bayesian calculator

Use this Bayesian calculator to update your belief after new evidence arrives. It is especially useful for medical tests, fraud detection, quality control, and any situation where a result can be positive or negative.

Example: if 1% of people have a condition, enter 1.
How often the test is positive when the hypothesis is true.
How often the test is negative when the hypothesis is false.
Formula used:
P(H|E) = [P(E|H) × P(H)] / [[P(E|H) × P(H)] + [P(E|¬H) × P(¬H)]]

What this Bayesian calculator does

A Bayesian calculator takes your starting belief (the prior) and adjusts it using observed evidence. The updated belief is called the posterior. This is often the right way to reason when information arrives in stages rather than all at once.

Most people intuitively overvalue a positive signal and undervalue base rates. Bayesian thinking solves that by forcing both pieces into one coherent update.

Quick intuition: why priors matter so much

If something is rare, even a strong test can produce many false alarms. For example, with a rare disease, the number of healthy people is so large that a small false positive rate can still create a surprising number of positive results. That is why this calculator asks for prior probability first.

  • Low prior + decent test = many positives can still be false.
  • High prior + same test = a positive result is much more convincing.
  • A negative result may be very reassuring when specificity is high.

How to use the calculator

1) Enter prior probability

This is your base rate for the hypothesis before seeing the latest result. In medical screening, this could be prevalence in the screened population.

2) Enter test accuracy

Use sensitivity and specificity from the test’s validation data. If you only have false positive rate, remember: false positive rate = 100% − specificity.

3) Choose observed result

Select whether the latest evidence was a positive (+) or negative (−) signal.

4) Read posterior and natural frequencies

The output includes both a percentage and a “per 10,000” interpretation. Natural frequencies are easier to reason about and reduce mistakes in decision-making.

Worked example

Suppose:

  • Prior probability: 1%
  • Sensitivity: 99%
  • Specificity: 95%
  • Observed result: Positive

Even with a strong test, the posterior after one positive result is not 99%. Why? Because the base rate is low. Many healthy people are tested, and 5% of them can still show false positives. The calculator captures exactly this trade-off.

Common interpretation errors this avoids

Confusing P(+|H) with P(H|+)

Sensitivity tells you the chance of a positive result given the condition is present. It does not directly tell you the chance the condition is present given a positive result.

Ignoring prevalence

Prevalence (prior) changes everything. The same test can imply very different posteriors across different populations.

Binary thinking

A positive test is not certainty, and a negative test is not impossibility. Bayesian output helps maintain probabilistic thinking.

Where Bayesian updates are useful

  • Medical diagnosis and screening programs
  • Spam, fraud, and anomaly detection
  • A/B test interpretation with prior assumptions
  • Reliability engineering and defect testing
  • Risk analysis and legal evidence reasoning

Final thought

Bayesian reasoning is less about complicated math and more about disciplined belief updating. If you consistently combine priors with evidence quality, your decisions become more calibrated, more transparent, and usually much better.

🔗 Related Calculators