How to Understand Probability: Rules and Examples

The direct answer: probability measures how likely an outcome is, from 0 (impossible) to 1 (certain). You combine probabilities with the addition rule for unions, the multiplication rule for independent events, and conditional probability for dependent ones. This guide builds each rule with examples you can copy, then shows the mistakes that cost points on tests.

Probability at a Glance

What Probability Really Measures

An experiment has a sample space of all possible outcomes. If those outcomes are equally likely, the probability of an event is the count of favorable outcomes divided by the total count. Probability is a ratio, so it never carries units like centimeters or seconds.

Worked example. Roll a fair six sided die. The chance of a 4 is 1 out of 6, because one face is favorable out of six. The chance of an even number is 3 out of 6, or 1 out of 2, since 2, 4, and 6 are favorable. Notice the logic is just counting, then dividing.

Two facts frame everything else. First, every probability lands between 0 and 1. Second, the probabilities of all outcomes in a complete sample space add to exactly 1. If you compute probabilities that sum to more than 1, you have double counted something. Khan Academy's probability library lays out this sample space approach clearly if you want more drills Khan Academy.

The Addition Rule

To find the probability of A or B, add their probabilities but subtract the overlap so you do not count it twice. The general form is:

P(A or B) = P(A) + P(B) minus P(A and B)

The subtraction matters whenever the two events can both happen. The overlap is the part counted once in P(A) and again in P(B).

Worked example. In a deck of 52 cards, the chance of a heart is 13/52 and of a king is 4/52. One card, the king of hearts, is both. So P(heart or king) = 13/52 + 4/52 minus 1/52 = 16/52.

When events cannot overlap

If A and B cannot happen together, they are mutually exclusive. The overlap is zero, so the rule shrinks to P(A or B) = P(A) + P(B). The classic case is a single die roll: you cannot roll a 2 and a 5 at once, so P(2 or 5) = 1/6 + 1/6 = 2/6.

The trap is assuming events are mutually exclusive when they are not. "Card is red" and "card is a face card" overlap on the red jacks, queens, and kings. Always ask whether the two groups share members before dropping the subtraction term.

The Multiplication Rule and Independence

For independent events, the chance of both is the product. The rule is:

P(A and B) = P(A) times P(B) when A and B are independent.

Independence means the result of one event does not change the chance of the other. Coin flips are the standard example because a tails on the first flip tells you nothing about the second.

Worked example. Flip a fair coin twice. The chance of heads then heads is 1/2 times 1/2 = 1/4.

Dependent events need conditional probability

If events are not independent, you cannot just multiply the raw probabilities. You adjust with conditional probability: P(A and B) = P(A) times P(B given A), where P(B given A) is the chance of B after A has already occurred.

Worked example. Draw two cards without replacement. The chance the first is a king is 4/52. Given that, only three kings remain among 51 cards, so the chance the second is a king is 3/51. Both kings is (4/52) times (3/51) = 12/2652, which reduces to 1/221.

The "without replacement" detail is what makes the events dependent. With replacement, the second draw would again be 4/52 and the events would be independent.

Conditional Probability and Bayes

Conditional probability is defined as:

P(A given B) = P(A and B) divided by P(B), provided P(B) is greater than 0.

The denominator, P(B), rescales the joint probability so the answer stays a proper fraction of the smaller world where B is already true.

Bayes' idea in plain terms

Bayes' theorem reverses a condition:

P(A given B) = P(B given A) times P(A) divided by P(B).

It is useful when you want to update a belief after seeing new evidence. A medical screening example makes this concrete. Suppose a test is 95 percent accurate and a condition affects 1 in 100 people. A positive result does not mean a 95 percent chance you have the condition, because the rare base rate pulls the true probability down sharply. Working the numbers with Bayes shows why a positive screen often warrants a second test. The intuition: rare conditions need strong evidence before you should believe them likely.

I always tell students that conditional probability is where probability stops feeling like counting and starts feeling like reasoning. Once you frame a question as "given that this already happened," the whole setup changes.

A numeric Bayes example

Suppose a condition affects 1 in 1,000 people, so the base rate is 0.001. A test is 99 percent likely to be positive if you have the condition, and 95 percent likely to be negative if you do not, which means a 5 percent false positive rate. Out of 1,000 people, about 1 truly has the condition and tests positive, while about 50 of the 999 healthy ones also test positive by false alarm. So of roughly 51 positive results, only about 1 is a real case.

The probability you actually have the condition given a positive test is about 1/51, near 2 percent, not 99 percent. The base rate dominates the answer. This is why screening for rare conditions pairs a first positive test with a confirmatory one, and it is the single best reason to learn conditional probability rather than trust the test accuracy number alone.

The Binomial Formula

When you repeat an independent trial with two outcomes, success or failure, the chance of exactly k successes in n trials is:

P(X = k) = (n choose k) times p to the k times (1 minus p) to the (n minus k)

where p is the success probability per trial and (n choose k) = n factorial divided by (k factorial times (n minus k) factorial).

A worked binomial example

A fair coin flipped 3 times. The chance of exactly 2 heads is (3 choose 2) times (1/2) squared times (1/2) = 3 times 1/4 times 1/2 = 3/8.

The (n choose k) term counts the arrangements. Two heads in three flips can happen as HHT, HTH, or THH, which is exactly three orders. That is why the coefficient is 3. Forgetting it is the most common binomial error: students compute p to the k times (1 minus p) to the (n minus k) but omit the number of ways to arrange the successes.

The binomial model only fits when trials are independent and each has the same success chance. Flips, yes. Free throws by a tired player whose accuracy drops, maybe not without adjustment.

Common Misconceptions

• Adding probabilities of overlapping events without subtracting the overlap. This double counts the shared cases.
• Treating dependent events as independent. Drawing without replacement is the usual culprit.
• Confusing "at least one" with "exactly one." "At least one" is often easier as 1 minus the chance of none.
• Forgetting that probabilities cannot exceed 1 or drop below 0.
• Misusing the binomial formula when trials are not independent or not identical.
• Believing a 95 percent accurate test means a 95 percent chance you have a condition. Base rates change that number.

Frequently Asked Questions

What does a probability of 0.5 mean?

It means the event is as likely to happen as not, like a fair coin landing heads.

Are independent and mutually exclusive the same?

No. Independent events do not affect each other's chances. Mutually exclusive events cannot occur together at all.

When do I use conditional probability?

When the outcome of one event changes the chances for another, such as drawing without replacement.

What is the difference between odds and probability?

Odds compare favorable to unfavorable outcomes. Probability compares favorable to all outcomes.

Why do probabilities add to 1?

Because the probabilities of every outcome in a complete sample space cover all possibilities, so together they total certainty, which is 1.

Where is probability used outside math class?

It underlies statistics, risk assessment, games of chance, weather forecasting, and the counting arguments common in discrete math.