Three measures of how well a company's reported profit is backed by cash and how it compares to its own history. We surface them as risk flags, not buy signals. In our own historical study of what actually predicted returns, high reported "quality" and fat margins were, if anything, a mild anti-signal — the cheap, high-margin, low-reinvestment large caps that screened best tended to underperform. So we show these to help you avoid low-quality earnings, never to rank names as buys, and we never colour them green. Where an input is missing we leave the figure blank rather than guess.
The accrual anomaly (Sloan, 1996): the part of earnings not backed by operating cash flow tends to reverse, and companies with high accruals have historically gone on to lower returns. We measure how far a company's reported profit runs ahead of the cash it actually generated, scaled by its asset base, from its audited annual filings. A high reading is the caution end (earnings leaning on accruals); a negative reading means cash exceeded earnings. Because the measure is naturally negative for most mature, capital-intensive businesses, it is best read across comparable companies rather than against any single line — the screener lets you sort and filter on it directly.
From Ball, Gerakos, Linnainmaa and Nikolaev (2016): operating profitability stripped of the working-capital accruals — the receivables, inventory, payables and deferred-revenue swings — that can inflate reported operating profit. What remains is profitability on a cash basis, scaled by assets. We compute it from consecutive annual filings using the working-capital components a company reports; when a component is genuinely unavailable we note that on the figure, and when too much is missing we suppress it rather than publish a hollow number. We present the level neutrally — the useful reading is the gap between reported operating profitability and the cash-based version, which widens when reported profit isn't backed by cash.
Standardized unexpected earnings — the post-earnings-drift family — answers "how big was the last quarter's surprise, relative to this company's own track record of surprises?" We compare the latest quarter's reported earnings per share to the same quarter a year earlier and express that change in units of the company's own seasonal-surprise volatility, using only reported figures (no analyst estimates, which we do not license). It is a transparency flag — last surprise vs. own history — shown neutrally with direction only, not a rating. A large surprise is a prompt to check for one-off accounting items before reading it as momentum, and is most informative read alongside the accruals flag. It needs several years of quarterly history; for newly public or sparsely reported names it is left blank.
Each figure is computed from the same audited SEC filings the rest of the terminal uses, so the screener column always agrees with the company page. The exact averaging conventions, thresholds, minimum-history requirements and suppression rules are proprietary and not enumerated here; the measures are presented as descriptive risk flags and are not investment advice.
