The Beneish M-score runs eight forensic ratios on a company's filings — receivables vs. sales, margin drift, accruals, asset quality — to estimate the probability that reported earnings are being dressed up. Below ~−1.8 is the clean zone; above it, manipulation risk rises. This heatmap colors US stocks by M-score, sized by market cap and grouped by sector.
Greener tiles score cleaner (lower M-score = fewer red flags), redder tiles show more manipulation-risk patterns — a forensic audit of the market in one view.
The original model uses roughly −1.8 as the cutoff — scores above it carry elevated manipulation probability. It catches statistical patterns, not proof: a red tile is a reason to read the 10-K, not an accusation.
Receivables growing faster than sales, falling gross margins, rising accruals, deteriorating asset quality and soft-asset growth — the classic tell-tales of aggressive revenue recognition and capitalization games.
Yes — fast growers and acquisition-heavy firms naturally trip the ratios. High M-score says "verify", not "guilty".
Heatmap data updates through the trading session. Fundamentals come from Finnhub and SEC filings; scores are simplified heuristics for screening — not investment advice.