Last week we looked at some new data from James O’Shaughnessy’s updated version of his classic book What Works on Wall Street. This week we’ll examine some more of O’Shaughnessy’s findings, including an interesting conclusion about which valuation metrics work best over the long haul.

In past editions of What Works on Wall Street, O’Shaughnessy found that, when it comes to valuation, the price/sales ratio (PSR) was the best predictor of future performance. In the book’s updated version, however, he taps into expanded historical data that leads him to a different conclusion. Using return data from 1964 through 2009, he looks at how stocks that were in the top decile based on a number of valuation metrics have fared historically. Here are the average annual compound returns for some of the more popular measures:

 Top Decile Based On

Enterprise Value/EBITDA — 16.58%

Price/earnings — 16.25%

Price/operating cash flow — 16.25%

Buyback Yield — 15.81%

Shareholder Yield — 15.56%

Price/book value — 14.53%

PSR — 14.49%

Dividend Yield — 13.30%

(A few definitions may be necessary here: Enterprise value is equal to common equity at market value, plus debt, minority interest, and preferred equity at market value, and minus associate company at market value and all cash and cash equivalents; EBITDA is earnings before interest, taxes, depreciation, and amortization; buyback yield is the change, percentage-wise, between the amount of shares outstanding now and a year ago; shareholder yield is buyback yield plus dividend yield.)

Of course, raw returns aren’t everything; risk is also a big factor to consider, so O’Shaughnessy looks at factors like standard deviation of returns, biggest annual declines, percentage of years in which returns were positive, and consistency of returns. One measure that takes into account both risk and return is the Sortino ratio. Here’s a look at how the various valuation metrics stacked up on that basis.

Top Decile Based On

Enterprise Value/EBITDA –0.50

Shareholder Yield — 0.47

Price/earnings — 0.46

Price/operating cash flow — 0.45

Buyback Yield — 0.44

Dividend Yield — 0.31

Price/book value — 0.30

PSR — 0.29

So, which valuation does O’Shaughnessy recommend? None, in a manner of speaking. As the data above shows, the higher-returning strategies aren’t always the best when you factor in risk. On top of that, different approaches fare better in different environments. So O’Shaughnessy decided to find out how stocks that had the best composite valuation based on a number of metrics fared.

He found that combining value metrics could produce even better results. An approach that incorporated the price/book, price/earnings, price/sales, EBITDA/EV, and price/cash flow ratios as well as buyback yield was a top performer, for example. To do this, he ranked all stocks by percentile in each of those categories, with those with the worst of a particular ratio (say, the PSR) getting 1, and those with the best getting 100. Then he simply added up the six scores, and split the results into deciles. He found that stocks in the top decile based on those six factors historically averaged a compound annual return of 17.30%, with a Sortino ratio of 57 — much better than any of the individual metrics.

O’Shaughnessy also lays out some of his broader investing advice in the updated version of his book. Here are some of his keys to success:

Always Use Strategies: “You’ll get nowhere buying stocks just because they have a great story,” O’Shaughnessy writes. These stocks usually come with huge price tags, and usually go down in flames. “You must avoid them,” he says. “Always think in terms of overall strategies and not individual stocks.”

Ignore the Short Term: “When you look only at how your investment portfolio has performed for the last quarter, year, and three- and five-year period, you are looking at a tiny snapshot of time,” he writes, saying that that snapshot can be very misleading. He recommends focusing on rolling returns vs. a benchmark over time.

Use Only Strategies Proven Over The Long Term: O’Shaughnessy says to make sure you use an approach that has proven its worth over several different market environments. Short-term performance, or even the fact that a strategy might make intuitive sense, are no match for a long-term track record. “Stocks change. Industries change,” he says. “But the underlying reasons certain stocks are good investments remain the same. Only the fullness of time reveals which are the most sound.”

Invest Consistently: “If you use even a mediocre strategy consistently, you’ll beat almost all investors who jump in and out of the market, change tactics in midstream, and forever second-guess their decisions,” O’Shaughnessy writes.

Always Bet With The Base Rate: “Base rates are essentially the odds of beating the market over the time period you plan to invest,” O’Shaughnessy says. “If you pay attention to the odds, you can put them on your side.”

Never Use The Riskiest Strategies: O’Shaughnessy says to always focus on strategies with the highest risk-adjusted returns.

Always Use More Than One Strategy: O’Shaughnessy says that combining strategies that focus on different types of stocks (i.e., growth, value, large-cap, small-cap, geographic regions) can allow you to do much better than the broader market while not taking on more risk.

Use Multifactor Models: “You should always make a stock pass several hurdles before investing in it,” O’Shaughnessy says.