Probability foundations and expected value
In a minute
Probability measures how likely an is, on a scale from to . When every is you count: . When you can't count, you estimate it from experiments, and either way you can predict how often the event should happen in many .
You'll be able to
- Use the words , , , and , and list or tabulate a sample space.
- Find a theoretical probability by counting, , when the outcomes are equally likely.
- Estimate a probability from experimental results using , and tell it apart from a theoretical probability.
- Use , , including for "at least one" questions.
- Find the expected number of times an event happens in a given number of trials.
Before you start3 quick questions on what this builds on
Trials, outcomes and sample spaces
Before you roll a die, you can't say which number will land face up. You can say exactly which numbers could land face up: , , , , or . Probability starts with that list.
For one roll of a die, and . The event , "an even number", is , so .
When a trial has two stages, such as a coin and then a spinner, list the outcomes as pairs. A list works when there are only a few. For two stages with several results each, a table is quicker and you can't miss any.
Try it first, then reveal one step at a time.
- Write each outcome as a pair: the coin result, then the spinner number. Take heads first and go through every spinner number.
- (a) With heads: . With tails: .
- (b) Count them: .
- Check: the coin has results and the spinner has , and each coin result pairs with every number, so .
Try it first, then reveal one step at a time.
- (a) Put the first spinner down the side and the second along the top. Fill each cell with the total of its row and column.
- Each cell is one outcome, so .
- (b) Find every cell holding : first with second , first with second , first with second .
- Two different spins can give the same total. Each cell counts once, so .
Theoretical and experimental probability
A fair die has , and of them are even. So you expect an even number half the time: . Counting like this gives a theoretical probability.
This formula is in the formula booklet, so you don't need to memorise it. You do need to know its condition: it works only when the outcomes in are equally likely.
Every probability lies between and . An impossible has probability , and a certain event has probability . If your answer is negative or bigger than , look for the slip.
Try it first, then reveal one step at a time.
- The is choosing one card. There are cards, each equally likely, so .
- Let be the event "the letter is a vowel". Count every vowel card, repeats included: three A cards, one E and one O. So .
- Use the formula: .
Try it first, then reveal one step at a time.
- The cells of the table are equally likely, because each spinner is fair. So .
- The totals that can occur are to . The primes among them are , , and .
- Count the prime cells row by row. Row : , which is cells. Row : , which is . Row : , which is .
- So and .
Some events can't be counted. A drawing pin can land point up or on its side, but the two outcomes are not equally likely. To find the probability, you do the experiment many times and see what happens.
A relative frequency gives an experimental probability: an estimate from data. A theoretical probability comes from counting equally likely outcomes, with no experiment.
The more trials you do, the closer the relative frequency tends to get to the true probability. So an estimate from many trials is more reliable than one from a few. A computer or GDC can simulate thousands of trials in seconds.
Try it first, then reveal one step at a time.
- The outcomes are not equally likely, so you can't count them. Use the experiment instead.
- Check the data: , so every throw is counted once.
- Let be the event "lands on its side". Its relative frequency is . This is the estimate of .
Try it first, then reveal one step at a time.
- (a) Mia: . Ravi: .
- (b) Ravi's estimate is better, because it comes from times as many trials.
- (c) If the die were fair, the theoretical probability would be .
- Ravi's is close to . So the data give no strong reason to think the die is unfair. Mia's is far off most likely because rolls is a small experiment, and small experiments vary a lot.
Complementary events
Every time you roll a die, either you get a or you don't. Exactly one of the two happens. So their probabilities must add up to .
Every outcome is in exactly one of and . So . Divide by to get the rule below.
This is in the formula booklet. You will mostly use it rearranged: . It holds for experimental probabilities too, not just counted ones.
Try it first, then reveal one step at a time.
- The cards are , so .
- The multiples of up to are . So and .
- Use the complement: .
- Check by counting: cards are not multiples of , and .
The complement is most useful for "at least one". "At least one" covers many cases. Its complement, "none", is a single case, and that is much quicker to count.
Try it first, then reveal one step at a time.
- The has equally likely outcomes: .
- Let be "at least one tail". Then is "no tails", which is only . So .
- Use the complement: .
- Check: of the outcomes contain a T, which agrees.
Try it first, then reveal one step at a time.
- Let be the event "the train is on time", so is "not on time". Write the given fact as an equation: .
- and are complementary, so .
- Substitute the first equation into the second: , so and .
- Then . Check: .
Expected number of occurrences
Roll a fair die times: that is . The "a six" has probability , so you would expect about of the rolls to be sixes: . You might get or , but is the long-run average.
This one is not in the booklet, so learn it. The expected number does not have to be a whole number. It is an average over many repeats, so is a perfectly good answer.
Try it first, then reveal one step at a time.
- The question asks about "on time", which is the complement of "late". So .
- Multiply by the number of parcels: .
- Leave the answer as . It is an average, so it need not be a whole number of parcels.
Try it first, then reveal one step at a time.
- No theoretical probability is given, so estimate it from the test. The of failing is .
- Use it as and multiply by the number of bulbs: .
- This is an estimate twice over: the probability came from an experiment, and the real count will vary around the average.
Common mistakes
- Treating outcomes as equally likely when they are not, such as the totals of two spinners. List or tabulate the outcomes that really are equally likely, then count those.
- Counting " then " and " then " as one outcome. In a two-stage trial, order matters: they are two cells of the table.
- Using where the question needs . Write down which event each probability belongs to before you substitute it.
- Taking the complement of "at least one" as "exactly one", or of "at least " as "at most ". The complements are "none" and "at most ".
- Calling a relative frequency the exact probability. It is an estimate, and one from a few trials can be far off.
- Rounding an expected number to a whole number, such as to . An expected number is an average, so give the exact value or 3 s.f. ().
Exam tips
- Write down a probability: no working is needed, but on Paper 1 give an exact value, such as or , not a rounded decimal. Check it lies between and .
- List or tabulate a sample space: show every outcome, in pairs for a two-stage trial, so the examiner can see none is missing.
- Comment on an experimental result: compare the relative frequency with the theoretical probability using both numbers, and say which estimate uses more trials.
- Find an expected number: write with the numbers in, then the value, unrounded unless the question says otherwise.
Key takeaways
Practise this
4 past-paper questions on 4.5 in the AA HL bank, each with a mark scheme and a worked explanation.