Thanks for a great post. I am trying to understand how the single number F-score is calculated from the various “elements”. The “Elements” are tests with boolean result True or False. For example “Return On Assets > 0” Suppose the result for ROA is True, then what? How much is True worth? How much is False worth for ROA? Then what does one do for all elements, sum up the results and average? Thanks again
In your example ROA > 0 is true is worth 1. ROA > 0 is false is worth 0. For all 9 tests, each true test is worth 1, each false test is worth zero. The final score is the sum of the tests. Or in other words, the F-Score equals the number of tests that are true. Consequently with 9 tests, the F-Score can range from 0 to 9.
Nice explanation of the F score! Quick question: does SR have a function where one can “backtest” the results of a screener like the F score mentioned in the article? Thanks
Yes – you can use the equation screener to select stocks by the Piotroski score they had in the past (e.g. 3 years a go, 5 yeas ago etc.)
It would be instructive to have a company’s Piotroski score followed by its sector’s composite score in parentheses in SR’s listings, e.g. for McDonald’s (as above) it would be: 6.0 (5.6).
Interesting. I quickly reviewed the stocks that came up 8/6/19 as a perfect 9 above from AIT – WINA and it looks like many of them had a short-term up move from Aug 9th but after that the number of ups/downs/side-ways were about the same. It’d be curious to follow their performance through the end of the year.