Conditional Probability: The "Given That" Idea
Imagine you're rolling a fair six-sided die. The probability of getting a 4 is 61 — straightforward.
Now suppose I roll the die, you don't see it, but I tell you: "The number is even." Given that piece of information, what's the probability that it's a 4? Suddenly, the only possible outcomes are 2, 4, and 6. Among these three equally likely cases, exactly one is a 4. So the probability becomes 31.
That's conditional probability in action: the probability of one event, recalculated under the knowledge that another event has already occurred.
The Intuition: Shrinking the Sample Space
When you condition on an event, you're not changing the die or the randomness. You're simply restricting your attention to a smaller set of possibilities — the part of the original sample space where the condition holds.
In the die example:
- Original sample space: {1,2,3,4,5,6} (size 6)
- Condition "even": {2,4,6} (size 3)
- Favorable outcome "4": {4} (size 1)
- Conditional probability = 31
The key shift: you're no longer asking "out of all possible outcomes, how many are favorable?" You're asking "out of only those outcomes that satisfy the condition, how many are favorable?"
The Precise Definition
Let A and B be two events, with P(B)>0. The conditional probability of A given B is:
P(A∣B)=P(B)P(A∩B)
Read "P(A∣B)" as "probability of A given B."
P(A∣B)=P(B)P(A∩B)
Why this formula? Because:
- P(A∩B) counts outcomes where both A and B happen — these are the only ones that matter once B is known.
- P(B) is the total probability of the condition — we divide by it to rescale the probabilities so that the condition's total probability becomes 1.
In the die example: A={4}, B={2,4,6}.
- P(A∩B)=P({4})=61
- P(B)=63=21
- P(A∣B)=1/21/6=31
--- …