How the StockData Score works

The score runs 30 yes-or-no checks on every US company we cover, six in each of five areas. A company earns one point for each check it passes, so each area scores 0 to 6 and the total 0 to 30. The checks only use numbers from SEC filings and daily share prices, and every check shows the numbers behind it.

Missing data is shown, not punished. When a check cannot run — a company with losses has no P/E, a bank has no gross margin — it is marked as not enough data. It earns no point, but the page says why.

Industries are compared like for like. We group companies by SIC code and use the narrowest group with at least eight companies reporting the figure: first the three-digit SIC group, then the two-digit group, then one of eleven sectors, then the whole market. Multiples only include positive values.

The labels are plain. 22 or more is Strong, 15 to 21 Solid, 9 to 14 Mixed and below 9 Weak. They describe the checks, not whether to buy.

Valuation

Is the price reasonable against the company's own history, its industry and our fair value estimate?

1
P/E below its 5-year median
Today's price ÷ trailing diluted EPS, against the median of weekly P/E ratios over five years. Each week uses the earnings that had been published by then.
2
P/E below the industry
Against the median P/E of profitable companies in the same industry group.
3
EV/EBITDA below its 5-year median
Enterprise value ÷ trailing EBITDA. Banks and insurers use price-to-sales against the industry instead.
4
P/B below the industry
Market value ÷ book value, against the industry median. When book value is negative we compare price-to-sales.
5
Trading below fair value
The share price is below our fair value estimate.
6
Margin of safety over 20%
The price is at least 20% below fair value.

Growth

Are sales and earnings growing, and is the business getting better?

1
Revenue growing 10%+ a year
Compound growth of fiscal-year revenue over three years.
2
EPS growing 10%+ a year
Compound growth of diluted EPS over three years (both ends must be positive).
3
Earnings growth accelerating
Net income over the last twelve months grew faster than its five-year average rate.
4
Growing faster than the industry
Trailing revenue growth above the industry median.
5
Guidance raised
The company raised an outlook in an earnings release in the last twelve months (a cut fails; an unchanged or first-time outlook is neutral).
6
Gross margin improving
Trailing gross margin above the level a year earlier.

Past performance

Has the company earned good returns on the money invested in it?

1
Return on equity over 15%
Both over the last twelve months and as a five-year average.
2
Return on capital over 10%
After-tax operating income ÷ invested capital. Banks and insurers use return on assets over 1%.
3
Return on equity above the industry
Against the industry median.
4
Earnings grew over 5 years
Positive compound growth of net income over five fiscal years.
5
Net margin above its 5-year average
Trailing net margin against the average of the last five fiscal years.
6
Net margin above the industry
Against the industry median.

Financial health

Could the balance sheet carry the business through a bad year?

1
Low debt
Debt under 40% of shareholders' equity (no debt passes).
2
Debt falling against equity
Debt to equity lower than five years ago.
3
Debt covered by cash flow
Operating cash flow at least 20% of total debt.
4
Interest well covered
Operating income at least five times interest expense.
5
Short-term assets cover short-term debts
Current ratio of 1 or more.
6
Low bankruptcy risk
Altman Z-score above 3. Banks get six bank-specific checks instead: equity of 8%+ of assets, loans funded by deposits, a loan-loss reserve of 1%+, an efficiency ratio under 60%, growing deposits and modest borrowings.

Dividend

Does it pay a dividend that is attractive and that it can keep paying?

1
Pays a dividend
Dividends paid in the last twelve months. Companies that pay none score zero here.
2
Yield above Treasuries
Trailing yield above the 1-year Treasury yield.
3
Yield in the top quarter
At or above the 75th percentile of US dividend payers.
4
Dividend not cut
Dividend per share never fell more than 10% in a year over the last ten fiscal years (needs five years of history).
5
Covered by earnings
Dividends under 90% of net income.
6
Covered by free cash flow
Dividends under 90% of free cash flow.

Does the score separate good stocks from bad?

Every month since 2016 we rebuilt the score for every company using only what had been published by then: quarterly figures 45 days after the quarter ended, annual figures after 75 days, 13F holdings after 45 days and insider trades on the day they happened. We then split the stocks with enough trading into five equal groups by score and held each group for a month in equal weights, using dividend-adjusted prices.

Score group	Per year	Worst fall	Stocks held
1 (lowest scores)	+2.2%	-57%	534
2	+7.7%	-52%	535
3	+10.4%	-35%	535
4	+12.1%	-32%	535
5 (highest scores)	+14.0%	-32%	535
All stocks, equal weight	+9.4%	-35%

The same method for every idea

The lists on the ideas page are backtested the same way: each month the rules run on that month's data, the stocks that pass are held in equal weights for a month, and a month with no stocks holds cash. Lists that depend on data we only have for recent periods (daily volume, short interest, superinvestor holdings) are not backtested.

Alongside the yearly return we show volatility (the standard deviation of monthly returns, scaled to a year), the Sharpe ratio (monthly return above the one-year Treasury yield, divided by that standard deviation, scaled to a year) and how much of the list is replaced at a typical monthly rebalance. A list that swaps most of its stocks every month would lose more to trading costs than its backtest shows.

Live results, kept apart from the backtest

A backtest is chosen with hindsight, and studies of published screens find their edge tends to shrink once they are public. So from the day we first published each list we also track it as published: at each month end we buy that day's list in equal weights and hold it to the next, against all stocks with enough trading on the same day and the S&P 500. The published lists themselves are never rebuilt.

What the backtests leave out

Companies that were delisted, merged or went bankrupt before today are missing from our price history. This survivorship bias flatters every result, including the all-stocks comparison.
No trading costs, spreads or taxes. Lists with high turnover or tiny companies would do worse in practice.
Restated figures replace the originals, so a few months use numbers that were corrected later.
The S&P 500 line is the price index without dividends; the all-stocks line includes dividends.

The score is a summary of reported numbers, not a recommendation to buy or sell. See also how we estimate fair value.