JustilityBayesian belief updater
Update a binary belief using explicit priors and conditional evidence likelihoods.
State a binary hypothesis, enter a prior probability, and update it using observed evidence and explicit conditional likelihoods.
All inputs are percentages (0–100). For several items, enter likelihoods conditional on previous evidence, or assume conditional independence within both hypotheses. Reusing correlated evidence can overstate confidence. Up to 50 evidence items; calculations use log likelihoods.
Your work stays in this tab. Nothing is saved automatically. Save a readable, unencrypted project file before closing. Downloads remain wherever you save them.
Processed on your device. Your files are never uploaded by Justility.
How privacy worksHow it works
Open bayesian belief updater, enter your own project details or load the labeled example, then enter conditional likelihoods and update the belief. Review the results and download your project to continue later.
Supported formats
Editable JSON project up to 1 MB and CSV results; up to 50 evidence items, probabilities entered as percentages.
Good to know
Multiple evidence rows require conditional likelihoods given earlier evidence, or conditional independence under both hypotheses.
Worked example
Before
Prior 10%; P(E | H) 80%; P(E | not H) 20%
After
Posterior 30.769231%
Bayes theorem accounts for both the prior and how expected the evidence is under each hypothesis.
Bayesian belief updater: interpret the result
Update a binary belief using explicit priors and conditional evidence likelihoods. Multiple evidence rows require conditional likelihoods given earlier evidence, or conditional independence under both hypotheses. All project files are readable and unencrypted. Save and open them explicitly; there is no automatic browser persistence.
Scenario planning studioTroubleshooting
Why is my Bayesian evidence impossible?
The entered prior and likelihoods may assign zero probability to the evidence under both hypotheses. Revise the model; dividing by zero has no valid posterior.
Common questions
Can I use bayesian belief updater offline?
Prepare this tool using Offline help while online. After preparation, it runs locally with fresh input. Browser cache eviction remains possible; keep project downloads separately.
Can I use this without an internet connection?
Yes, after you prepare this tool from Offline help and the browser retains its assets. Installation is optional. All processing dependencies are bundled locally.