Complement probability equals one minus P because an event and its complement cover every possible outcome without overlapping. The whole sample space has probability 1, so P(A) + P(Ac) = 1. Subtracting P(A) leaves P(Ac) = 1 − P(A).
The complement of an event means everything in the sample space that does not belong to that event. If your event is rolling at least 5, its complement is rolling less than 5. It includes all four results from 1 through 4, not just one different result.
This lesson explains why the subtraction works and how to choose the right event to subtract. For more probability lessons, visit the statistics and probability hub. The examples below use defined probability models, not observed results from an experiment.
Why Does The Whole Space Have Probability One?
The sample space is the complete set of outcomes allowed by your model. One of those outcomes must occur, so the chance of landing somewhere in that set is 1. This is the same amount as 100%, even when individual outcomes have different chances.
Call the event you care about A. The notation Ac means its complement, which you can read as “not A.” Each outcome belongs to A or to Ac, and no outcome belongs to both.
Those two facts explain the entire rule. There are no gaps because the groups cover the sample space, and there is no double counting because they do not overlap. Their probabilities therefore add to the probability of the whole space.
P(A) + P(Ac) = 1
P(Ac) = 1 − P(A)
Stanford’s introductory probability lecture defines a complement as all outcomes outside the original event and states this subtraction rule. The examples here apply that idea to separate, explicitly defined models.
Two groups that cannot occur together are called mutually exclusive. But that property alone does not make them complements: they must also include every possible outcome between them. Our lesson on mutually exclusive events explains that separate distinction.
Think of sorting every possible result into two boxes labeled “meets the condition” and “does not meet the condition.” You do not need the boxes to hold the same number of results. You only need each result to go into exactly one box.
How Does The Rule Work With A Die?
For one roll of a fair six-sided die, let A mean “the result is at least 5.” The event contains 5 and 6, so P(A) = 2/6. Its complement contains 1, 2, 3, and 4, giving P(Ac) = 4/6.
Subtracting gives the same result as counting: 1 − 2/6 = 6/6 − 2/6 = 4/6. The 1 becomes 6/6 because the complete set has six equally likely results. You can simplify 4/6 to 2/3.
Watch the boundary word “at least.” Since it includes 5, the complement must exclude 5. “Less than 5” is correct, while “at most 5” is not, because that group would overlap with the original event.
It helps to write both groups before using a formula. If a result appears in both groups, you have an overlap; if it appears in neither, you have a gap. Either problem means that your two groups are not complements.
Rolling 1 and rolling 6 cannot happen on the same roll, but those events are not complements. Results 2 through 5 belong to neither event. The complement of rolling 1 includes all five of the other die results.
Unequal Chances Still Follow The Rule
The complement rule does not require equally likely outcomes. Counting outcomes is one way to find a probability when all outcomes have equal chances. When they have unequal chances, add their assigned probabilities instead, then subtract the event’s total from 1.
Suppose a model has four possible outcomes named amber, blue, coral, and green. Their probabilities are 0.1, 0.2, 0.3, and 0.4, which add to 1. Let A contain amber and blue, so P(A) = 0.1 + 0.2 = 0.3.
| Outcome | Probability | Group |
|---|---|---|
| Amber | 0.1 | Event A |
| Blue | 0.2 | Event A |
| Coral | 0.3 | Complement of A |
| Green | 0.4 | Complement of A |
The complement contains coral and green, whose probabilities total 0.7. Subtraction agrees: 1 − 0.3 = 0.7. Each group contains two outcomes, but the groups do not have a 50% chance each.
The important amount is probability, not the number of labels. A group with many rare outcomes can be less likely than a group with one common outcome. Equal counts imply equal chances only when the underlying outcomes are equally likely.
The same rule works for a probability stated without an outcome list. If a complete model gives P(A) = 0.73, then P(Ac) = 0.27. In percentage form, the calculation is 100% − 73% = 27%.
Use one scale throughout the subtraction, and label the answer clearly. A decimal probability pairs with 1, while a percentage pairs with 100%. Writing 1 − 73 mixes those scales and produces a number that cannot be a probability.
Why Is The Complement Of At Least One None?
“At least one” includes every count from one upward, so its complement is zero. For three coin tosses, at least one head includes one, two, or three heads. The only result outside that group is no heads, meaning all three tosses show tails.
Assume each coin toss is fair and the tosses are independent. There are eight equally likely ordered outcomes, and only one has no heads. Therefore P(no heads) = 1/8, and P(at least one head) = 1 − 1/8 = 7/8.
You could count the seven outcomes directly, but the complement takes less work. This advantage grows when an event covers many different ways to succeed. Finding the single group called “none” often avoids a long list of cases.
“Exactly one head” is a different event with three outcomes in this model. Its complement includes zero heads, two heads, and three heads. Subtracting the chance of exactly one head would not answer a question about at least one head.
For n independent trials, each with the same success probability p, the chance of no successes is (1 − p)n. You can then subtract that amount from 1. The independence and fixed probability assumptions justify that product, not the complement rule itself.
Ask which calculation has fewer cases before you begin. If the desired event has one simple case, count it directly; if its complement has one simple case, use subtraction. Both paths describe the same probability when the model and event definitions match.
Dependent Draws And Conditional Probability
Complement subtraction still works when trials depend on each other. What changes is how you calculate the event you plan to subtract. You must account for earlier results instead of assuming that every trial has the same chance.
Consider a bag with three blue balls and two red balls. Draw two balls at random without replacement, and find the chance of at least one blue ball. The complement is drawing no blue balls, which means drawing red twice.
The first red draw has probability 2/5. After that red ball leaves the bag, one red ball remains among four balls, so the next red draw has probability 1/4. Thus P(no blue) = (2/5) × (1/4) = 1/10.
Now subtract: P(at least one blue) = 1 − 1/10 = 9/10, or 90%. Squaring 2/5 would be wrong because it treats the second draw as though the bag still contained two red balls. Removing the first ball changes that chance.
A condition can also change the sample space you are using. Given that a fair die result is even, only 2, 4, and 6 remain possible. Within that smaller space, the event “at least 5” contains 6 and its complement contains 2 and 4.
Write this as P(Ac | B) = 1 − P(A | B), where B is the condition and P(B) is greater than zero. Both probabilities must use the same condition. Here the values are 2/3 and 1/3, rather than the original die probabilities.
Check The Event Before Subtracting
Before subtracting, state the event and its complement in plain words. Keep the same trial, time period, population, and conditions on both sides. For example, “no heads in three tosses” cannot serve as the complement of “at least one head in four tosses.”
Also check that your probability model is complete. If a table leaves out outcomes or contains unexplained missing data, do not label its leftover amount as the complement. First establish what the omitted part means and whether the stated probabilities cover the whole space.
The boundary values provide a quick check. An event with probability 0 has a complement with probability 1, while an event with probability 1 has a complement with probability 0. Any result below 0 or above 1 signals a mistake in the input or arithmetic.
Keep exact fractions until the final step when possible. For the die example, using 1/3 gives a complement of exactly 2/3. Using a rounded decimal such as 0.33 instead gives 0.67, which is only an approximation of that answer.
Use the probability calculator for the arithmetic after you have defined the events and checked your assumptions. The core reasoning stays simple: identify all outcomes outside A, find the probability of A, and subtract it from the whole probability of 1.