Statistics + Trade

Two fascinating worlds finally meet (and make for a beautiful offspring)

Screening is not enough

Almost half a year ago (October 29th, 2011), a well-known Israeli finance site has published an article demonstrating the use of stocks screening: You choose some fundamental and/or technical parameters, and remain with the stocks that satisfy your screening conditions.

Here is the link to the article (in Hebrew), where the author demonstrated the screening process using the following conditions:

  • Market cap > $1B
  • Div. yield > 3%
  • Sales grew by >= 20% in the last 5 years
  • EPS grew by >= 20% in the last 5 years
  • P/B < 3
  • P/E < 15

The author himself stressed: (translated from Hebrew)

It’s important to clarify that the action we’ve performed is purely technical and does not evaluate the companies business nor the stocks behavior. In other words, the screening reviews the company’s history, and cannot tell the future…

And indeed, at hindsight, it wasn’t such a good idea to invest in those stocks, whose tickers and data is summarized in the following table: (close prices in USD; commissions and dividends are omitted to simplify the evaluation)

Name (ticker) October 31, 2011 March 23, 2012 Change in %
BlackRock (BLK) 157.79 201.11 27.5%
KKR Financial Holdings LLC (KFN) 8.35 9.25 10.8%
Vale (VALE) 25.41 22.85 -10.1%
Yanzhou Coal Mining Co. (YZC) 24.7 21.54 -12.8%
NYSE Euronext (NYX) 26.57 29.57 11.3%
Elbit Systems Ltd. (ESLT) 45.63 36.79 -19.4%

Take what you want from the data, but here’s a possible bottom line to the story:

Measure October 31, 2011 March 23, 2012 Change in %
Average change 1.2%
S&P500 1253.3 1397.11 11.5%

The bottom line: An absolute difference of ~10%. Those recommendations were badly beaten by the market. And in this case, as the numbers are close to 1.0, the ratio of returns: 1.012/1.115 = 0.9076 also implies that those recommendations were ~9% worse than the return on the S&P500.

Finally, a few final beliefs/opinions we wished to share:

  • Screening based on some voodoo parameters is a good place to start at, but it’s not enough.
  • If you think screening (or other method, for that sake) is enough, ask yourself: Has it ever performed well for you? Have you ever carefully analyzed its historical success? Do you have any reason to believe it will succeed well in the future?
  • Recall that it’s easy, in theory, to check for yourself whether a specific method is any good: Maintain a “virtual portfolio” and do as that method tells you to. Monitor its historical success over time. Don’t get too excited if it performs well (or bad) over a period of 1 week, for example, when you’re more into longer term investments.
  • Anyway, good luck, don’t trust anyone else with your money, and trade responsibly. You’re more than welcome to take a look at our analysis results, and we hope you’ll find our data interesting. Nate‘s data is available to you for free these days. Evaluate it for a while. If he alone impresses you, then you should really try our big cannons.

Leave a Reply

Your email address will not be published. Required fields are marked *