On this concept: Formula

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

Statistics · Grade 6-8

Experimental Probability

Experimental probability is the probability of an event estimated from actual experimental data, calculated as the number of times the event occurred divided by the total number of trials.

P(E)=number of successesnumber of trials

Quick Answer

Experimental probability is the probability of an event estimated from actual experimental data, calculated as the number of times the event occurred divided by the total number of trials. Use it when you have results instead of a symmetric object: 27 point-up landings in 50 thumbtack drops gives 2750=0.54. It is the only way to estimate chances that cannot be reasoned out, such as a free throw or a bent coin. The estimate wobbles from one batch of trials to the next and steadies as the number of trials grows.

Why This Matters

Most real chances, such as a player making a shot, a seed sprouting or a machine part failing, have no symmetry to reason from, so the only way to put a number on them is to count what happened. Experimental probability also tests theory: if a supposedly fair die gives 25 sixes in 60 rolls, the gap from the expected 10 is worth checking. Without it you either guess, or trust a handful of trials too much.

Intuitive Explanation

Drop a thumbtack 50 times. It lands point-up 27 times and on its side 23 times, so your experimental probability of point-up is 27/50 = 0.54. No reasoning about its shape could have produced that number, because a thumbtack is not symmetric and its two ways of landing need not be equally likely. Data is the only route.

Drop it 50 more times and you might see 31 point-up, so the estimate becomes 58 out of 100, or 0.58. Neither number is wrong; both are estimates, and estimates from small batches swing more. With 10 drops, 7 point-up (0.7) would surprise no one; with 1,000 drops, the fraction usually moves only in the second decimal place. That is why an honest estimate always says how many trials it came from.

Rolling a fair die 60 times and getting 12 sixes gives an experimental probability of 12/60 = 0.2, while the theoretical probability is 1/6, about 0.167. These are different quantities: one comes from the die's symmetry, the other from these particular rolls, and a gap of this size is normal in 60 rolls, not proof of a loaded die. When a question gives counts of what happened, compute the experimental probability; when it gives a fair object and no data, reason it out theoretically.

Think of it like this. Judging a restaurant by its reviews instead of its menu is the experimental approach: you go by what actually happened to the people who ate there. Three reviews can mislead you; three hundred give a fair picture. The limit: reviews describe past meals, and if the kitchen changes, as when a coin gets bent, the old data stops applying.

Core idea. Experimental probability estimates a chance from what actually happened, successes divided by trials; it works for any process, symmetric or not, but it is only an estimate, and it becomes more reliable as the number of trials grows.


When to Use

Use experimental probability when the problem gives counts of results: "out of 80 spins, 26 landed on blue", "she made 45 of 60 shots". It is the right tool when the object is not symmetric or the process comes from the real world, like weather, sports or machines. Divide successes by total trials, never by failures. If the object is fair and no trials are given, use Theoretical Probability; if the question is about how estimates improve with more trials, use the Law of Large Numbers.

Tip. Ask: Is the chance estimated from results someone actually recorded, counting how often the event happened out of all the tries?

How to Recognize It

Before dividing, find the record of real tries and check that it comes from repeating one chance experiment.

  1. Does the problem give results from tries that were actually carried out, like 50 drops or 20 games?

    Yes means you are estimating from data. If you are only told the design, such as a fair die, compute the Theoretical Probability instead.

  2. Is each try a repeat of the same chance experiment, such as dropping the same thumbtack again and again?

    Yes makes the fraction an estimate of a probability. If the numbers describe a group, like favorite colors in one class survey, the fraction is a Relative Frequency that summarizes that data.

  3. Is the denominator the total number of tries, including the ones where the event did not happen?

    It must be. Dividing 27 successes by 23 failures gives the odds, not a probability; the correct estimate from 50 tries is 2750.

  4. Does the question ask how the estimate behaves as the number of tries keeps growing?

    Then the topic is the Law of Large Numbers, which explains why 500 tries give a steadier estimate than 20. A single estimate from the data you have is Experimental Probability.

Experimental Probability vs Theoretical Probability vs Relative Frequency vs Law of Large Numbers

These four meet whenever real results are compared with predictions, so they get used as if they meant the same thing. Ask whether you have trial data, a fair design, a summary of a group, or a claim about what happens as tries pile up.

Experimental Probability

Meaning
Experimental probability is an estimate of a chance found by dividing the number of times the event happened by the number of tries.
Key test
An estimate from repeated tries of one experiment; it moves as tries are added.
Formula
P^(E)=times E happenednumber of trials
Example
A paper cup tossed 40 times lands on its side 31 times, an estimate of 3140=0.775.

Theoretical Probability

Meaning
Theoretical probability is the exact chance predicted from equally likely outcomes, without collecting any data.
Key test
Needs a fair design and no trials; it stays fixed however many tries you make.
Formula
P(E)=favorabletotal equally likely
Example
A fair coin has a 12 chance of heads before it is ever flipped.

Relative Frequency

Meaning
Relative frequency is the share of observations in a data set that fall into one category, whether or not the data came from a chance experiment.
Key test
Summarizes any data set, such as survey answers, not only repeated tries.
Formula
category counttotal count
Example
If 18 of 30 students choose soccer as their favorite sport, its relative frequency is 0.6.

Law of Large Numbers

Meaning
The Law of Large Numbers says that as the number of tries grows, the experimental fraction tends to settle close to the true probability.
Key test
A statement about many tries, not a number calculated from one data set.
Example
A coin's share of heads might be 0.7 after 10 flips but is almost always within 0.02 of 0.5 after 10,000.

Formula & Notation

P(E)=number of successesnumber of trials
The experimental probability after n trials is P^(A)=count(A)n. By the Law of Large Numbers, P^(A)→P(A) as n→∞.

How to read it: P^(A) is the experimental (estimated) probability. n is the number of trials. As n increases, P^(A) converges to the true probability P(A).

Worked Examples

Example 1 — Estimate, then predict

Easy

A seed company plants 80 seeds of a new variety and 68 of them sprout. (a) What is the experimental probability that a seed sprouts? (b) About how many of 500 seeds should sprout?
Show the steps and the answer
  1. Identify the counts: 68 successes (seeds that sprouted) out of 80 trials (seeds planted).

    Each seed planted is one trial, and the event is "sprouts".

  2. Divide: 6880=0.85.

    Successes go over all trials. Dividing by the 12 seeds that failed would give 6812, a 'probability' above 1.

  3. Predict: 0.85×500=425 seeds.

    An experimental probability is a long-run share, so multiplying it by a new number of trials gives the expected count.

  4. Report the prediction as about 425.

    The 0.85 comes from only 80 seeds; another batch of 80 might give 0.80 or 0.90, so the prediction is approximate.

Answer: (a) 6880=0.85. (b) About 425 of the 500 seeds.

Takeaway: Divide successes by total trials, then multiply by a new number of trials to predict, and say 'about'.

Example 2 — Is the spinner unfair, or just a short run?

Standard

A spinner has 4 equal sections: red, blue, green and yellow. In 40 spins it lands on red 14 times, blue 8, green 9 and yellow 9. Compare the experimental and theoretical probabilities of red. Is this good evidence that the spinner is unfair?
Show the steps and the answer
  1. Experimental: P(red)=1440=0.35.

    This comes from what the spinner actually did in these 40 spins.

  2. Theoretical: P(red)=14=0.25, so a fair spinner would give about 0.25×40=10 reds.

    Four equal sections are equally likely, so this is the prediction to compare against.

  3. Measure the gap: 14−10=4 extra reds.

    The question is whether 4 extra reds in 40 spins is unusual for a fair spinner.

  4. Judge the gap: a fair spinner gives 14 or more reds in 40 spins about 1 time in 10.

    Something that happens 1 time in 10 is ordinary luck, so 40 spins cannot settle the question.

  5. Decide what to do: spin about 400 times. If red still comes up near 35%, about 140 times instead of 100, the spinner is likely unfair.

    Larger batches wobble much less, so a real bias would stand out.

Answer: Experimental 0.35 versus theoretical 0.25. Forty spins are not enough to call the spinner unfair; a gap this large happens by chance about 1 time in 10.

Takeaway: Experimental and theoretical probabilities usually differ in short runs; collect many more trials before deciding an object is unfair.

Example 3 — Find the mistake: dividing by failures

Application

A goalkeeper faced 40 penalty kicks and saved 10 of them. Ravi writes: "Saves to goals is 10 to 30, so her experimental probability of a save is 10/30 = 1/3." Find the mistake and correct it.
Show the steps and the answer
  1. Count the trials: every penalty kick is one trial, so there are 40.

    The denominator of an experimental probability is the total number of trials.

  2. Spot Ravi's denominator: 30 is the number of goals, the failures.

    '10 to 30' compares saves with goals, which is a ratio called odds, not a probability.

  3. Correct it: P(save)=1040=0.25.

    10 of the 40 kicks were saved.

  4. Check: P(save)+P(goal)=1040+3040=1.

    With Ravi's method, 1030+3010 is far more than 1, so it cannot describe the two outcomes.

Answer: P(save)=1040=0.25, not 13.

Takeaway: Put every trial in the denominator; successes compared with failures gives odds, not a probability.

Where You'll See It

  • Testing a thumbtack landing

    A thumbtack isn't symmetric, so you can't reason out P(point-up). You drop it 100 times, count point-up landings, and use successes/trials to estimate it.

  • Free throws over a season

    A player's experimental shooting probability is makes divided by attempts; with more games the estimate steadies toward their true skill.

  • Cereal box prizes

    Open 20 boxes, find 3 with the rare prize, and your experimental probability is 3/20 = 0.15, an estimate that sharpens as you open more.

Common Mistakes

  1. The slip: Experimental probability should equal the theoretical value.

    Why it's tempting: Both are 'the probability', so they seem like they must match.

    Instead: Experimental probability is an estimate from real trials and usually differs from theoretical; it only approaches it as the number of trials grows.

  2. The slip: A few trials are enough to nail down the probability.

    Why it's tempting: Getting a number quickly feels like you're done.

    Instead: Small samples are noisy; 3 heads in 4 flips (0.75) doesn't mean the coin is unfair — more trials are needed for a reliable estimate.

  3. The slip: The denominator is the number of successes.

    Why it's tempting: Students focus on the event that happened.

    Instead: Experimental probability = successes / total trials, so the denominator is every trial run, not just the favorable ones.


Mini Practice

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

  1. A spinner is spun 40 times and lands on blue 14 times. What is the experimental probability of blue?

    Hint: Divide the number of successes by the number of trials.

  2. A factory checked 250 cereal boxes and found a prize in 30 of them. Estimate the probability that a box has a prize, and predict how many prizes are in 1,000 boxes.

    Hint: Find the experimental probability first, then scale it up.

  3. Mia flips a coin 20 times and gets 13 heads. Find the experimental probability of heads and compare it with the theoretical probability.

    Hint: Theoretical probability for a fair coin is one of two equally likely outcomes.

  4. A number cube is rolled 60 times. The results are 1: 8 times, 2: 11, 3: 9, 4: 10, 5: 7, and 6: the rest. How many 6s were rolled, and what is the experimental probability of an even number?

    Hint: The counts must add to 60. Then add the counts for 2, 4 and 6.

  5. A café wants the probability that a lunch customer buys a drink. Can it work this out from equally likely outcomes, or should it use experimental probability?

    Hint: Is there a list of equally likely outcomes, like the faces of a die?

  6. Ben rolls a number cube 6 times and never gets a 4. He says, "The experimental probability of a 4 is 0, so this cube can never roll a 4." Correct him.

    Hint: Is 6 rolls enough to trust an estimate?

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

How many trials do I need before I can trust an experimental probability?

There is no magic number, but the typical error shrinks with the square root of the number of trials. Roughly, an estimate from 100 tries is usually within about 0.1 of the true chance, and one from 10,000 tries within about 0.01. So making the error 10 times smaller takes 100 times as many trials, which is why careful studies use thousands of tries.

If my experimental and theoretical probabilities are far apart, which one is right?

Look at the size of the gap and the number of trials. With 20 rolls, getting six or more 6s (0.30 or higher instead of about 0.17) happens with a fair die roughly 1 time in 10, so it proves little. With 2,000 rolls and 700 sixes, the gap is far too big to be luck, and the die is almost certainly unfair. A big gap over many trials points to a wrong assumption in the theory.

Can an experimental probability be 0 even though the event is possible?

Yes. If you open 10 cereal boxes and none has the rare prize, your estimate is 010=0\frac{0}{10} = 0, yet the prize exists. A zero estimate only says the event did not show up in your trials. Rare events need many trials before they appear at all, so a careful report would say the chance is probably small, not that it is impossible.

Why do two groups doing the same experiment get different experimental probabilities?

Because each batch of trials is a different random sample of what the object can do. Two classes each dropping the same thumbtack 50 times might get 0.54 and 0.62, and both are honest results. Pooling the data, 58 point-up landings out of 100, gives one estimate based on more trials, which is usually closer to the true chance than either alone.

Learning Path

Experimental Probability

You are here

You need Basic Probability (the 0-to-1 scale) and Data Collection (recording every trial honestly) first. Experimental probability adds estimating a chance from results and judging how far to trust an estimate from its number of trials. Next, the Law of Large Numbers explains why estimates settle down as trials pile up, and Statistical Simulation lets a computer run thousands of trials for you.

See Also