On this concept: Formula

Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 6-8

Probability as Expectation

Probability as expectation reads P as a long-run relative frequency: $P(\text{heads})=0.

Expected count=n⋅P(event)
E = 1/6 · n06121824303642485460(0, 0)

A point at 0 rolls on the expected-sixes line: every 6 rolls trades for exactly 1 expected six.


Quick Answer

Probability as expectation reads P as a long-run relative frequency: P(heads)=0.5 means about half of many flips are heads. Use it to turn a probability into a prediction for many trials by multiplying the number of trials by the probability, n⋅P. A basketball player who makes 80% of free throws should make about 40 of 50. The result is a count to expect on average, not a guarantee; real results land near it.

Why This Matters

This reading lets you plan with probabilities. A bakery that knows 1 customer in 20 orders a birthday cake can expect about 15 cake orders from 300 customers and bake for that. It also lets you check a game's fairness, estimate how many faulty parts a factory will make, and notice when observed counts are far from what chance predicts.

Intuitive Explanation

Roll a fair die 6 times and you might get no sixes, or three. Roll it 600 times and the number of sixes will usually be somewhere near 100. The probability 16 does not tell you what any single roll will do; it tells you the share of sixes you will see over a long run. That is why you can multiply: 600 rolls times 16 of them being sixes gives about 100.

The expected count is a center, not a promise. In 600 rolls the actual number of sixes usually lands somewhere between about 82 and 118, so 91 or 108 would be perfectly normal. What tightens as you roll more is the share: 108 sixes out of 600 is 18%, and 1,020 out of 6,000 would be 17%, both closing in on 16≈16.7%.

This reading does not apply to one trial. A 30% chance of rain cannot be checked on one day, because it either rains or it does not; it describes what happens over many days with that forecast. Keep the two kinds of answer apart as well: P is always between 0 and 1, while n⋅P is a count that can be any size, like 100 sixes.

Think of it like this. Probability as expectation works like a recipe scaled up for a crowd. If one cookie in every 4 gets sprinkles, a batch of 200 should have about 50 with sprinkles. Where the recipe idea breaks: a baker decorates exactly every fourth cookie, while chance only makes the count land near 50, not exactly on it.

Core idea. P(event) is the fraction of trials the event happens over very many repetitions.


When to Use

Use n⋅P when a question asks how many times an event should happen over a number of trials: sixes in 60 rolls, faulty bulbs in a box of 500, heads in 1,000 flips. Words like expect and about how many are signals. If the question asks about one single trial, give the probability itself; if outcomes carry different values, like prize money, you need expected value instead.

Tip. Ask: Am I predicting a long-run count or share, not a single outcome?

How to Recognize It

Before you multiply, check what the question hands you, a probability and a number of trials, and what it wants back: a count, a single chance, or an average amount of money or points.

  1. Is a number of trials given, such as 60 rolls or 500 bulbs?

    If yes, multiply it by the probability to get the count to expect: 60×16=10 sixes. If there is only one roll, the answer is the probability itself, 16, not a count.

  2. Do different outcomes carry different amounts, such as $5 for one prize and $20 for another?

    If yes, you need expected value: multiply each amount by its probability and add. Probability as expectation counts how often one event happens; it does not weigh outcomes by what they are worth.

  3. Were you given results that already happened, like 18 heads in 40 flips?

    If yes, 1840=0.45 is an experimental probability, measured from data. Predicting future counts uses a probability you trust, theoretical or from a long record, times the number of new trials.

  4. Does your answer make sense as a count, between 0 and the number of trials?

    It should: 500 bulbs with a 2% fault rate gives 500×0.02=10 expected faults. An answer like 0.02 means you stopped at the probability, and one over 500 means a slip.

Probability as Expectation compared with Theoretical probability, Expected value, and Experimental probability

All four rows start from a probability, so a fraction gets given where a count was wanted, or a count where money was. Ask whether the answer is one likelihood, a share measured from past results, an expected count over many trials, or an average value.

Probability as Expectation

Meaning
Use this when you want to predict how many times an event occurs over many trials from its probability.
Key test
Am I predicting a long-run count or share, not a single outcome?
Formula
Expected count=n⋅P(event)
Example
You roll a fair die 60 times. About how many sixes should you expect?

Theoretical probability

Meaning
Is the single 0-to-1 likelihood, not a predicted count over trials.
Key test
Use when you just need the chance of one event.
Formula
P(E)=favorabletotal
Example
P(6)=16

Expected value

Meaning
Weights outcomes by value, not just counts occurrences.
Key test
Use when each outcome carries a payoff or score.
Formula
E[X]=∑xiP(xi)
Example
Average winnings per game

Experimental probability

Meaning
Comes from actual observed results, not the theoretical P.
Key test
Use when you have real trial data, not a known P.
Formula
observedtrials
Example
47 heads in 100 real flips

Formula & Notation

Expected count=n⋅P(event)
P(A)=lim⁡n→∞count of A in n trialsn; expected count in n trials =n⋅P(A)

How to read it: n is the number of trials; P is the probability per trial; n⋅P is the expected count

Worked Examples

Example 1 — Expected sixes

Easy

You roll a fair die 60 times. About how many sixes should you expect?
Show the steps and the answer
  1. Find the probability for one roll: P(6)=16.

    One face out of six equally likely faces is a six.

  2. Note the number of trials: n=60.

    The question asks for a count over many rolls, so you need both the probability and how many times you roll.

  3. Multiply: 60×16=10.

    In the long run one roll in six is a six, and 60 rolls make ten groups of six rolls.

  4. Read the result as a typical value: getting 7 or 13 sixes would be normal.

    The count varies from one set of 60 rolls to the next; 10 is where it centers.

Answer: About 10 sixes

Takeaway: Multiply the number of trials by the probability to get the expected count, and expect the real count to land near it.

Example 2 — One roll, not many

Standard

What's the probability the very next single roll is a six?
Show the steps and the answer
  1. Count the trials: just 1 roll.

    With one trial there is no long run to average over, so a count like 10 makes no sense here.

  2. Use the per-roll probability: 16≈0.17.

    The question asks how likely one result is, and that is the probability itself.

  3. Check the form of the answer: it lies between 0 and 1.

    A probability always does; an expected count, like 10, need not.

Answer: 16

Takeaway: For one trial, give the probability. Multiply by n only when the question is about many trials.

Example 3 — Find the mistake: exactly 50 heads?

Application

A fair coin is flipped 100 times and lands heads 46 times. Zoe writes: "The probability of heads is 0.5, so there should have been exactly 50 heads. Getting 46 proves the coin is unfair." Find the mistake and correct it.
Show the steps and the answer
  1. Compute the expected count: 100×0.5=50 heads.

    Zoe's expected count is right; her error is treating it as a guarantee.

  2. Find the gap: 50−46=4 heads, a share of 46÷100=0.46.

    To judge a gap, compare it with how much a fair coin's count normally varies.

  3. Use the normal spread: in 100 fair flips the head count usually lands between about 40 and 60.

    A fair coin rarely hits 50 exactly; the chance of exactly 50 heads is only about 8%.

  4. Correct the conclusion: 46 is well inside that range, so it gives no evidence that the coin is unfair.

    Only a count far outside 40 to 60, such as 75, would point to a biased coin.

Answer: 46 heads is a normal result for a fair coin; 50 is the expected count, not a promise.

Takeaway: n⋅P is the center of the likely results. Real counts land near it, and hitting it exactly is uncommon.

Where You'll See It

  • Airlines overbooking flights

    If 5% of ticket holders usually miss their flight, an airline with 200 seats expects about 200×0.05=10 empty seats. Some airlines sell a few extra tickets for that reason, and accept that on some days more passengers show up than expected.

  • A basketball season of free throws

    A player who makes 75% of free throws and takes 120 in a season should expect about 120×0.75=90 makes. If she makes 88, she performed about as expected; if she made 70, the coach would look for a real problem.

  • Quality checks on light bulbs

    If 1 bulb in 500 is faulty, a shipment of 10,000 bulbs should contain about 10000÷500=20 faulty ones. A store uses that count to keep enough replacements in stock.

Common Mistakes

  1. The slip: Expecting the exact expected count

    Instead: n⋅P is the long-run average, not a promise for any one run.

  2. The slip: Applying it to a single trial

    Instead: probability-as-expectation describes many repetitions, not one outcome.

  3. The slip: Confusing the predicted count with a probability

    Instead: n⋅P is a count, while P stays between 0 and 1.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. A spinner is one quarter blue. If you spin it 80 times, about how many times should it land on blue?

    Hint: Multiply the number of spins by the probability.

  2. A basketball player makes 70% of her free throws. About how many will she make in 50 shots?

    Hint: Change 70% to a decimal and multiply.

  3. A factory finds that 2% of its light bulbs are faulty. About how many faulty bulbs should a box of 400 contain?

    Hint: Find 2% of the number of bulbs.

  4. Two dice are rolled 360 times. About how many times should the sum be 7?

    Hint: First find the probability of a sum of 7 out of 36 outcomes.

  5. A thumbtack is tossed 100 times and lands point-up 62 times. Is 0.62 a theoretical or an experimental probability? How many point-up landings would you predict in 500 tosses?

    Hint: Was the 0.62 found by reasoning or by trying it?

  6. Jon rolls a fair die 60 times and gets 7 sixes. He says: "The die must be unfair, because I should get exactly 10." Correct him.

    Hint: Is n⋅P a promise or a long-run average?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

If I expect 10 sixes in 60 rolls but get 15, should I expect fewer sixes later?

No. The next 60 rolls still have an expected count of 60×16=1060 \times \frac{1}{6} = 10 sixes, so after 120 rolls you should expect about 15+10=2515 + 10 = 25, not 20. The die does not owe you fewer sixes. The extra 5 simply stay in the total, and the share of sixes drifts back toward 16\frac{1}{6} only because the total keeps growing.

Does n⋅P work when the chance changes from one trial to the next?

Not as a single product, but adding works. If you take 3 free throws with chances 0.9, 0.8 and 0.7 because you are getting tired, the expected number made is 0.9+0.8+0.7=2.40.9 + 0.8 + 0.7 = 2.4. The formula n⋅Pn \cdot P is the shortcut for adding the same PP to itself nn times. Adding the chances gives the expected count even when the trials affect each other.

If each child in a family has a 1-in-4 chance of a trait, will 1 child in 4 have it?

Only on average across many families. When two parents each carry one hidden copy of a gene, each child has a 14\frac{1}{4} chance of showing the trait, so a family of 4 children expects 4×14=14 \times \frac{1}{4} = 1. Real families vary: about 32% of such families have no child with the trait, about 42% have exactly one, and about 26% have two or more.

Learning Path

← Before

Probability
Probability as Expectation

You are here

You need Probability first: a number from 0 to 1 for one trial. This concept reads that number as a long-run share and multiplies it by the number of trials to predict a count. Next, Expected Value weights each outcome by what it is worth, and the Law of Large Numbers (Intuition) explains why observed shares close in on P.

See Also