Sounds exactly like the Dogs of the Dow.  Dividend yields are pretty conservative to where they have been in the past, some point to this as an indication of the stock market being overvalued.  Anytime anything is backtested becareful you run into data mining issues, as in you are finding correlation instead of causation.  If you run a set of data through enough screens you are bound to find some patterns, some that will continue in the future, and others that are flukes.

 Retireontime makes a good point, I am a big fan of dividends.  Why do, (or SHOULD) companies pay dividends?  Because the return on the cash when they invest is less than their cost of capital, i.e. they are essentially investing at a negative rate.  So you can look at it two ways, a company that doesn't pay dividends has great growth prospects, and the return on the cash they invest is going to be net positive.  OR The company has ignorant or worse yet, greedy management, that is going to chase returns to try to boost their stock price so they can kick in their options.  

Becareful with the Random Walk Down Wallstreet.  It was required reading in some of my finance classes, and as such I never read it.  However, from what I hear it pitches the idea of efficient markets.  If the current market has proved anything, it is the market is inefficient.  People act on imperfect knowledge and many times, act on emotion.  Some of the best advice the grandmaster has ever said "the stock market in the short run is a voting machine in the long run it is a scale".  Eventually things come back to reality.

 Hopefully this helps  and wasn't too lengthy.