Commentary

Anne Plested
Traders Magazine Online News

Bottlenecks Ahead

Anne Plested, head of Fidessa's EU Regulation Change programme, has written a short blog arguing that although we should be thankful that ESMA have taken a pragmatic approach to moving things along, more bottlenecks could appear in the future.

Traders Poll

Would you feel better if the Chicago Stock Exchange were purchased by U.S. firm or consortium rather than a foreign one?




Free Site Registration

July 31, 2004

Algorithms Made Simple

By Ingrid Eisenstadter

Sam (Samuel) Eisenstadt, chairman of research for Value Line's Investment Survey, is one of Wall Street's all-time great stock pickers. That's because Eisenstadt, with a 40-year record and an early user of algorithms, has superb numbers over a long period.

Despite intermittent down years, Value Line has left all major market indexes in the dust: Since 1965, when the algorithm in use today was created, Value Line's recommended stocks have yielded (excluding dividends) a 46,060-percent gain, or 17 percent a year during the same period. That compares to 1,044 percent for the DJIA and 1,194 percent for the S&P 500, or 6.4 percent and 6.8 percent a year, respectively. How's that for an efficient market?

Early in the 1970s, Eisenstadt, an avid chess player, started beating the market. But then a prominent skeptic, University of Chicago's Fisher Black of Black Sholes Option Model fame, spent months at Value Line studying the company's stock-picking formula.

"He was from the enemy camp," Eisenstadt explains, "a mathematical economist who was an efficient market' theorist." Along with many academics, Black believed it was impossible to beat the market consistently.

When his lengthy analysis was completed, however, he reached this conclusion: It "appears that most investment management organizations would improve their performance if they fired all but one of their analysts and then provided the remaining analyst with the Value Line service," according to Black.

Today, Wall Street is awash in stock-picking algorithms, but Eisenstadt's was the first one widely available to the public, and may well have been the first successful algorithm altogether. Mark Hulbert's monthly Financial Digest tracks the performance of more than 150 market newsletters and the 500 portfolios they recommend. He tells Traders Magazine, "We've tracked newsletters for 24 years and Value Line is in first place on a risk-adjusted basis." That means higher performance with lower volatility than all other newsletters Hulbert analyzes.

In 1931, Arnold Bernhard, a former Moody's analyst, founded Value Line and began publishing his weekly picks. Bernhard had no mathematics training and made his buy-sell recommendations based on visual examination of price and earnings charts.

The company was in business for 15 years when Eisenstadt, a soldier who survived the World War II Battle of the Ruhr, returned home. To no job. It was 1946 when, armed with a statistics degree, after a long and fruitless search, Eisenstadt answered a Value Line want-ad.

So, was the ad for a statistician? Had Bernhard decided to graduate from visual interpretations of stock charts to mathematical analysis? Why, no. He had not. The ad Eisenstadt answered was for a copyboy. A proofreader.

"I took the job because at least I would be in an environment that was around numbers, even if my job would not make me part of them," Eisenstadt says. Within two years, however, he had indeed persuaded Bernhard to use statistical, not visual, data to make his recommendations. Results quickly improved.

By 1950, Eisenstadt was chief statistician. "By then we were pounding away on calculators," he recalls. "We analyzed four variables for each stock, and that took one day per company."

The labor-intensive work limited the survey to several hundred stocks under year-round review. Each stock was compared only to its own history. It was not until 1965 that the survey took a quantum leap and began comparing all stocks to each other to produce the current ranking method. Today, millions of calculations each week are behind the system, which covers 1,700 stocks.

Eisenstadt continues to fine-tune his formula, "which is always a work in progress," he explains. When not at his desk at work, he's at his desk at home in Brooklyn, New York.

"I play a lot of chess on the computer to keep my brain cells active. Unfortunately, I find the computer a formidable competitor" he adds, "but I can always shut it off when I make stupid mistakes."