Piotroski F-Score

The Piotroski F-Score is a nine point financial strength scorecard. Each of the nine tests is a simple yes or no question about a company's most recent year of financial statements compared to the year before. If the company passes a test, it earns one point. If it fails, it earns zero. The scores sum to a total that ranges from zero (failed every test) to nine (passed every test). The model was introduced by Joseph Piotroski, at the time an accounting professor at the University of Chicago and later at Stanford, in a 2000 paper titled "Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers."

Piotroski was trying to solve a practical problem inside the value investing community. The Fama French research had shown that cheap stocks, measured by a low price to book ratio, outperformed expensive ones on average over long periods. But the outperformance came from a minority of the cheap basket. Most cheap stocks were cheap because they deserved to be, and dragged down the average. Piotroski asked whether you could use simple, fundamental, backward looking accounting data to separate the winners in the cheap basket from the losers, without having to do a full discounted cash flow on every name. The answer turned out to be yes.

The nine tests are grouped into three categories that collectively capture the dimensions of business health that matter to a value investor. Four tests measure profitability. Three tests measure leverage, liquidity, and the source of capital. Two tests measure operating efficiency. In Piotroski's original backtest, running from 1976 to 1996, a portfolio that went long the high F-Score cheap stocks (scores of eight or nine) and short the low F-Score cheap stocks (scores of zero or one) produced 23 percentage points of annual excess return before transaction costs. That is a remarkable result for a filter that any investor can compute in ten minutes on the back of an envelope.

Since publication, the F-Score has been independently replicated across international markets, across different definitions of "cheap", and across different decades. The result is robust: within a universe of statistically inexpensive stocks, high F-Score names outperform low F-Score names by a meaningful margin, and the effect does not arbitrage away once the paper is widely known. The likely reason is that the filter demands a kind of patience that most active investors do not have. By the time a company has passed all nine tests, the easy story is over. The price has often moved. The F-Score helps you stay disciplined about buying the boring, quietly improving company instead of the exciting, still declining one.

The F-Score is only expected to add value inside a universe of statistically cheap stocks. Running it on the most expensive stocks in the market is technically possible but statistically unsupported. High F-Scores on expensive stocks do not predict outperformance in the historical data. The F-Score is a filter, not a standalone rating.

As a screen on top of a value screen: this is the original and canonical use. Filter the investable universe to the bottom quintile or decile by price to book (or by enterprise value to EBITDA, or by price to free cash flow, depending on your value definition), then rank that subset by F-Score, and focus your research on the names with scores of seven or higher. The F-Score does the work of excluding the obvious value traps without requiring you to do a deep dive on every name.

As a monitoring tool on your existing holdings: compute the F-Score on your portfolio once a year, after the 10 K season. A stock whose F-Score has dropped from seven to four is telling you the fundamentals are moving against you, regardless of what the stock price has done. That is often the cleanest early warning the accounting statements will give you.

As a catalyst signal: the transition from a mid score (four or five) to a high score (seven plus) is often a fundamental turning point. In Piotroski's data, companies that improved their F-Score from one year to the next outperformed companies that declined, by a meaningful margin. Watching for that transition inside the value basket is a disciplined catalyst for incremental buying.

The F-Score is backward looking. It compares the most recent reported year to the year before. That works well in steady state businesses but poorly in businesses undergoing large structural changes, spin offs, major acquisitions, or accounting restatements. Read the 8 K filings before taking the F-Score at face value.

The accrual test (cash flow from operations higher than net income) is the single most important test in many practitioners' experience, and the one most commonly gamed. A company can pass the test by pulling cash into the period from receivables or pushing expenses out. The Beneish M-Score and the Sloan ratio catch some of that; the F-Score on its own does not.

The F-Score does not evaluate valuation. A score of nine on a stock at fifty times sales is not a buy signal. Piotroski's result is only documented inside the value universe. Outside that universe, the F-Score is a health check, not an alpha signal.

Finally, the F-Score treats all nine tests as equally weighted. There is academic work that suggests reweighting the tests, or using continuous rather than binary versions of them, produces better out of sample results. TickerXray reports both the classical nine point score and a continuous version scaled to the same range, so you can see both.

Joseph Piotroski published "Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers" in the Journal of Accounting Research in 2000, while he was on the faculty at the University of Chicago. The paper's central claim was that simple, fundamental accounting signals could separate the winners from the losers inside the high book to market universe, and that a strategy built on this signal would have earned substantial excess returns over the 1976 to 1996 period Piotroski studied. The paper was unusual for its combination of academic rigor and practical simplicity. The nine tests can be computed by any investor with a copy of the annual report. Subsequent research has replicated the result in international markets and later periods, and the F-Score has become one of the few academically supported fundamental investing filters that works reliably out of sample. Piotroski is now a professor at the Stanford Graduate School of Business.