes the cohort accurately.
This is not a defect in the mean. It is the mean keeping its promise — every value gets an equal vote — in a set where one value is more than thirty times the largest of the others. The moment a single figure can do that, the question to ask is not "what is the average" but "what is the middle".
You Cannot Average Two Averages
A class of 10 students averages 70; a class of 30 averages 90. The average across all 40 students is not 80. Add the totals instead — 10 × 70 = 700 and 30 × 90 = 2,700 — and divide by the head count: 3,400 ÷ 40 = 85. Treating the two class means as two equal values throws away the fact that three times as many students sat in the second room.
A mean of means is only correct when the groups are the same size, which they rarely are. When they are not, each group mean has to be weighted by its own size before averaging, which is the job of the Weighted Average Calculator.
Rates Need the Harmonic Mean
Drive 60 miles at 30 mph, then a further 60 miles at 60 mph. The mean of 30 and 60 is 45, but the journey took 2 hours plus 1 hour to cover 120 miles, so the average speed was 40 mph. The arithmetic mean of two speeds is only correct when equal time is spent at each; across equal distances the harmonic mean is the right measure, and 2 ÷ (1/30 + 1/60) gives exactly 40.
The same trap is waiting in price per unit, fuel economy and typing speed. The test is simple: if the quantity is already a ratio of two things, averaging the ratios is not the same as taking the ratio of the totals.
Growth Needs the Geometric Mean
An investment gains 50% in one year and loses 40% in the next. The arithmetic mean of +50% and −40% is +5% a year, which sounds like a gain. Follow the money instead: 1,000 becomes 1,500, then 900. That is a 10% loss across the two years, or −5.13% a year compounded.
Growth rates multiply rather than add, so their correct average is the geometric mean — here the square root of 1.5 × 0.6 — and it is always at or below the arithmetic one. Quoting the arithmetic average of a series of returns overstates performance every time the returns vary at all.
The Mean May Not Be a Possible Value
A mean does not have to appear in the data, and it does not have to be achievable. The set 1, 2, 2 returns Average: 1.67 | Sum: 5.00 | Count: 3, and 1.67 is not a possible count of anything countable. That is not an error to be rounded away: the mean of counts is a rate — five items shared between three — rather than a value any member of the set holds. Round it for display if you must, but never round it before using it in further arithmetic.