ValidAnytime
  • Try it free
  • Observatory
  • Pricing
  • Docs
Sign inStart free
ValidAnytime

A guaranteed false-alarm budget across your whole fleet, valid no matter how often you look.

Made by Compiled Intelligence — a frontier AI lab working on quantitative finance from first principles; ValidAnytime is the monitoring we built for our own model fleets, productized.

Product

  • How it works
  • Pricing
  • Observatory
  • Try it free
  • Dashboard

Solutions

  • Who it's for
  • For LLM engineers
  • For MLOps engineers
  • For SRE & platform
  • LLM eval regression
  • Model drift detection
  • Alert fatigue

Developers

  • Docs
  • Quickstart
  • Python SDK
  • API reference

Learn

  • Benchmark
  • Detector guides
  • Glossary
  • Compare
  • Why dashboards lie
  • Blog
  • Anytime-valid 101
  • Winter Storm Uri case study
  • Incident ledger
  • Observatory methodology

Company

  • Contact
  • Privacy
  • Terms

Observatory alarms & product updates, by email

Rare emails only. We store your address and the tag “footer” — nothing else. Unsubscribe anytime.

© 2026 ValidAnytime. All rights reserved.

Every alarm ships with its guarantee_tag and theorem_ref.

Glossary

Ville's inequality

Ville's inequality is the theorem that makes 'valid no matter how often you look' true rather than wishful.

Also known as: Ville's maximal inequality

Ville's inequality is the theorem that a non-negative process that does not grow in expectation — a martingale starting at one — has only an α chance of ever reaching the level 1/α, across all time at once. Not at one chosen moment: ever. It is the quiet engine under everything anytime-valid.

That single guarantee is what lets you check a monitor as often as you like. If you raise the alarm when the evidence crosses 1/α, the theorem caps your false-alarm rate at α no matter how many times you peeked along the way. The peeking problem simply does not apply, because the bound was proven over every look simultaneously.

In practice you never compute Ville's inequality by hand; it is the reason the e-process threshold is what it is. When ValidAnytime says an alarm carries a guarantee, this is the theorem the guarantee traces back to.

Go deeper

  • The math behind the demo (docs)
  • The coverage e-process — Ville's bound as an alarm

Related terms

  • E-processAn e-process is a running score of evidence against 'nothing has changed'; its value at any moment is an e-value, and it stays valid at every look.
  • Anytime-valid inferenceAnytime-valid inference is a way of testing that stays statistically valid no matter how often you look at the results.
  • Test martingaleA test martingale is a running evidence score that, if nothing has changed, is not expected to grow — the honest core of an e-process.
  • Always-valid p-valueAn always-valid p-value is a p-value you are allowed to read at any moment — it never gets less trustworthy the more often you check.

Put the theory to work.

ValidAnytime turns these ideas into a live alarm you can trust — valid no matter how often you look. Prove it on your own data, free.

Start freeTry the detector