The Piotroski F-Score
A nine-question fundamental-strength scan published by Joseph Piotroski in 2000. The original paper showed that applying the F-Score filter to high book-to-market stocks raised the cohort’s annualized return by 7.5 percentage points — the most-cited evidence in academic finance that fundamentals still matter inside a value screen.
Each of nine questions is binary. Yes scores one point, no scores zero. The maximum is nine; a score of seven or higher is the “strong” cohort that drives Piotroski’s published outperformance. Below four is the weak cohort, where you’re probably looking at a value trap.
The nine tests
Profitability (4 points)
- Net income positive this year?
- Operating cash flow positive this year?
- Return on assets higher than last year?
- Operating cash flow higher than net income (cash quality of earnings)?
Leverage, liquidity, source of funds (3 points)
- Long-term debt ratio lower than last year?
- Current ratio higher than last year?
- No new shares issued this year?
Operating efficiency (2 points)
- Gross margin higher than last year?
- Asset turnover higher than last year?
What a score actually means
The score isn’t predictive in the way a discounted-cash-flow model pretends to be. It’s a checklist that catches the specific failure modes that turn cheap stocks into permanently impaired stocks: declining margins, rising leverage, dilutive equity raises, deteriorating asset utilization. A high F-Score doesn’t say “this stock will rise.” It says “this stock has not, in the most recent year, exhibited any of the eight mechanical signs that usually precede further deterioration.”
Worked example
Apple’s most recent F-Score is 8 of 9. Net income positive, cash flow positive, return on assets up, operating cash flow comfortably above net income, no debt expansion, current ratio improved, no significant share issuance, gross margin up, asset turnover up. The one missed point is usually a quirk of the year-over-year comparison rather than a flag.
NVIDIA scores 6 of 9 — high but not maxed. The misses tend to be on leverage and efficiency tests that get tripped by a company expanding its balance sheet faster than its revenue base, which is exactly what NVIDIA is doing. A 6 in this context is still in the “mixed but not weak” band; the misses are growth-induced, not deterioration-induced.
What to take away
The F-Score is a sanity check, not a thesis. It tells you whether a cheap-looking stock is also a deteriorating one. Pair it with the Beneish M-Score — together they catch most value traps before they trap you.