How to read an HDB Score

You are looking at a stock page and it says 8.4 / 10. This is about what that number is telling you, what it isn't, and how two companies with almost identical growth can land far apart. How the score is built, the eight signals and their exact weights, is on the methodology page.

It's a rank, not a grade

The first thing to know is that the score is comparative. A company is measured against the whole universe I track, so 8.4 means “near the top of this field” rather than “passed a fixed test”. There is no absolute threshold a business clears to earn an 8. It earns one by having a stronger, steadier record than most of the several thousand companies it's ranked against.

That matters when you compare scores across periods. A 9 over one year and a 9 over twenty aren't the same achievement, because the field is different and the bar for twenty consistent years is far higher.

What high, middling and low actually look like

• High. Steady, compounding revenue and profit over many years. Few down years, and none of them catastrophic.
• Middling. Real growth undercut by something: lumpiness, thin profitability, or simply a short history.
• Low. Stagnant, shrinking or loss-prone financials. Losses and negative years count against a company rather than being averaged away.

If a period doesn't have enough data to score honestly, it isn't scored at all. A company without ten years of history has no ten-year score rather than a guessed one, so a blank is information, not an omission.

Two companies, same growth, very different scores

This is the case that explains the score better than any description of it. Picture two businesses over ten years. The first grew profit in eight of those ten years, along a fairly smooth line. The second doubled profit in one blockbuster year and drifted for the rest of the decade.

Their total ten-year growth might be identical. Their scores won't be close. The first ranks far higher, because consistency and predictability are scored separately from growth, so the eight-out-of-ten record is rewarded and the single spike isn't.

That gap is precisely what a lone growth percentage hides, and it's the reason to open the chart before trusting any single number. What to look for is in reading a fundamental history.

Change the period and you change the question

The period selector isn't a display preference. A one-year score answers “how was last year?”, a question a single strong cycle can answer flatteringly. A twenty-year score answers “has this business delivered across recessions, cycles and management changes?”

My own research found the shorter windows carry much less signal about which companies turned out to be the long-run winners. Over one year, the best-performing stocks sat at roughly the 54th percentile on fundamentals, a coin flip, while over a decade or more they sat near the 80th and above. The evidence, including the parts that argue against this tool, is in I backtested my own stock rankings.

Practically: if you're looking at a long holding period, weight the long windows. The one-year ranking is the noisiest thing on the site.