Probability diagrams, rules and independence
In a minute
, and tables turn a wordy probability question into regions and branches you can read off, often with no formula at all. A few rules cover the rest: the addition rule for "or", , for "given that", and multiplying for .
You'll be able to
- Read probabilities straight off a Venn diagram, a sample-space diagram or a table of outcomes, often with no formula.
- Use the addition rule , knowing that "or" includes both.
- Recognise mutually exclusive events and use .
- Find a conditional probability from a table or the formula, and use .
- Draw and use tree diagrams for two stages, with and , and interpret the result.
- Use independence to multiply probabilities, and test whether two are independent.
Before you start3 quick questions on what this builds on
Revise first: 4.5 Probability foundations and expected value
Venn diagrams and the addition rule
A station survey asks commuters whether they own a bike or a car. Some own both, so adding the two groups counts them twice. A picture stops that.
In maths, "or" includes both. " or " means only, only, or both. So is three regions: , and .
covers two regions, and so does . Adding them counts the overlap twice. Take it away once and you get the addition rule.
This is in the formula booklet. Given any three of its four probabilities, you can find the fourth.
Try it first, then reveal one step at a time.
- (a) Fill the overlap first: commuters are in .
- Bike only: . Car only: .
- Outside both circles: . Check: .
- (b) "Neither" is the region outside both circles, . Read it off: .
- Interpret: about of these commuters own neither a bike nor a car.
Try it first, then reveal one step at a time.
- Let be "watches drama" and be "watches comedy". "Drama or comedy" includes both, so .
- (a) Put the known values into the addition rule: .
- Rearrange: .
- (b) Exactly one is the union without the overlap: .
- Interpret: about of subscribers watch one of these genres but not both.
Mutually exclusive events
On one journey, a commuter either takes the train or takes the bus, never both. On a Venn diagram, those two circles don't overlap.
Put into the addition rule and the last term disappears. This version is in the booklet too, but use it only when and can't happen together.
Try it first, then reveal one step at a time.
- Draw a sample-space diagram of the products. Its cells are equally likely.
- (a) A multiple of is even, so never odd. No cell is in both events, so they are mutually exclusive.
- (b) Count the cells: odd products () and multiples of (; ; ; and all of row ).
- The events are mutually exclusive, so add: .
- Interpret: in the long run, players win three games in every four.
Try it first, then reveal one step at a time.
- Let , so .
- Nobody takes both, so and are mutually exclusive and .
- Every customer either takes a contract or doesn't, so .
- Solve : . So . Check: .
- Interpret: about of customers take contract .
Conditional probability
A phone repair shop records repairs by phone brand and whether the repair was finished the same day.
A random repair was finished the same day with probability . If you know the phone is brand Y, only that row matters: .
Knowing that an event has happened shrinks the sample space to that event. The new probability is a conditional probability.
Given , only the outcomes in are left, so divide the overlap by all of . The given event always goes on the bottom. This is in the booklet.
Multiply both sides by for a second form, which finds the probability of both. Learn it: the booklet shows only the first form.
Try it first, then reveal one step at a time.
- Let be "brand Z" and be "not finished the same day". The given event is .
- From the table: only the Later column matters. It has repairs, and of them are brand Z.
- So .
- The formula gives the same: .
- Interpret: about a quarter of the slow repairs were brand Z phones.
Try it first, then reveal one step at a time.
- Let be "has solar panels" and be "has a home battery". Then .
- " of the households with solar panels" means , not the probability of both.
- (a) Use the second form, with as the given event: .
- Interpret: about of all households have both.
- (b) The is only for households with solar panels, so the claim is not supported.
Tree diagrams, with and without replacement
Some experiments happen in stages: first the serve, then the rally. What happens at the first stage can change the chances at the second.
1. Multiply along a path. This is , with the first stage.
2. Add the paths you want. Different paths can't both happen, so they are mutually exclusive.
Try it first, then reveal one step at a time.
- Let be "first serve in" and be "wins the point". Draw the tree: or , then or , with and .
- (a) Two paths end in a win. Multiply along each: and .
- Add the two paths: .
- (b) Now is the given event, so it goes on the bottom:
- So . Interpret: about of the points she wins began with a first serve in.
With replacement, you put the first item back, so the second pick has the same chances. Without replacement, you keep it out, so the totals and the second-stage probabilities change.
Try it first, then reveal one step at a time.
- Let be "a teenager". First pick, from : and .
- "At least one teenager" is three paths. Its complement, "no teenager", is one: then .
- Second pick, after a non-teenager: of the left are not teenagers. So .
- (a) .
- (b) With replacement it would be . Keeping a non-teenager out leaves a larger share of teenagers for the second pick.
Independent events
Roll a die and check tomorrow's weather. Rain doesn't change the chance of a six: it is still .
Put into and you get the multiplication rule below. It is in the booklet.
If you're told the events are independent, multiply. To test whether they are, compare with .
Try it first, then reveal one step at a time.
- From the table: , and .
- Multiply: .
- Compare, showing both values: , so and are not independent.
- Interpret: , so brand X repairs were more often finished the same day.
Try it first, then reveal one step at a time.
- Let be "generator 1 starts" and be "generator 2 starts". Let , so .
- "At least one starts" is . Independence gives .
- Write the addition rule with these in: .
- Rearrange to a quadratic and write it down before using the GDC: .
- The polynomial solver gives or
- A probability can't be above , so reject . Then , which is valid. Generator 1 starts in about of power cuts.
Common mistakes
- Adding and and forgetting the overlap. Unless the events are mutually exclusive, subtract .
- Using the wrong denominator for "given that". Divide by the given event's total, or its probability.
- Multiplying for events not known to be independent. Use a diagram, the addition rule or .
- Treating picks without replacement as if the item went back. After each pick, reduce the total, and that type's count, by one.
- Calling events independent, or not, with no numbers. Compare with first.
Exam tips
- Write down a probability from a table or diagram: no working needed; give a fraction or 3 s.f.
- Find a conditional probability: show the fraction before the decimal. The numerator and denominator can each earn a mark.
- Determine whether events are independent: write and as numbers, compare them, then conclude.
- Solve for an unknown probability with the GDC: write the equation first, then reject any root outside to , saying why.
Key takeaways
Practise this
14 past-paper questions on Conditional probability in the AI HL bank, each with a mark scheme and a worked explanation.