A deep-value screen usually arrives as a list of 16 or so metrics with a threshold beside each one, split across valuation, quality, financial health, income and growth. Read that way it looks like 16 separate tests. It is not, and treating it as though it were will make a screen far more restrictive than intended.
Start with the clearest case. A price-to-earnings threshold of 15 or below and an earnings yield threshold of 0.0667 or above are the same test written twice. Earnings yield is EPS divided by price; P/E is price divided by EPS. One is the reciprocal of the other, and 1 divided by 15 is 0.0667. Screening on both does not double-confirm anything.
The same holds for price-to-book at 1.5 or below and book-to-market at 0.667 or above: 1 divided by 1.5 is 0.6667. Two more rows, one test.
So of the six valuation metrics in a list like this, two pairs collapse, and what looks like six independent valuation hurdles is really four.
There is a more interesting relationship hiding in the same two numbers. Multiply the P/E threshold by the price-to-book threshold: 15 times 1.5 is 22.5. That is the classic combined criterion from the Graham tradition, applied to the product rather than to each ratio separately. It is a deliberately looser test, and the difference matters: a company at a P/E of 10 and a price-to-book of 2.0 fails the individual price-to-book screen but passes the combined one at a product of 20. Which behaviour you want is a real decision, not a detail.
The other categories are genuinely independent of valuation, and that is the point of including them. Free cash flow yield at 0.08 or above asks whether the earnings are converting into cash. Debt-to-equity at 0.5 or below, current ratio at 2.0 or above and interest coverage at 5x or above ask whether the company survives long enough for the valuation to matter. A stock can clear every valuation hurdle precisely because the market has concluded it will fail those tests, which is the difference between a cheap company and a value trap.
Worth being explicit about what these thresholds are. The formulas are definitional, and the reciprocal relationships above are arithmetic. The threshold values themselves are a screening convention, in the Graham deep-value tradition, not a consensus standard and not a measurement. Reasonable screens use different cutoffs, and a threshold that suits an asset-heavy industrial will reject most software businesses outright, whose book value understates what they own.
None of this is investment advice, and no figure above says anything about what any particular stock is worth.
A screener built on these metrics is at deepvalueradar.com.