Don’t compare Close or EPS to a constant
A lesson on absolute prices versus relative ones
You’ve been referred to this page by something saying “You can’t (or shouldn’t) use Close > 5.00” in a stock screener.
Imagine you are looking at a share of GoodStock (yeah, I know that’s a lame name for a fake stock!). It goes up and up, and up! If it follows the example of Berkshire Hathaway, it soon sports a 5 or 6 digit closing price, and people can’t buy the stock. So, it splits, and splits again. On the other hand, BadStock starts at $10, and soon enough it has a price less than a dollar. So, it does a reverse split, gets its closing price up to $10 and goes on with its life.
So what does this mean for me? That you shouldn’t compare per share quantities (Close, Open, High, Low, Volume, EPS, Sales_per_share, etc versus constants. Close > Close as of 200 days ago is fine, but Close > 10 is not. Similarly, Close > Open is fine (they both have this history element included).
What should I do?
You can convert Close to RawClose, Open to RawOpen, etc. Or you can swap out your constant for something that factors the splits out. Here are some example problems and how you can fix them.
High prices stocks are terrible!
Type “Close > 444” into a stock screener, and observe the best short screen ever. Ever!
All the biotechs that reverse split again and again help push this down to new lows. Unless you’re psychic and can predict reverse splits and FDA failures, I wouldn’t try shorting this screen. If you swap the Close for RawClose (RawClose > 444), you get this.
Which is still nothing to write home about. But it’s not magical.
I want to get rid of penny stocks!
The next case this comes up is with the natural urge to get rid of illiquid penny stocks. Here you have the opposite problem: it looks like penny stocks sometimes outperform. What do do?
The obvious, “Jam the Radar” solution is to again use RawClose. But that does nothing for illiquid stocks. So we use Close * Volume > 200k instead. Why is this OK? Because volume goes down with reverse splits, even as the price goes up. The two effects cancel. Heck, why choose? Use both!
Large negative earnings
Sometimes a picture is worth a thousand words. Unfortunately, EPS is also split adjusted, and so we again are being psychic without meaning to.
We solve this one by looking for companies with high negative PE values (we have to compute them manually since the built in PE excludes money losing companies.
This may be the beginnings of a good short strategy, but it’s not magical. Luckily.
What is EPS 1Y? How are restated earnings handled? Is this a PIT database? Why show equity curves on uninvestable stocks (pink sheet, penny stocks)? Some of the posts here have merit, but the inconsistencies do not inspire confidence.
This is a PIT database (as much as anything is! There are some known flaws, such as that the industries stocks are assigned to do not update over time. If I were to find a historical source of industries, I’d fix that. Generally, we use as-reported earnings, and ignore restatements, and we use quarterly and annual financials, and create our own trailing twelve months from the quarterlies.
What are the inconsistencies that do not inspire confidence? It’s very hard to usefully address such a claim without more details.