About

Bubble Pressure publishes daily gauges of possible market bubbles, built from free public data. Each gauge shows its workings, including where it has been wrong.

Who runs it

Bubble Pressure is built and edited by Brendan Faulds, an enterprise architect and software developer in South Australia. He is responsible for everything published here.

He is not a licensed financial adviser and does not work in finance. He holds no shares or other investments in any company covered here, or in any listed company. The site is an independent project. It is not affiliated with any of the companies it covers, any data provider, or any financial institution.

What it does

Each gauge follows a fixed basket of listed companies. A data pipeline runs twice a day, at 8:45am and 4:45pm Sydney time. It collects share prices, company filings and market data, then recalculates the gauge.

The first basket is AI stocks. Its gauge combines six signals into a Bubble Pressure score. It also fits an experimental bubble model, LPPLS, and estimates the chance of a burst from past market episodes. The method is set out in full on the gauge's page.

How it keeps itself honest

• Every forecast goes into a forecast log when it is published. Entries are never rewritten, and each one is hash-chained to the one before.
• Probability forecasts are scored with a Brier score once their horizon has passed.
• The model's backtest is published, including results that show the evidence is weak.
• Errors are fixed and listed, with dates, in each gauge's corrections section.
• The gauges have been through several external reviews. Each claim in a review was checked before anything was changed.

Use of AI

A daily commentary and a weekly reading of news headlines are written by an AI model, Anthropic's Claude. Each is labelled as AI-written where it appears. Every figure the commentary quotes is checked automatically against the data, but no person edits it before it is published. The weekly news reading can lean the burst estimate higher or lower in words. It never changes the number.

The pipeline code and parts of this site were also written with AI assistance and checked by the editor.

Money

There are no ads, no affiliate links, no sponsors and no paywall. The running costs are paid by the editor. If that changes, this page will say so.

Data

The data comes from free public sources, including Nasdaq, the US Securities and Exchange Commission, the Federal Reserve Bank of St. Louis (FRED), Cboe, the US Treasury, Naver Finance for Japan and Korea, and the Taiwan Stock Exchange. Each gauge lists its sources with the date of the latest observation. The data behind each gauge can be downloaded as JSON.

Get in touch

Corrections, questions and suggestions for new baskets are welcome. See Contact.

Last updated 8 November 2026.