Recent research into growth investing by New York-based Bernstein Investment Research and Management has produced new tools to help analysts discern when general expectations for a prospective investment in a company are too low.

The approach, known as “aerodynamic gap investing,” tries to identify the attributes displayed by growth companies whose growth and stock prices will probably exceed consensus expectations.

The major obstacle to successful growth investing has always been regression to the mean, with most companies tending toward mediocrity over time, taking earnings with them as they go.

When it comes to growth companies, Bernstein’s research reveals that earnings increases of at least 10% a year are hard to sustain for even three years. Only 13% of growth companies have been able to do it, with 75% of growth companies experiencing a decline in earnings or an actual loss in at least one year out of three.

To focus the search for growth stars, Bernstein studied the performance of a universe of U.S. firms, big and small, over more than 20 years. It found Wall Street professionals’ forecasts of corporate growth were often off the mark.

In fact, the companies most analysts expected the least from — those in the bottom 20% for growth expectations — performed better than the 20% of companies analysts had heavily touted. This held true whether the forecast period was five years or even just one.

The explanation, Bernstein says, is that most forecasts simply extrapolate past results. Wall Street analysts tend to base future expectations on the status quo, thereby underestimating imminent change.

Even if the research suggests a company’s earnings growth is about to take off, analysts probably won’t go out on a limb. Instead, they will predict a small to moderate increase, often aiming too low, and so the company’s next earnings report is again a surprise, surpassing published forecasts. As often as not, Bernstein maintains, the subsequent earnings report contains yet another happy revelation. And so it goes, with each surprise a bit more predictable than the previous one.

As a result of this forecasting bias, one upward move will probably be followed by a chain of surprises in the same direction as the first, Bernstein maintains. Companies in its study that had not yet experienced a positive surprise subsequently announced one about 33% of the time.

The probability that such a company would issue another monthly positive revision was 48%. If the firm had six consecutive upward revisions, the odds increased to 72% that it would experience another. And after 11 consecutive monthly revisions, the chance of another one was more than 76%, the data suggest. Momentum clearly does count, Bernstein argues.

The firm also found that when a company was enjoying positive earnings surprises, stock performance tended to follow suit. And the more cheery revelations a firm posted over the period, the better its stock performed. According to Bernstein, companies that announced at least three positive surprises in a row outperformed a broad basket of stocks by almost three percentage points a year on average.

Bernstein posits that the earnings of such a firm would exhibit distinct characteristics relative to the broad stock market. Earnings growth would be positive, the earnings growth rate would be accelerating from, say, 10%-11%, and the pace of acceleration would increase measurably.

Bernstein then grouped sets of stocks by these characteristics and saw that the stocks returned almost two percentage points more than the market on an annualized basis as long as the chain of surprises was underway. Those stocks exhibiting none of the attributes underperformed by 1.2 percentage points.

For advisors looking to track successful firms, Bernstein suggests emphasizing these five criteria: an earnings growth rate increasing faster than in previous history; positive stock-price momentum in the previous 12 months; year-over-year earnings growth; increasing returns on equity; and increasing free cash-flow margins. IE