Picking & holding stocksScreening9 min readBeginner friendly

Two Numbers: Cheap, and Good

Key takeaway

Rank every company twice — how cheap and how good — then add the ranks. The winner is rarely best at either.
Return on capital is the quality measure: how much profit the business produces from the money tied up in it.
The formula spends years underperforming. That is not a flaw in it; it is the reason the advantage still exists.

Learning pathResearch and value a company from scratchStep 11 of 13

Read before this:The Four-Number Stock Checklist — Which Parts Survive

Based on The Little Book That Beats the Market — Joel Greenblatt, 2005

The whole method, in two numbers

Buy above-average businesses at below-average prices. The formula is just a way of ranking that sentence.

Joel Greenblatt wrote The Little Book That Beats the Market for his children, which explains the tone and also the discipline: he refused to use anything that could not be explained simply. What survived was two measures — earnings yield for price, return on capital for quality.

Rank on each measure separately, then add. Company C wins without topping either list, which is the mechanism: it selects for the combination rather than for an extreme.

The addition is doing real work. Ranking on cheapness alone finds businesses that deserve to be cheap. Ranking on quality alone finds excellent companies at prices that already assume excellence. The sum finds the overlap, which is where the disagreement between price and quality lives.

What return on capital is telling you

Two businesses given the same money. One returns $25 a year, the other $6. Over a decade of reinvestment that difference is most of what separates them.

This is the number Greenblatt uses as a proxy for a good business, and the logic is compounding. A company that earns 25% on the capital it employs can reinvest its profits at 25%; one earning 6% has to find somewhere else to put the money, or return it. The first compounds internally, and that is what makes the years pass in your favour.

The two axes are independent, and the two failure modes sit in opposite corners. The bottom-right corner is the one that punishes beginners: cheap because it deserves to be.

Why publishing it did not kill it

A strategy anyone can copy but almost nobody can hold is not really available to everyone.

The stretch in the middle is the price of admission. Three years of trailing the market is long enough for most people to conclude it stopped working.

Greenblatt is unusually direct about this: the formula's periods of underperformance are what protect it. If it worked every year, capital would flood in, the mispricing would close, and there would be nothing left. The discomfort is the barrier to entry.

Which reframes the practical question. It is not "does this work?" — it is "can I run something for five years while it looks broken for three of them?" That is a question about you rather than about the method, and it is the same question every rules-based approach eventually asks. See does your trading system fit you.

What the formula does not handle

It uses trailing figures. A cyclical company at a peak looks cheap and good simultaneously, right before earnings collapse — the classic false positive.
It excludes whole sectors. Financials and utilities do not fit the capital measures, so the screen is silent on a large part of the market.
It needs breadth. The results come from holding twenty or thirty names, not from picking the top two. Concentrating it removes the statistical basis it relies on.
It says nothing about the future. Both numbers describe what has already happened, which is why the scuttlebutt work still has a job.

Treated as a screen that produces a shortlist, it is a genuinely good use of an afternoon. Treated as an oracle that removes the need to understand what you own, it will hand you a cyclical at the top of its cycle and call it a bargain.

Try this week

Take five companies you already follow and rank them on earnings yield, then separately on return on capital.
Add the two ranks. Note whether the winner is one you would have chosen on either measure alone.
For the cheapest name on your list, write one sentence on why it is cheap. If you cannot, that is the finding.
Look up how the formula's published results did in its three worst consecutive years, and ask whether you would have stayed.

Common questions

What is Greenblatt's magic formula?

A screen that ranks companies separately on earnings yield and return on capital, adds the two ranks, and buys from the top of the combined list. It is designed to find good businesses at below-average prices rather than extremes of either.

What is return on capital and why does it matter?

It measures how much operating profit a business generates from the capital tied up in it. A high figure means the company can reinvest its own earnings at an attractive rate, which is what allows value to compound internally over time.

Does the magic formula still work?

Its published long-run results predate wide adoption, and returns have been less impressive in some later periods. The underlying idea — combining cheapness with business quality — remains sound, but treat it as a shortlist generator rather than a guaranteed outcome.

Why does the magic formula underperform for years at a time?

Because that is what keeps it available. Greenblatt argues that if the approach worked consistently, enough money would adopt it to eliminate the mispricing, so the multi-year stretches of trailing the market are the barrier that preserves the advantage.

How many stocks should I hold using this approach?

The published version relies on holding twenty to thirty names so that individual mistakes average out. Concentrating into a handful removes the statistical basis the method depends on and turns it into ordinary stock picking.