Conditional probability feels counterintuitive because it asks you to replace the original sample space with a smaller one after receiving new information. Your first probability may have been perfectly correct, but it answered the question before the condition was known. Once the evidence changes, the relevant group of outcomes changes too.
For IB students, the difficulty is rarely arithmetic. It is recognizing what the condition means, identifying the correct denominator, and resisting familiar but unreliable assumptions such as treating every remaining outcome as equally likely. This guide explains the concept without duplicating the wider syllabus coverage available in RevisionDojo's IB Maths AA probability explained and exam-focused notes.
What conditional probability actually means
The notation
P(A | B)
is read as “the probability of A given B.” It means that B is now known to have occurred, so outcomes outside B are no longer relevant.
The formal definition is
P(A | B) = P(A ∩ B) / P(B), provided P(B) > 0.
This formula is included in the official IB Mathematics: Analysis and Approaches guide. The guide also identifies Venn diagrams, tree diagrams, sample-space diagrams and outcome tables as valid ways to calculate these probabilities.
The numerator P(A ∩ B) represents outcomes satisfying both A and B. The denominator P(B) represents every outcome still possible after B is known.
A useful verbal translation is:
Among all outcomes satisfying B, what proportion also satisfy A?
That sentence often makes the calculation clearer than memorizing the formula alone.
Why the denominator changes
Suppose a school has 100 students:
GroupStudies chemistryDoes not study chemistryTotalStudies biology302050Does not study biology104050Total4060100
Let:
-
A be the event that a student studies chemistry.
-
B be the event that a student studies biology.
Before receiving any additional information,
P(A) = 40/100 = 0.4.
Now suppose you are told that the selected student studies biology. The 50 students who do not study biology are eliminated from consideration. Among the 50 biology students, 30 study chemistry, so
P(A | B) = 30/50 = 0.6.
The probability has changed from 0.4 to 0.6 because the reference group has changed from all 100 students to the 50 biology students. Nothing about the selected student physically changed. What changed was your information about the student.
This is the central idea behind conditional probability explained correctly: conditioning does not alter the event itself; it alters the set of outcomes against which the event is measured.
Why human intuition struggles with conditional probability
We keep using the original sample space
Students become accustomed to dividing by the total number of outcomes. In the example above, that encourages an answer of 30/100 because 30 students satisfy both conditions.
However, 30/100 is P(A ∩ B), the probability that a randomly selected student studies both subjects. It is not P(A | B) because the condition tells us to work only within B.
The vertical bar should therefore trigger a deliberate question: What is my new total? The event after the bar supplies that total.
We confuse P(A | B) with P(B | A)
Conditional probability is directional. In general,
P(A | B) ≠ P(B | A).
Using the same table:
-
P(chemistry | biology) = 30/50 = 0.6
-
P(biology | chemistry) = 30/40 = 0.75
The numerator is the same because both calculations involve students studying both subjects. The denominators differ because each question conditions on a different group.
This distinction is especially important in diagnostic-test questions. “The probability of testing positive given that a person has a condition” is not the same as “the probability that a person has the condition given that they tested positive.” The second calculation must also account for how common the condition is and how often people without it test positive.
We assume the remaining possibilities are equally likely
After some possibilities are removed, it is tempting to divide probability equally among what remains. That is valid only when the remaining outcomes genuinely have equal conditional probabilities.
Information is often produced by a non-random process. If someone deliberately reveals an outcome they know to be impossible, their action carries information. The remaining outcomes do not automatically receive equal probabilities merely because the same number of visible options remains.
This tendency is central to the Monty Hall problem. Research reviewed in a systematic study of Monty Hall reasoning identifies equal-probability assumptions and failure to represent conditional information as recurring sources of error.
We focus on evidence while ignoring base rates
A base rate is the initial prevalence or probability of an event before new evidence is considered. People often notice strong evidence but overlook how rare the underlying event is.
Imagine a condition affecting 1% of a population. A test correctly returns positive for 90% of affected people but also returns positive for 10% of unaffected people. Out of 10,000 people, the expected counts are:
GroupPositive resultNegative resultTotalHas the condition9010100Does not have the condition9908,9109,900Total1,0808,92010,000
Given a positive result, the relevant sample space contains the 1,080 positive results, not just the 100 people with the condition. Therefore,
P(condition | positive) = 90/1080 = 1/12 ≈ 0.0833.
The result can feel surprisingly low because the test is described as detecting 90% of affected people. That figure is P(positive | condition), not P(condition | positive). The large group without the condition generates many more false positives despite its lower individual positive rate.
Natural language hides the mathematical structure
IB questions may express a condition using phrases such as:
-
“given that”
-
“knowing that”
-
“of those who”
-
“if it is known that”
-
“selected from the students who”
These phrases all restrict the sample space. Under exam pressure, students sometimes skim them and calculate an intersection or an unconditional probability instead.
Rewrite the request symbolically before calculating. If a question asks, “Find the probability that the student takes chemistry, given that the student takes biology,” write P(C | B) immediately.
A better mental model: filter, then count
Instead of beginning with the formula, use this two-stage model:
-
Filter: Retain only outcomes satisfying the condition after the vertical bar.
-
Count or measure: Within that restricted set, find the proportion satisfying the event before the bar.
For P(A | B), filter to B first and then look for A within B. This explains the formula directly:
-
The filtered total is P(B).
-
The successful part of that total is P(A ∩ B).
-
Therefore, P(A | B) = P(A ∩ B)/P(B).
Natural frequencies, such as “90 out of 10,000,” are often easier to interpret than abstract percentages. A 2024 study on conditional reasoning and the Monty Hall problem describes how representing hypothetical repetitions and eliminating cases inconsistent with the evidence can improve understanding. The full method is discussed in the Cambridge paper on likelihood neglect and mental simulations.
How diagrams make the condition visible
The IB guide permits several representations because different structures suit different questions.
RepresentationBest used whenHow conditioning appearsTwo-way tableData are divided into two categorical variablesUse the conditioned row or column as the new totalVenn diagramEvents overlap within one populationRestrict attention to the region inside the conditioning eventTree diagramEvents happen in stagesFollow branches consistent with the known informationSample-space diagramOutcomes can be listed systematicallyCross out outcomes that fail the condition
For broader diagram and notation practice, see RevisionDojo's combined events and probability diagrams topic and its worked probability diagram notes.
Example with a tree diagram
A bag contains 3 red counters and 2 blue counters. Two counters are drawn without replacement. Find the probability that the second counter is red given that the first counter was red.
Once the first counter is known to be red, the bag contains 2 red and 2 blue counters. Therefore,
P(second red | first red) = 2/4 = 1/2.
The original probability of red was 3/5, but removing a red counter changes the composition of the bag. This is dependence: knowing the first outcome changes the probability of the second.
Conditional probability and independence are closely connected
Two events A and B are independent when knowing that B occurred does not change the probability of A. Symbolically,
P(A | B) = P(A), provided P(B) > 0.
An equivalent test is
P(A ∩ B) = P(A)P(B).
Conditional probability therefore measures whether information matters. Compare the original and updated probabilities:
-
If P(A | B) = P(A), B provides no probabilistic information about A.
-
If P(A | B) > P(A), knowing B makes A more likely.
-
If P(A | B) < P(A), knowing B makes A less likely.
Do not confuse independent with mutually exclusive. Mutually exclusive events cannot occur together, while independent events do not affect one another's probabilities. If two events with positive probabilities are mutually exclusive, learning that one occurred makes the other impossible, so they cannot be independent.
RevisionDojo's explanation of why independence matters in IB probability develops this distinction further.
How conditional probability appears in IB Maths AA
In the current published IB Mathematics: Analysis and Approaches syllabus, statistics and probability is one of five syllabus topics. The shared course content includes conditional probability, combined events, independence and diagram-based methods. Additional higher-level content develops the formal use of conditional probabilities, testing for independence and Bayes' theorem.
The official IB Mathematics AA subject brief outlines the course and assessment structure. Students should use the guide and instructions supplied by their school for their examination session, since official documents determine requirements.
This distinction matters: understanding the formula is not the whole skill. An IB question may require you to extract probabilities from a table, complete a tree, form an equation from a condition or decide whether events are independent.
A reliable exam method
For every conditional probability question:
-
Define the events. State clearly what A and B represent.
-
Translate the wording. Write the required probability as P(A | B).
-
Circle the condition. The event after the bar controls the denominator.
-
Choose a representation. Use a table for categories, a tree for stages or a Venn diagram for overlapping events.
-
Calculate the intersection. Identify outcomes satisfying both the event and the condition.
-
Divide by the conditioned total. Use P(A ∩ B)/P(B).
-
Check the result. It must lie between 0 and 1 and should match the restricted sample space.
If your answer uses the original total despite being given a smaller subgroup, reconsider the denominator. RevisionDojo's step-by-step probability problem guide provides a broader routine for mixed probability questions.
Common exam mistakes and how to correct them
MistakeWhy it failsCorrectionDividing by the original totalThe condition has restricted the sample spaceUse the total for the event after the barReversing A and BConditional probability is directionalTranslate the sentence before substituting valuesTreating “and” as “given”P(A ∩ B) is a joint probability, not a conditional oneDivide the intersection by P(B) for P(AAssuming independenceMany processes, especially without replacement, are dependentCompare P(AAssuming remaining options are equally likelyThe process revealing information may be selectiveModel how the evidence was producedRounding too earlyIntermediate rounding can alter the final answerRetain exact fractions or full calculator values
A further mistake is writing only a decimal with no setup. Even when calculator use is allowed, notation such as P(A | B) = P(A ∩ B)/P(B) makes the reasoning visible and reduces the chance of reversing the condition.
How to make conditional probability intuitive through practice
Start with tables containing whole-number frequencies. Ask “out of which group?” before calculating anything. Then move to percentages, Venn diagrams and two-stage tree diagrams.
A useful revision sequence is:
-
Translate ten verbal statements into conditional notation.
-
Solve table questions using only frequencies.
-
Solve the same questions using the formula.
-
Compare P(A), P(A | B) and P(A | B′).
-
Practise identifying dependence without calculation.
-
Finish with mixed exam-style questions where the required representation is not stated.
The RevisionDojo conditional probability and independence topic supports focused review, while the conditional probability Questionbank is useful once you can explain why the denominator changes. Jojo AI can help diagnose an incorrect setup, but you should first state the event definitions and your chosen sample space so the feedback addresses your reasoning rather than only the final number.
Conclusion
Conditional probability feels counterintuitive because ordinary intuition tends to preserve the original total, reverse the direction of evidence or distribute probability equally among visible options. The mathematical remedy is to treat new information as a filter: restrict the sample space to the condition, then find the proportion that also satisfies the target event.
For IB Maths AA probability, the most reliable habits are translating the wording into notation, identifying the conditioned denominator and using an appropriate diagram. RevisionDojo's Study Notes and Questionbank can then help turn that reasoning into a consistent exam method.