Criteria decide who. The score decides order. Signals decide when.

1. The quality gate — who is eligible

Every night the screen reads annual reports for thousands of listed companies across the US, Canada, Hong Kong, mainland China, the UK and more. Each company faces eight tests covering 5–10 years of history:

• Revenue growing year after year
• Earnings per share growing (on core earnings, not one-off investment gains)
• Net margin above a floor
• Total assets growing
• Shareholders' equity growing
• Return on capital and return on invested capital above a floor
• Free cash flow growing
• Market capitalisation of at least US$1 billion

A company can miss at most two tests on real data. A test that can't be measured — a bank without a classified balance sheet, a company with too short a history — never counts against it.

First screen: the 15-year industry score

Before the tests, we ask whether the company earns its money in an activity that is likely to take a bigger share of the world economy over the next 15 years. Each of 58 economic activities is scored 0–10 on ten factors — how unavoidable it is, whether it is gaining share of global GDP, how much runway is left, technology and demographic tailwinds, and how hard it is to replace. A company's revenue is split across the activities it earns from, giving two numbers: its Ind 15Y score (0–10) and its exposure, the share of revenue in activities scoring 7 or more. Companies scoring below 6.0 don't reach the shortlist; a company not scored yet is never held back. Above that floor the industry score counts for 10 of the 100 score points, so a stronger industry ranks higher. Like the moat rating and the management score, it judges the business, not the share price.

2. The score — what order to look in

Companies that clear the gate are scored out of 100. Seventy points come from the fundamentals — growth, profitability and cash generation — and 10 each from the AI moat rating, the management score and the 15-year industry score. Within the fundamentals, recent years count far more than old ones, so a business that is getting better ranks above one that was great a decade ago. Businesses whose profits swing with the economic cycle are demoted, not removed. The score contains no price of any kind — a great company doesn't become a better business because its shares fell.

3. The buy signal — when the price is right

Only now does price matter. A 4/4 buy signal needs all four conditions at once:

1. Cheap: the P/E is low compared with the company's own 5-year range, not the market's.
2. Near its low: the price sits near the bottom of its recent range.
3. Turning: monthly MACD momentum has turned up.
4. Confirmed: trading volume backs the move.

The 3/4 and 2/4 views show which quality companies are getting close.

4. The sell watch — when to look again

For the stocks you hold, the screen watches every night for signs the business is deteriorating or the price has run far ahead of it. Any one raises a flag, shown as n/3 so you can see how serious it is. A flag is a prompt to review, never an order.