- Foundation
- About 17 min
- 5 practice questions
Theoretical probability, fairness and the law of large numbers
Find theoretical probabilities when outcomes are equally likely, use P(not E) = 1 − P(E), compare theoretical and experimental probability, and avoid the gambler's fallacy.
If something looks wrong, please tell us.
By the end of this lesson you can
- Find P(E) = favourable outcomes ÷ possible outcomes when all outcomes are equally likely, and use P(not E) = 1 − P(E).
- Decide whether the outcomes of an experiment are equally likely.
- Explain the law of large numbers and why past results of a fair coin or die do not change the next result.
For a fair die we do not need to roll it a thousand times to know that a 4 comes up about one time in six. We can reason it out: the die is symmetrical, so no face is favoured. Probability found by reasoning, without doing the experiment, is theoretical probability.
Equally likely outcomes
A coin or die is fair (unbiased) when it is symmetrical and tossed freely, so no outcome has any reason to occur more often than another. Its outcomes are then equally likely. The outcomes that make an event happen are its favourable outcomes.
When all outcomes are equally likely, P(E) = number of favourable outcomes ÷ number of possible outcomes. Also, P(not E) = 1 − P(E).
Worked example 1: a die
A fair die is rolled. Find (a) P(4), (b) P(a number greater than 4), (c) P(not greater than 4).
Working: there are 6 equally likely outcomes. (a) favourable {4}: P = 1/6 ≈ 0.167. (b) favourable {5, 6}: P = 2/6 = 1/3. (c) P = 1 − 1/3 = 2/3; check: {1, 2, 3, 4} is 4 of 6 outcomes, 4/6 = 2/3.
Worked example 2: letters of a word
One letter is picked at random from the 11 letters of MATHEMATICS. Find P(M) and P(a vowel).
Working: count each letter: M 2, A 2, T 2, H 1, E 1, I 1, C 1, S 1, total 11. P(M) = 2/11 ≈ 0.18. Vowels are A, A, E, I: 4 letters, so P(vowel) = 4/11 ≈ 0.36.
Worked example 3: balls in a bag
A bag has 3 red, 2 blue and 1 green ball. One is taken without looking. Each ball is equally likely, even though the colours are not. P(red) = 3/6 = 1/2, P(blue) = 2/6 = 1/3, P(green) = 1/6. P(not red) = 1 − 1/2 = 1/2. Check: 1/2 + 1/3 + 1/6 = 3/6 + 2/6 + 1/6 = 1.
When outcomes are not equally likely
'A car either starts or does not start' has two outcomes, but a well-kept car starts far more often, so each is not ½. 'The paper cup lands on its bottom, top or side' has three outcomes that are not equally likely. For such experiments we must use experimental probability.
Two ways to find a probability
-
Experimental
- Do the experiment or use data
- Times the event happened ÷ number of trials
- Works for any situation, such as a paper cup
- Changes a little from one experiment to another
-
Theoretical
- Reason, no experiment needed
- Favourable ÷ possible outcomes
- Needs equally likely outcomes, such as a fair die
- Always the same answer
The law of large numbers
Even with a perfectly fair die, 12 rolls might give three 4s (relative frequency 0.25) instead of exactly two. With few trials, experimental probability can differ a lot from theoretical probability. As the number of trials grows (60, 600, 6000 rolls), the relative frequency of 4 tends to settle closer and closer to 1/6 ≈ 0.167. This is the law of large numbers. Probability does not predict the next result; it describes what happens in the long run.
The gambler's fallacy
Suppose a fair coin shows heads six times in a row. Many people feel tails is now 'due'. It is not: the coin has no memory, and each toss is independent of the others. The probability of tails on the next toss is still ½. In the same way, after three 6s in a row in Ludo or Snakes and Ladders, the chance of another 6 is still 1/6. Believing otherwise is called the gambler's fallacy.
Key takeaways: for equally likely outcomes, P(E) = favourable ÷ possible, and P(not E) = 1 − P(E). Use theoretical probability only when the outcomes really are equally likely. Experimental probability gets closer to theoretical probability as the number of trials grows, but past results of independent trials never change the next one.
Check your understanding
-
1. A fair die is rolled once. What is the probability of getting an even number?
- 1/2
- 1/3
- 1/6
- 2/3
Show answer
Answer: A. Favourable outcomes are 2, 4 and 6: 3 out of 6. P = 3/6 = 1/2.
-
2. A letter is chosen at random from the word SCHOOL. What is the probability that it is O?
- 1/6
- 1/2
- 2/5
- 1/3
Show answer
Answer: D. SCHOOL has 6 letters and O appears 2 times. P(O) = 2/6 = 1/3.
-
3. A box has 4 red, 5 blue and 3 white beads. One is picked at random. What is the probability that it is not blue?
- 5/12
- 7/12
- 1/3
- 4/7
Show answer
Answer: B. Total = 4 + 5 + 3 = 12. P(blue) = 5/12, so P(not blue) = 1 − 5/12 = 7/12 (the 4 red and 3 white beads).
-
4. A fair coin has shown heads 5 times in a row. What is the probability that the next toss shows tails?
- More than 1/2, because tails is due
- Less than 1/2, because heads is on a streak
- 1/2
- 1/6
Show answer
Answer: C. Tosses of a fair coin are independent; the coin has no memory. The probability of tails on any toss is 1/2. Thinking otherwise is the gambler's fallacy.
-
5. For which experiment are the outcomes equally likely?
- A fair die is rolled: 1, 2, 3, 4, 5 or 6
- A car is started: it starts or it does not
- A paper cup is tossed: bottom, top or side
- A marble is drawn from 3 red and 7 blue: red or blue
Show answer
Answer: A. A fair die is symmetrical, so each face is equally likely. A car usually starts, a cup lands on its side more often, and blue is more likely than red when there are more blue marbles.
Sign in to keep track of the lessons you finish.