Computational Stock-Picking

In today's confusing market, the best strategy may be to sit back and let the computer decide where to invest.

At least that's the approach Mike Mara, portfolio manager for the (PSRPX Quote)Penn Street Advisors Sector Rotational fund, has successfully taken since developing his stock-picking model at Vanguard a few years ago.

Mara takes the top 1,250 U.S. companies, with $1 billion or more market capitalization, and runs them through his quantitative model. He then ranks them based on valuation metrics such as absolute price to earnings, price to book, earnings surprise, growth and price trends.

Then he looks at the top 125 by pure count to determine in which sectors they fall. For instance, how many are technology, how many are financials and so on, and he sets his target weights for the sectors this way. Finally, he sets target weightings in 11 sectors, and selects the securities to fill the allotted spaces, being careful not to be redundant in his choices.

Lately Mara's model has favored the homebuilding sector including stocks like DR Horton(DHI Quote) and KB Home(KBH Quote).

His computer is also keen on energy companies like Valero(VLO Quote) and ConocoPhillips(COP Quote).

Of course, those sectors may shift with the next earnings period. But his model has been prescient thus far. Mara's fund is up 8% year to date, about four percentage points higher than the S&P 500.

Mara stopped by TheStreet.com's offices to chat about his mathematical methods for choosing stocks and discussed whether the impact of Hurricane Katrina will confuse the computer.

To view Gregg Greenberg's StreetWatch interview with Mara, click here.