Lesson 6 of 7
The Piotroski F-Score in practice
The F-Score sums nine pass or fail tests from a company's own accounts. Piotroski built and tested it only inside the cheapest fifth of the market.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- The F-Score adds nine pass or fail tests, taken from a company's own filed statements, into one number from 0 to 9.
- Piotroski built and tested it only inside the cheapest fifth of the market by book-to-market, not across every stock.
- He classified 0 or 1 as low and 8 or 9 as high, leaving a middling score outside both of his own labels.
The Piotroski F-Score sums nine yes-or-no tests, taken straight from a company's own financial statements, one point for each test passed. Piotroski built it to sort within stocks already cheap by book-to-market, not to replace a cheapness test on its own.
The nine tests, in the three areas Piotroski used
| Area | Test | Scores one point if |
|---|---|---|
| Profitability | Return on assets | Positive this year |
| Profitability | Cash flow from operations | Positive this year |
| Profitability | Change in return on assets | Higher than the year before |
| Profitability | Accrual | Operating cash flow exceeds net income |
| Leverage and liquidity | Change in leverage | Long-term debt to assets fell |
| Leverage and liquidity | Change in current ratio | Current ratio improved |
| Leverage and liquidity | New shares issued | None issued this year |
| Operating efficiency | Change in gross margin | Gross margin improved |
| Operating efficiency | Change in asset turnover | Asset turnover improved |
Built inside the cheapest fifth, not the whole market
Piotroski's own tests were run only inside the highest book-to-market quintile of the market, the fifth of companies priced cheapest against their book value. His stated purpose was narrower still: to see whether a high F-Score identified the stronger businesses within that already-cheap group, not to test the score against the whole market.
What "high" and "low" meant in his own terms
Piotroski classified a score of 0 or 1 as low, expecting those firms to have the weakest subsequent returns, and a score of 8 or 9 as high, expecting the strongest. He described scores from 3 to 7, where most of his sample fell, as carrying conflicting signals rather than a clear one.
One-year market-adjusted return by F-Score group
- High F-Score (8 or 9)13.4%
- All high book-to-market5.9%
- Low F-Score (0 or 1)−9.6%
Running it outside his sample
Using the F-Score on the whole market, rather than inside the cheapest fifth by book-to-market, asks a different question than the one Piotroski tested. This product's own F-Score screen makes that substitution explicit. It stands in for his book-to-market quintile with a price-to-book ceiling of 1.0. The stand-in is a loose one: in his 1976 to 1996 sample, the median company in that quintile had a book-to-market ratio of 1.721, a price-to-book near 0.58, noticeably cheaper than 1.0.
Run it on a real 10-K
The nine tests, one filing at a time
- Return on assets: net income before extraordinary items, divided by beginning total assets. Score a point if positive.
- Operating cash flow: from the cash flow statement. Score a point if positive.
- Change in return on assets: this year's ROA against last year's. Score a point if it rose.
- Accrual: this year's operating cash flow against this year's net income. Score a point if cash flow is higher.
- Change in leverage: long-term debt divided by average total assets, this year against last year. Score a point if it fell.
- Change in current ratio: current assets divided by current liabilities. Score a point if it improved.
- Shares issued: common shares outstanding, this year against last year. Score a point if none were issued.
- Change in gross margin: gross profit divided by sales. Score a point if it improved.
- Change in asset turnover: sales divided by beginning total assets. Score a point if it improved.
Check yourself
Piotroski built the F-Score to sort companies...
- Across the whole market, cheap and expensive alike
- Inside the cheapest fifth of the market by book-to-market
- Only among companies carrying no debt
Show the answer
Answer: B.
His tests ran inside the highest book-to-market quintile, the fifth of the market priced cheapest against book value, not across every stock.Operating cash flow is 80 and net income is 100. Does the accrual test score a point?
- Yes, because both figures are positive
- No, because operating cash flow does not exceed net income
- Yes, because the gap is less than 25 percent
Show the answer
Answer: B.
The accrual test scores a point only when operating cash flow exceeds net income. Here it does not, so no point.In Piotroski's own study, how did high F-Score firms perform against the whole high book-to-market quintile they were drawn from?
- About the same, 5.9 percent against 5.9 percent
- Better: 13.4 percent against 5.9 percent, one year out
- Worse: high F-Score firms trailed the quintile
Show the answer
Answer: B.
High F-Score firms returned 13.4 percent against 5.9 percent for the entire high book-to-market quintile, and low F-Score firms returned -9.6 percent.Sources
- Piotroski, "Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers," Journal of Accounting Research, 2000.
What this leaves out
- Piotroski tested the score only inside the cheapest fifth of the market by book-to-market. A high score computed outside that group has not been tested the same way.
- The nine tests read one or two years of a company's own filed accounts. They say nothing about its price.
- The two-year table is an invented example, not any filed company's actual figures.
- Piotroski found the gains concentrated in small, thinly traded firms with no analyst following, and named possible data snooping himself.
- Nothing here is advice about any security.