Skip to main content
Message Font: Serif | Sans-Serif
No. of Recommendations: 2
The "monthly" thing over your time period more or less just shows that there are no "January effects" and such.

But, have you taken each of your metrics and checked to see if the constants are correct?

For instance, you played with market caps of $200MM and $1000MM. Try adjusting it to $500MM and $2000MM. You used ROE >13. Try various other values (5, 10, 17).

I'm not surprised by your market cap trick by the way. But you may try cutting off the top and bottom. It is well known that midcaps tend to underperform both "small caps" and "large caps". So try again with say market cap between $200MM and $1BB or $2BB, and market cap in the range of $4BB+.

I suspect that your stock screen is biased towards catching recovering large caps for two reasons. First, you screen on P/BV and dividends. Book value is more meaningful in certain industries (manufacturing) but almost meaningless in others (financial, software). This tends to orient your screen towards capital-intensive businesses.

Second, more mature firms which tend to be older and larger as a group also tend to pay substantially higher dividends because they find it more difficult to spend retained earnings; plus they can expense many small projects which a smaller firm would have to capitalize. Dividend screens tend to pick up large caps.

Testing in isolation is also not always a good idea. It is a good acid test to verify that certain criteria are more or less useless. But I always try to run tests like this where I change one parameter but leave all the others fixed within a "blended test" as you call it.

Why did you cut off the stocks at 25 (TOP 25)? When you are backtesting a STRATEGY, don't bother with a cutoff. The idea is to try to test the strategy using the entire stock universe. During excessively bad years, there may be very few stocks meeting your criteria since items like quarterly growth tend to fail across the board when the economy takes a huge nose dive (2001/2002).

You are correct by the way that the reason you are not allowed to see the stock symbols is to keep Valueline happy that their proprietary data is not being distributed freely.

Two more criteria that I routinely use are a penny stock filter (price >$XX) and if possible (I can't remember if does this), a trading volume filter (knocking off poor liquidity stocks). It's been a while since I've run a backtest so I forget the exact criteria.

Once you have your backtested criteria, then you can saunter over to Reuter's stock screener and plug in the exact same criteria to use it as a real stock screen. Judging by the fact that you are looking for a large (25) number of stocks vs. a small group (<10), you are clearly looking to test a screen and not for a mechanical investing filter.

Now that you've more or less tested the strategy, there are two more things to look for. By varying the constants, you can widen/narrow the screen. The goal I always use with a stock screen is to simply throw out as many losers as possible while not throwing out any more winners than necessary. I'm looking for screens that give me 25-100 stocks which I can then go through and scrutinize to find a potential winner. This second manual pass is necessary because there are tons of criteria which can't be screened for currently. For instance: too many restatements, management not following through on their strategies, questionable off-balance-sheet games, members of the board also run the vendor/customer companies (undue influence problems).

Print the post  


What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.