Skip to content

Geometric Distribution Calculator

Enter the chance of success on each trial and the trial number, then choose exact, at-most, or at-least. The calculator returns all three probabilities plus the mean and standard deviation. With a 20% chance per trial, the first success lands exactly on trial 3 about 12.8% of the time, and within 3 trials about 48.8% of the time.

Updated 2026-11-19 Editorial estimate Calccc Editorial Team Free · No sign-up
%
Scenarios A/BSave two input sets and compare the results. Stored only in this browser.
AEmpty
BEmpty

Your inputs stay in this browser. No account, no uploads.

Professional worksheet

Detailed schedule & assumptions

Inputs used, method, sensitivity, and any year-by-year table. The summary above stays compact; the audit lives here.

Professional worksheet

Inputs, scope, and sensitivity

Inputs used

Method and scope

Result audit

MetricValue

What this calculator does

The geometric distribution describes the number of independent trials up to and including the first success. It answers questions such as how many sales calls before the first appointment, how many attempts before a component passes a test, or how many rolls before a six appears. The distribution has one parameter, the probability of success on each trial, and its shape is a decaying curve: the first success is most likely on trial one and becomes steadily less likely with each additional trial. This calculator returns the exact probability, the cumulative probability of success within a number of trials, the probability that the first success needs at least that many trials, and the summary statistics. It assumes independent trials with the same probability each time, which is the key limitation in real applications.

How to use it

  1. Enter the probability of success on a single trial as a percentage.
  2. Enter the trial number you want to analyse.
  3. Choose whether to show exact, at-most, or at-least probability.
  4. Read the other two probabilities in the secondary results.
  5. Compare the mean number of trials with your expectation.
  6. Check that the independence assumption is reasonable for your process.

How it works

P(X = k) = (1 - p)^(k - 1) × p. P(X ≤ k) = 1 - (1 - p)^k. P(X ≥ k) = (1 - p)^(k - 1). Mean = 1 / p. Variance = (1 - p) / p². Standard deviation = √((1 - p) / p²). In the default case with p = 0.2 and k = 3: exact = 0.8² × 0.2 = 0.128; at most = 1 - 0.8³ = 0.488; at least = 0.8² = 0.64.

Example calculations

Example 1: A 20% chance per trial

If each trial has a 20% chance of success, the first success lands on trial 3 about 12.8% of the time. Within three trials the chance is 48.8%, and the chance that it takes at least three trials is 64%. The mean number of trials is 5, so a run of five attempts is normal, not unlucky.

Example 2: A rare event

A process with a 2% success rate has a mean of 50 trials and a standard deviation of about 49.5. The chance of success on the first trial is 2%, and the chance of waiting more than 100 trials is 0.98¹⁰⁰, about 13.3%. Rare successes are heavy-tailed: long dry spells are expected, not exceptional.

Example 3: A high success rate

With an 80% success rate, the first success is on trial one 80% of the time, and within two trials 96% of the time. The mean is 1.25 trials. When the per-trial chance is high, extra attempts add little risk, which is why a process with a high pass rate rarely needs a long test plan.

What the result means

The mean is the average number of trials over many repetitions, not a prediction for a single run. A process with a 20% success rate has a mean of five attempts, but it can easily take two or twelve. The distribution is skewed: the most likely single outcome is always one trial, yet the average is higher because of the long tail. That is why a run of failures does not mean something is wrong with an independent process, and why a short lucky streak is not evidence of improvement. The independence assumption is the weak point in most real applications: if each attempt changes the chance of the next, or if the probability is not constant, the geometric model overstates the chances of a quick success and understates the tail.

Common mistakes

  • Using the distribution when trials are not independent.
  • Treating the mean as the most likely outcome.
  • Forgetting that the trial number includes the success itself.
  • Confusing the chance of success on a specific trial with the chance of success within a range.
  • Using a success probability that changes between attempts.
  • Assuming a long run of failures means the process is broken.

Assumptions and limits

  • Each trial has exactly two outcomes: success or failure.
  • The probability of success is the same on every trial.
  • Trials are independent of each other.
  • The trial number includes the successful trial.
  • The number of possible trials is not capped.
  • The success probability is known rather than estimated from small samples.

Frequently asked questions

What does the geometric distribution model?

It models the number of independent trials up to and including the first success, when each trial has the same probability of success. It is the natural model for repeated attempts with a fixed chance, such as sales calls or quality tests.

Why is the most likely outcome always one trial?

Because the probability of succeeding immediately is p, and every later trial requires a failure first, which multiplies the probability by less than one. The distribution decays from trial one, even though the mean is higher because of the tail.

What is the difference between exact and at-most probability?

Exact is the chance the first success happens on that specific trial. At-most is the chance it happens on that trial or any earlier one. At-most is always larger than exact and reaches 100% as the trial number grows.

How do I estimate the success probability?

Use historical data: the number of successes divided by the number of trials, measured over a period long enough to average out variation. If the success rate changes over time, use the most recent comparable period instead of a long-run average.

Can I use this for sampling without replacement?

No. When items are drawn without replacement, the probability changes after each draw and the hypergeometric distribution applies instead. The geometric distribution requires a constant probability with independent trials.

What is the expected number of failures before the first success?

The mean number of trials is 1 / p, so the expected failures before the first success are (1 - p) / p. With a 20% success rate that is four failures on average, followed by the success.

Sources

This calculator is an educational planning tool, not financial, legal, insurance, medical, or professional advice. Results depend on the inputs and on rules that vary by place and provider. Confirm important decisions with a qualified professional and the relevant documents.

2026 update

How to use the professional output

This page is not just a formula box. The result is designed to be checked, compared, and taken away.

  • Professional worksheet. The result includes the inputs used, the method and scope, and a one-variable sensitivity check for finance calculators.
  • A/B scenarios. Save two input sets in the browser and compare the primary, secondary, and breakdown figures side by side.
  • Print / PDF and CSV. Print a branded worksheet, or export the year-by-year table when the calculator produces one.

The formulas and assumptions above still govern the number. The worksheet makes the estimate auditable; it does not make it a lender quote, tax return, or personalized advice.

Calccc Editorial Team · Updated 2026-11-19

Prepared by Calccc Editorial Team as an educational estimate. This page is not a licensed professional review and is not financial, medical, legal, or tax advice.

Editorial estimate prepared by the Calccc Editorial Team; not a licensed professional review.