Skip to main content

John Reese is CEO of and Validea Capital, and portfolio manager for National Bank Consensus funds. Globe Investor has a distribution agreement with, a premium Canadian stock screen service. Try it.

Have you ever followed a recipe in a cookbook and thought, "I can make this better"?

Maybe you add a little more butter than is called for. Maybe you substitute asiago cheese for mozzarella, and throw in some cayenne pepper as well.

You might luck out, and your dish turns out great. But unless you're a gifted chef, you may well be disappointed. After all, if it's made its way into a cookbook, the recipe was likely created by someone who has a lot of experience and expertise, and who has honed the recipe through practice.

The kitchen isn't the only pace people tinker. In the stock market, many investors who use quantitative screens do it, too. And, quite often, they end up with a "stock stew" that leaves a bad taste in their mouths.

Here's what happens: An investor starts with a solid stock-picking strategy that has a proven long-term track record, either through real-time performance figures or back-testing. But then, the investor will tinker, perhaps changing the model's P/E requirement a bit, or switching one variable for another, or adding in a new metric. While these moves may seem logical, the fact is that the resulting strategy hasn't been tested.

Investors get off track in another way, as well. Rather than tinkering with the strategy, they veto some of the stocks it selects, usually because some negative news about the company makes them think the stock's too risky. Turmoil in the Middle East leads them to scratch an oil stock off the list, or new proposed financial regulations leads them to veto a bank stock.

Again, such decisions may seem logical. The problem is that if a particular negative story is catching your eye, the rest of the investing world probably sees it as well, meaning that the risk is already baked into the stock's price. In fact, the price probably includes an exaggerated sense of the risk, given how humans are prone to overreaction.

In one study, hedge-fund guru Joel Greenblatt found that over a two-year period investors who were able to pick and choose between stocks his formula approved of and pick the timing of their trades fared far worse than those who had their buying and selling done on automated fixed intervals, with no ability to veto picks. While the latter beat the market by 21.4 percentage points, the former actually lagged the market by about three points. One big reason: They tended to miss out on many of the best-performing stocks – beaten-down value plays that were the subject of scary headlines.

Bottom line: Tinker with a proven strategy and you just don't know what you are creating. Moreover, if you get lucky and your tinkering works in the short term, that can lead you to be overconfident going forward. That can lead you to make some big bets that go very badly.

Believing in a strategy is hard, to be sure. Even after years of quantitative, systematic investing, I still cringe when I see some of the names that my strategies turn up. But I stick to the strategy and buy the stock. I trust the numbers and reams of historical data, not my own fickle emotions.

Here are three stocks my models picked that made me cringe, but which I bought (and still own) anyway:


Russia's political situation is certainly cringe-worthy, but my James O'Shaughnessy-based value model likes this oil giant's size ($52-billion market cap), $14.82 in cash flow per share, and 5.4 per cent dividend.

USANA Health Sciences

Fellow multilevel marketer Herbalife's business model has come under much scrutiny, and this nutritional and personal care products firm has suffered, by guilt by association. But my Peter Lynch-, Warren Buffett-, and Joel Greenblatt-based models all like the stock. It has a 0.57 P/E-to-growth ratio, no long-term debt, and just one decline in annual earnings per share over the past decade.

Bed Bath & Beyond

Shares of this home-goods firm were hit hard after some disappointing recent results. But its valuation (0.61 PEG ratio) and balance sheet (no long-term debt) help it get high marks from my Lynch-based model.