How To Do Conditional Probability | Solve Dependent Events Clearly

Conditional probability sounds harder than it is. Once you know what changes the sample space, the steps become clean and repeatable. You are not finding a brand-new kind of probability. You are updating the probability after new information comes in.

That update is the whole point. If you know someone already drew a red card, or you know a student already passed one test section, your next probability changes. The pool of possible outcomes is smaller, and the math must match that smaller pool.

This article shows the full method in plain language, then walks through examples with cards, dice, classes, and medical screening style numbers. You’ll also see where people slip up, especially when they mix up “A given B” with “B given A.”

What Conditional Probability Means In Plain Words

Start with two events:

• Event A: the result you want to measure
• Event B: the condition that is already true

“Conditional probability of A given B” means: what is the chance of A now that B has already happened?

In notation, people write it as P(A | B). The vertical bar means “given.” So you can read it as “probability of A given B.”

That bar is where most confusion starts. It does not mean “and.” It does not mean “or.” It means “we already know B is true, so use only outcomes inside B.”

That last line matters most. The sample space changes. If the original problem had 100 possible outcomes, and B narrows it to 20 outcomes, your denominator is now 20, not 100.

The Core Formula

The standard formula is:

P(A | B) = P(A and B) / P(B)

Use it when P(B) > 0. If B can’t happen, you cannot condition on it.

This formula works because the overlap of A and B is the part you care about, and P(B) is the new total space after the condition is known.

A Fast Visual Way To Think About It

Picture a box of outcomes. Event B cuts out a smaller box inside it. Once B is known, toss out everything outside that smaller box. Then ask: what fraction of the remaining outcomes also belong to A?

That fraction is your answer.

How To Do Conditional Probability Step By Step

Use the same sequence each time. This keeps you from mixing up the numerator and denominator.

Step 1: Name The Events

Write A and B in plain words. Short labels help. If you skip this, it gets messy fast.

Say:

• A = “second card is a king”
• B = “first card is a king”

Now the direction is clear. You are finding P(A | B), not some other version.

Step 2: Identify The New Sample Space

Ask what remains true after B happens. This is the step people rush past, and it causes most wrong answers.

If B says one card is already removed, the deck size changes. If B says a student is already in the honors group, the class totals change. Your denominator must match this updated pool.

Step 3: Count Or Compute The Overlap

Find the outcomes where A and B are both true. This is the numerator part in the formula, written as P(A and B) if you are using probabilities, or a raw count overlap if you are using counts.

When a problem gives raw counts, a count-based shortcut is often easier:

P(A | B) = (number in both A and B) / (number in B)

Step 4: Divide And Simplify

Once you have the overlap and the “given B” total, divide. Then simplify the fraction or convert to a decimal or percent if your class or site style prefers that format.

Step 5: Sanity Check The Result

Your answer must be between 0 and 1 (or 0% and 100%).

Also check whether the condition should raise or lower the chance. If the condition makes A more likely, your result should be larger than the original probability of A. If it makes A less likely, the result should be smaller.

How To Do Conditional Probability With Cards, Classes, And Dice

This section uses short examples with the same method. The wording changes, but the structure stays the same.

Example 1: A Standard Deck Of Cards

You draw two cards without replacement. What is the probability the second card is a king, given the first card is a king?

Let:

• A = second card is a king
• B = first card is a king

Once B is true, one king is already gone. There are 51 cards left, and 3 kings left. So:

P(A | B) = 3/51 = 1/17 ≈ 0.0588

Notice what changed: the denominator became 51 because the condition removed one card from the deck.

Example 2: Class Roster Counts

A class has 30 students. 18 play a sport. 12 are in band. 8 are in both groups. What is the probability a student is in band, given the student plays a sport?

Use counts:

• A = in band
• B = plays a sport

Among the 18 students who play a sport, 8 are also in band. So:

P(A | B) = 8/18 = 4/9 ≈ 0.444

The original class size (30) is not the denominator after the condition is known.

Example 3: Two Dice

You roll two fair dice. What is the probability the total is 8, given at least one die shows 3?

Let:

• A = total is 8
• B = at least one die is 3

First list B outcomes (ordered pairs):

(3,1), (3,2), (3,3), (3,4), (3,5), (3,6), (1,3), (2,3), (4,3), (5,3), (6,3)

That is 11 outcomes, since (3,3) was listed once.

Now find outcomes in B that also have total 8: (3,5) and (5,3). There are 2.

P(A | B) = 2/11 ≈ 0.1818

This one trips people because they keep 36 as the denominator. Once “at least one die is 3” is known, you must work only inside those 11 outcomes.

Problem Type	Condition (B) Changes	Conditional Setup
Cards Without Replacement	Deck size and card counts drop after each draw	Use remaining cards as denominator
Class Or Survey Counts	Total shifts to the selected subgroup	Use “both groups” over “given group”
Dice Or Coin Outcomes	Sample space shrinks to listed outcomes that satisfy B	Count overlap inside the reduced list
Medical Test Tables	Denominator depends on condition wording	“Given positive test” means test-positive row or column
Defective Item Checks	Inventory changes if sampling without replacement	Track remaining items after each draw
Weather Statements	Condition picks a subset of days	Use days with B, then count A among them
Exam Performance	Condition picks students with one outcome already true	Use intersection count over condition count
Venn Diagram Problems	Condition points to one circle/region as new total	Overlap area divided by conditioned area

Using The Formula With Probabilities Instead Of Raw Counts

Some problems give probabilities, not counts. In that case, use the formula straight away:

P(A | B) = P(A and B) / P(B)

Say a report gives:

• P(A and B) = 0.18
• P(B) = 0.30

Then:

P(A | B) = 0.18 / 0.30 = 0.60

So once B is true, A happens 60% of the time.

If you need a formal reference for the notation and rule, the conditional probability definition is stated in standard probability texts with the same structure used here.

When You Can Rearrange The Formula

You can also rearrange it to get the overlap:

P(A and B) = P(A | B) × P(B)

This is handy in multi-step problems. If a question gives a conditional probability and a base probability, you can rebuild the intersection and move on.

Independent Events Vs Conditional Events

If A and B are independent, then knowing B does not change the chance of A. In that case:

P(A | B) = P(A)

Say two coin tosses are fair. Let A = “second toss is heads,” B = “first toss is heads.” The first toss does not change the second toss. So the conditional probability stays 1/2.

Card draws without replacement are not like that. The first draw changes the deck, so the events are dependent.

Common Mistakes That Ruin Conditional Probability Answers

Most errors come from setup, not arithmetic. Catch these early and your answers improve fast.

Mixing Up P(A | B) And P(B | A)

These are not the same in most problems.

“Probability of disease given a positive test” is not the same as “probability of a positive test given disease.” One speaks from the test-positive group. The other speaks from the disease group.

Write the phrase in words before you touch the numbers. That one habit saves a lot of marks.

Keeping The Old Denominator

If B is known, the denominator becomes the number (or probability) of outcomes in B. People often keep the full class size, full deck, or full sample space. That gives a ratio for a different question.

Forgetting Whether Sampling Is With Or Without Replacement

This detail changes the answer.

With replacement, the pool resets after each draw. Without replacement, the pool shrinks. Card and ball-draw problems often hide this line in one short phrase, so read the wording with care.

Reading “At Least One” As “Exactly One”

These phrases are not interchangeable. “At least one” includes one, two, three, and so on. “Exactly one” means only one.

In dice and card problems, this wording changes the reduced sample space a lot.

Mistake	What Goes Wrong	Fix
Swap the condition	You solve P(B | A) instead of P(A | B)	Write the question in words before using symbols
Use the full total	Denominator ignores the condition	Use only outcomes where B is true
Miss replacement rule	Counts or probabilities change the wrong way	Mark “with” or “without” replacement first
Misread “at least”	Reduced sample space is too small	List all outcomes that satisfy the phrase
Use overlap as final answer	You stop at P(A and B)	Divide by P(B) to finish the conditional probability
Skip a reasonableness check	Answer falls outside 0 to 1 or clashes with context	Check bounds and whether B should raise or lower A

How To Work Faster On Conditional Probability Problems

Speed comes from structure, not from doing math faster. Use a repeatable layout on paper or screen:

1. Write the question in words.
2. Define A and B.
3. Mark the condition: “Given B.”
4. Build the reduced sample space (or subgroup total).
5. Count the overlap with A.
6. Divide overlap by the “given B” total.

If the problem uses a table of counts, scan the row and column labels before you compute. Many school and exam mistakes come from pulling one number from the wrong row.

If the problem uses percentages, convert them to decimals before dividing. It cuts clutter and makes your final check easier.

Use A Two-Way Table For Real-World Data

In school, sports, survey, and health examples, a two-way table keeps the setup clean. You can build one quickly with rows for one event and columns for the other.

The Khan Academy conditional probability lesson is a solid refresher if you want a second explanation with worked examples and notation practice.

When The Formula Feels Hard, Use Counts First

Many learners do better with counts than with symbols. That is fine. If a problem gives counts, start with:

(number in both) / (number in the condition group)

Then, if needed, translate it into notation after you already know the answer.

Practice Walkthrough With A Full Two-Step Setup

Here is one more example with a little more detail.

A jar has 5 red marbles and 7 blue marbles. You draw two marbles without replacement. What is the probability the second marble is red, given the first marble is blue?

Let:

• A = second marble is red
• B = first marble is blue

Step 1: Condition on B. If the first marble is blue, one blue marble is gone.

Step 2: Count the new jar contents. There are now 11 marbles left total, with 5 red marbles still in the jar.

Step 3: Compute the conditional probability.

P(A | B) = 5/11 ≈ 0.4545

That’s it. The condition changed the denominator to 11. The red count stayed 5 because the removed marble was blue.

Now compare it to a different condition: “given the first marble is red.” Then the answer would be 4/11. Same jar, same target event shape, new condition, new result.

This is why conditional probability matters. New information changes the chance, and the method gives a clean way to update it.

Final Check Before You Submit Or Publish

Before you move on, run this short check on your own answer:

• Did you read the direction correctly: A given B?
• Did you shrink the sample space to only B outcomes?
• Did you count the overlap inside that reduced space?
• Is your answer between 0 and 1?
• Does the result make sense after the condition?

If all five are true, your setup is in good shape. Conditional probability starts to feel easy once you treat “given” as a sample-space change every single time.