• Practice
  • About 15 min
  • 5 practice questions

Empirical probability: relative frequency

Find the empirical probability of an event as r/n from the results of repeated trials or from recorded data, and explain why more trials give a better estimate.

If something looks wrong, please tell us.

Some chances cannot be found by counting equally likely cases. Is a new light bulb likely to be faulty? How often does a bus arrive late? For such questions we look at what has actually happened many times. This is the empirical (experimental) approach, also called the relative frequency approach.

The empirical formula

If an experiment is repeated n times and an event E happens r of those times, then the relative frequency of E is r/n. The empirical approach takes this as the probability of E.

Empirical probability P(E) = r / n, where r = number of trials in which E happened and n = total number of trials

— Relative frequency

Classical and empirical probability

  • Classical
    • Counts equally likely outcomes
    • P(E) = n(E)/n(S)
    • Worked out before any trial
    • Example: P(head) = 1/2 for a fair coin
  • Empirical
    • Uses results actually observed
    • P(E) = r/n
    • Needs many trials or records
    • Example: 27 heads in 50 tosses gives 0.54

Worked example 1: tossing a coin

A student tossed a coin 50 times and got 27 heads (a made-up record). Find the empirical probability of a head and of a tail. Working: P(head) = 27/50 = 0.54. Tails came 50 − 27 = 23 times, so P(tail) = 23/50 = 0.46. Their sum is 0.54 + 0.46 = 1. The classical value for a fair coin is 0.5; the experiment gives a value close to it, but not exactly equal.

Why more trials help

If the class pools its results to make 1000 tosses, the relative frequency of heads usually comes closer to 0.5. In general, the larger the number of trials, the better the estimate. There is no guarantee that the value settles at exactly one number; repeating the experiment can give a slightly different relative frequency each time.

Estimating a probability by experiment

  1. Repeat the trial Under the same conditions.
  2. Count successes r = number of times E happened.
  3. Compute r/n The relative frequency.
  4. Add more trials Recompute; the estimate becomes more reliable.

Then back to step 1: Repeat the trial

Worked example 2: faulty bulbs

A shop sold 500 bulbs in a month and 8 were returned as faulty (made-up figures). Estimate the probability that a bulb bought there is faulty, and that it is good. Working: P(faulty) = 8/500 = 0.016. Good bulbs = 500 − 8 = 492, so P(good) = 492/500 = 0.984.

Worked example 3: relative frequencies of several groups

In a class of 60 students (made-up data), 24 walk to school, 21 come by bus and 15 by bicycle. Find the relative frequency of each. Working: walk = 24/60 = 2/5 = 0.4; bus = 21/60 = 7/20 = 0.35; bicycle = 15/60 = 1/4 = 0.25. Check: 0.4 + 0.35 + 0.25 = 1. Every student is in exactly one group, so the relative frequencies add up to 1.

Worked example 4: a sports record

A team won 9 of its last 25 matches and lost the rest (a made-up record). What is the relative frequency of a win, and of a loss? Working: P(win) = 9/25 = 0.36; losses = 16, so P(loss) = 16/25 = 0.64. This describes the past record; it does not promise what will happen in the next match.

Key takeaways: empirical probability = r/n from observed trials; it is used when outcomes cannot be counted as equally likely; more trials give a more reliable estimate; the relative frequencies of all possible results add up to 1.

Check your understanding

  1. 1. A dice is rolled 60 times and shows 6 on 12 of them. What is the empirical probability of a 6?

    1. 1/6
    2. 12/48
    3. 1/12
    4. 1/5
    Show answer

    Answer: D. r = 12 and n = 60, so P = 12/60 = 1/5 = 0.2. 1/6 is the classical value for a fair dice, not the observed one.

  2. 2. In a survey of 200 people (made-up data), 46 said they read a newspaper daily. What is the probability that a person chosen from them does not read one daily?

    1. 0.23
    2. 0.77
    3. 0.46
    4. 0.54
    Show answer

    Answer: B. Non-readers = 200 − 46 = 154. P = 154/200 = 0.77.

  3. 3. Which statement about empirical probability is true?

    1. It is found before doing any trial
    2. It always equals the classical probability
    3. It usually becomes a better estimate as the number of trials grows
    4. It can be greater than 1 if the event happens often
    Show answer

    Answer: C. Empirical probability comes from observed trials and gets more reliable with more trials. It need not equal the classical value, and since r ≤ n it can never exceed 1.

  4. 4. A fielder held 18 of the 24 catches that came to him in a season (made-up record). What is the relative frequency of a dropped catch?

    1. 1/4
    2. 3/4
    3. 1/3
    4. 6/18
    Show answer

    Answer: A. Dropped catches = 24 − 18 = 6. Relative frequency = 6/24 = 1/4. (6/18 wrongly divides by the catches held instead of all 24 chances.)

  5. 5. Of 120 seeds planted (made-up data), 102 sprouted. What is the empirical probability that a seed sprouts?

    1. 0.15
    2. 0.102
    3. 0.85
    4. 0.18
    Show answer

    Answer: C. P = 102/120 = 0.85. 0.15 is the probability that a seed does not sprout (18/120).

Privacy · Terms · Refunds · Delivery