Excel ERF Function
last modified October 5, 2026
The ERF function returns the error function, a special function
that appears throughout probability theory, statistics, and engineering. It
integrates the Gaussian curve between two limits, which makes it a natural
tool for working with the normal distribution. This tutorial provides a
comprehensive guide to using the ERF function with detailed
examples. You'll learn basic syntax, the one- and two-argument forms, and
practical applications.
ERF Function Basics
The ERF function returns the error function integrated between a
lower and an upper limit. With a single argument it integrates from zero to
that value, and with two arguments it integrates between them. The result is
closely related to the cumulative normal distribution.
| Component | Description |
|---|---|
| Function Name | ERF |
| Syntax | =ERF(lower_limit, [upper_limit]) |
| Arguments | lower_limit (required), upper_limit (optional) |
| Return Value | Error function value between -1 and 1 |
This table breaks down the essential components of the ERF
function. It shows the function name, the syntax, the argument requirements,
and the range of the return value.
| Argument | Meaning |
|---|---|
| lower_limit | The lower bound of integration. When it is the only argument, the integration starts at zero and this value acts as the upper bound. |
| upper_limit | The upper bound of integration. If omitted, ERF
integrates between zero and lower_limit. |
Both arguments must be numeric. A non-numeric argument produces a
#VALUE! error, while negative limits are accepted and yield a
negative result.
The Error Function Definition
The error function is defined by an integral of the Gaussian bell curve. For a
real number x, it is:
ERF(x) = (2 / SQRT(PI)) * integral from 0 to x of EXP(-t^2) dt
Because the integrand is even, the function is odd: ERF(-x)
equals -ERF(x), and ERF(0) is exactly 0. As
x grows, the value approaches 1 from below, so ERF(2)
is about 0.9953 and ERF(3) is already about 0.99998. The
two-argument form follows directly from the definition: ERF(a, b)
equals ERF(b) - ERF(a).
Basic ERF Example
This example demonstrates the simplest use of the ERF function
with a single limit. The function integrates from 0 to the specified value.
=ERF(1)
This formula calculates the error function from 0 to 1. The result is
approximately 0.8427, the value Microsoft documents for =ERF(1).
This shows how ERF works with a single argument.
ERF with Two Limits
ERF can calculate the integral between any two points, not just from
zero. The second argument is the upper bound, so the result is the difference
of the two one-argument values.
| A | B |
|---|---|
| 0.5 | |
| 1.5 | |
| =ERF(A1, A2) |
The table shows a simple spreadsheet with limits in cells A1 and A2. The
ERF formula in B3 calculates the integral between these two
points.
=ERF(0.5, 1.5) ' 0.4456, equals ERF(1.5) - ERF(0.5) =ERF(1, 2) ' 0.1526, equals ERF(2) - ERF(1)
The first formula calculates the error function between 0.5 and 1.5, which is
0.4456. The second integrates between 1 and 2 and returns 0.1526. Note that
ERF(0.5) alone is 0.5205, so the two-argument form is not the
same as evaluating one limit by itself.
One-Argument vs Two-Argument Forms
The two forms answer different questions. The single-argument form always starts at zero, while the two-argument form measures the area between any two bounds.
| Form | Integrates | Example | Result |
|---|---|---|---|
| =ERF(x) | from 0 to x | =ERF(1) | 0.8427 |
| =ERF(a, b) | from a to b | =ERF(1, 2) | 0.1526 |
Because ERF(a, b) = ERF(b) - ERF(a), the two-argument form can
return a negative value when the upper limit is smaller than the lower limit.
For example, =ERF(2, 1) returns -0.1526.
ERF with Negative Values
ERF accepts negative limits. The function is odd, so a negative input
produces the negative of the corresponding positive result, and a symmetric
interval such as (-1, 1) returns twice the one-argument value.
| A | B |
|---|---|
| -1 | |
| =ERF(A1) |
This table demonstrates the calculation with a negative input value. The function keeps its odd symmetry for negative inputs.
=ERF(-1) ' -0.8427 =ERF(-1, 1) ' 1.6854, equals 2 * ERF(1)
This formula calculates the error function from 0 to -1. The result is
approximately -0.8427. Integrating from -1 to 1 covers both tails and returns
1.6854, which is twice ERF(1).
ERF in Probability Calculations
ERF is often used with the normal distribution. The standard normal
cumulative distribution is 0.5 * (1 + ERF(z / SQRT(2))), and
ERF(z / SQRT(2)) gives the probability that a standard normal
value falls between -z and z.
| A | B |
|---|---|
| 1.96 | |
| =ERF(A1/SQRT(2)) |
The table shows how to use ERF to calculate a two-sided
probability for a standard normal distribution. The formula converts a
z-score into an interval probability.
=ERF(1.96/SQRT(2)) ' 0.9500, P(-1.96 < Z < 1.96) =0.5*(1+ERF(1.96/SQRT(2))) ' 0.9750, the one-sided CDF
The first formula returns approximately 0.9500, the familiar 95% central interval found in statistical tables. The second adds the lower tail to obtain the one-sided cumulative probability of 0.9750. This is why the 1.96 z-score appears so often in confidence interval calculations.
ERF with Cell References
For practical applications, ERF is often used with cell
references rather than hard-coded values. This example demonstrates this
approach.
| A | B | C |
|---|---|---|
| 0.2 | 0.8 | |
| =ERF(A1, B1) |
The table illustrates using ERF with cell references for both
limits. This approach makes the formula dynamic and easily adjustable.
=ERF(A1, B1) ' 0.5194 for A1=0.2 and B1=0.8
This formula calculates the error function between the values in A1 (0.2) and B1 (0.8). The result is approximately 0.5194. Cell references make the formula more flexible.
ERF vs ERF.PRECISE
Excel also provides ERF.PRECISE. Microsoft replaced the legacy
ERF with this newer function, which provides improved accuracy
and a clearer name. The difference is that ERF.PRECISE takes only
one argument and always integrates from zero.
| Function | Arguments | Integrates | Status |
|---|---|---|---|
| ERF | lower_limit, [upper_limit] | 0 to x, or a to b | Legacy, kept for compatibility |
| ERF.PRECISE | x only | 0 to x | Modern replacement |
=ERF(1) ' 0.8427 =ERF.PRECISE(1) ' 0.8427, identical result =ERF.PRECISE(1, 2) ' error, only one argument is allowed
For a single limit the two functions agree exactly, so =ERF(1)
and =ERF.PRECISE(1) both return 0.8427. When you need an interval
between two non-zero bounds, use ERF(a, b) or compute
ERF.PRECISE(b) - ERF.PRECISE(a).
ERFC Complementary Function
Excel also provides ERFC, the complementary error function. It is
defined as ERFC(x) = 1 - ERF(x), so it returns the area in the
tails beyond x rather than the area from zero to x.
=1-ERF(1) ' 0.1573 =ERFC(1) ' 0.1573, same value
For x = 1, ERF(1) is 0.8427, so
1 - ERF(1) equals 0.1573, which matches ERFC(1).
The complementary function is useful when the quantity of interest is the tail
probability.
ERF Errors and Troubleshooting
| Cause | Symptom | Fix |
|---|---|---|
| lower_limit is text | #VALUE! error | Make sure the cell or literal is numeric |
| upper_limit is text | #VALUE! error | Use a number or a numeric cell reference |
| Too many arguments | Too many arguments error | ERF accepts at most two arguments |
| An argument is an existing error | The same error is passed through | Fix the source cell before calling ERF |
| Empty cell used as a limit | 0 is used for that limit | Fill the cell or supply a default value |
The ERF function is essential for probability and engineering
calculations in Excel. From a simple one-argument evaluation to two-sided
normal probabilities, it maps the Gaussian integral onto the spreadsheet grid.
Remember that the single-argument form integrates from zero, that
ERF(a, b) equals ERF(b) - ERF(a), and that
ERF.PRECISE is the modern replacement for the single-argument
case. Mastering these relationships lets you move confidently between error
functions, z-scores, and cumulative probabilities.
Author
List all Excel Formulas.