Sum Calculator

What does this sum calculator compute?

A sum is the total of a set of numbers added together. Summation adds each value xᵢ once: Sum = x₁ + x₂ + … + xₙ. Long addition is the same total computed column by column with carrying.

The published formula or method is Sum = Σ xᵢ; x̄ = (1/n) Σ xᵢ; Mean = Sum / n. The following references support the stated mathematics or method:

• Margaret Stevens — The Maths Workbook (University of Oxford): Addition of algebraic terms, decimal fractions and equation transformations. Supports the addition model; does not validate CalculatorX implementation or the compensated-summation algorithm and is not an endorsement.

• Arnold Neumaier (1974): Primary compensated-summation and rounding-error source, ZAMM 54, 39–51; not a validation or endorsement of this implementation.

Which inputs and limits should you check?

Sum accepts 1–10,000 values, at most 65,536 characters in the list and 512 per token. Plain decimals allow 48 significant digits and 48 fractional places. Long addition accepts 1–32 non-negative plain decimals with at most 48 aligned columns. Units are not parsed. Preserve exact strings, warnings and numerical flags when citing results.

The worked example uses Sample data set with −66.25 + (−37.5) + 40 + 42 + 47 + 53 + 54 + 59 + 65. Compare it with the Example calculations and the complete result for the same inputs.

How can you verify and cite this result?

CalculatorX is responsible for this page. The mathematical references above explain the model; implementation checks and their scope are documented separately in the math.sum v1.0.3 Evidence and capability contract. Read the displayed verification and production state together with the specification's numerical limits.

Use the clean calculator URL to cite the tool. A shared input link identifies a particular calculation. Content updates, numerical verification, and domain reviews have separate dates and scopes.

What this calculator does

Use it to:

1. Sum a data set — any count of numbers, including decimals and negatives.
2. Long addition — the same total for non-negative numbers, with stacked place-value steps and carries.

A sum is simply the total after addition. You might sum people per floor to get building occupancy, line items on a receipt, or a statistics sample before computing a mean.

How to enter numbers

Accepted layouts (all become the same list):

Do not use a comma as a thousands separator. Write 7890, not 7,890. If you enter 7,890, it is read as 7 + 890 = 897 and the result includes a non-blocking warning.

Sum formula

Sum = sum(i=1)ⁿ xᵢ = x₁ + x₂ + ⋯ + xₙ

The arithmetic mean is the same total divided by the count:

barx = 1/nsum(i=1)ⁿxᵢ

which is also:

Mean = Sum/n

Sum mode shows both the total and the mean so you can move straight into average-style checks without a second tool. When the mean is a repeating decimal it is labelled with ≈ (for example 256.25 / 9 ≈ 28.4722).

Long addition

Long addition is addition by place value:

1. Stack the numbers and align columns (ones under ones, tenths under tenths).
2. Add each column from right to left.
3. Write the ones digit of the column total; carry any tens to the next column on the left.

Whole-number example: 937 + 129 + 2248

    937
    129
+  2248
  -----
   3314

• Ones: 7 + 9 + 8 = 24 → write 4, carry 2
• Tens: 2 + 3 + 2 + 4 = 11 → write 1, carry 1
• Hundreds: 1 + 9 + 1 + 2 = 13 → write 3, carry 1
• Thousands: 1 → write 1

Decimal example: 37.5 + 129.44 + 48.123

Align decimal points (pad with zeros if needed), then add the same way:

   37.500
  129.440
+  48.123
  -------
  215.063

How the engine adds

• Plain decimals (including mixed signs) are added in base-10, so 0.1 + 0.2 = 0.3 exactly — not the IEEE-754 sequential total 0.30000000000000004.
• Scientific notation (1e3, 2e-12) uses Neumaier compensated summation, which is more stable than naive left-to-right addition when large and small terms cancel.
• Long addition is the schoolbook place-value algorithm with carrying. It matches the exact decimal total for non-negative addends.

sumStr and display are authoritative exact strings for plain decimals and Long addition. Numeric sum, value, mean, values and chart are binary64 conveniences. sum_numeric_approximate and mean_numeric_approximate describe numeric representation; mean_rational_exact (compatibility alias mean_exact) describes an exact source, not an exact JSON number. Terminating rational means display in full; repeating means round half away from zero to four decimal places. Long mode exposes only a convenience numeric mean and no exact-mean claim.

A scientific token selects the floating model for the whole list. Cancellation can amplify input/conversion errors, even with compensation; no universal ULP guarantee or measurement uncertainty model is provided. Scientific totals use shortest binary64 round-trip text, and rounded means use ≈. chart[] contains every numeric prefix (exact-decimal prefixes converted to binary64, or compensated binary64 prefixes); it is machine data, not a displayed interactive chart.

Typing computes locally. Sharing a URL sends its inputs to the server and may put them in browser history, analytics or infrastructure logs. REST/MCP receive submitted inputs on the server. Recent calculations may persist in local browser history and can be cleared. See the privacy policy for service handling; fixed retention periods, deletion guarantees and processing regions are not asserted here. Shared links execute the current engine, not a preserved historical runtime. Links are limited to 8,000 encoded bytes; larger local lists keep complete JSON/CSV exports but disable link copying and history replay. Reload does not retain that local list.

Worked sum example

Numbers: -66.25, -37.5, 40, 42, 47, 53, 54, 59, 65

−66.25 + (−37.5) + 40 + 42 + 47 + 53 + 54 + 59 + 65 = 256.25

Count = 9 · mean ≈ 28.4722.

Agent / API notes

• Capability engine sum · page /calc/math/add · REST /api/v1/calc/sum (alias /api/v1/calc/add) · calculation_version 1.0.3
• Share URL: ?mode=sum|long&sumInput=… is server-rendered (the HTML includes the computed sum, not an empty result shell).
• Input separators: commas, whitespace (spaces, tabs, new lines), semicolons and pipes. sumInput, input and a are list-text aliases. values also accepts an array of numeric strings or safe JSON numbers. Numeric array members are converted to shortest binary64 text first, which may select scientific mode; use strings to preserve notation. Multiple aliases must contain identical trimmed tokens in the same order; raw presentation uses sumInput, then values, input, a. Unknown fields, null/blank array members and unsafe numeric integers are rejected; send exact large integers as strings.
• Non-blocking warning: THOUSANDS_SEPARATOR_AMBIGUITY when a comma group looks like thousands grouping (7,890 → 7 + 890).
• Stable errors: MISSING_REQUIRED_INPUT, INVALID_NUMBER, INVALID_MODE, VALUE_MUST_BE_NON_NEGATIVE, UNSUPPORTED_FORMAT, TOO_MANY_VALUES, TOO_MANY_DIGITS, INVALID_INPUT, CONFLICTING_INPUT, UNKNOWN_INPUT, NUMERIC_UNDERFLOW, NUMERIC_RANGE, INPUT_TOO_LONG. Nonzero scientific underflow and nonfinite intermediate totals are errors; rescale extreme inputs.

Tips

• Prefer Sum when you only need a total or mean.
• Prefer Long addition when teaching place value or checking handwritten work.
• For percent of a total, use the Percentage Calculator after you have the sum.
• Pair with Subtraction, Multiplication, or Division for the other basic operations.