Our protagonist, the first machine that can feel love, manages to escape when security guards mistake him for a human.

In the investment world, it may be possible for a machine to dupe technical analysts as easily as those guards were fooled, or so believes Joshua Reich, a researcher for Gilder Gagnon Howe, a brokerage in New York. One weekend this past winter, Reich, the type of person who makes calculators in his spare time, put aside his efforts to build the perfect calculator (one with “all the bells and whistles of modern technology”) and turned his attention to another project: debunking the notion that technical analysis, or chartistry (no relation), has any validity.

Technical analysis has never enjoyed much support from academia or the proponents of f undamental analysis. A technical analyst’s methods are often seen as the voodoo of the investment world and have been compared with fortune-telling. Unlike analysts who look at fundamentals—a company’s sales trends and earnings, the underlying economic conditions—technicians ignore such data and trade only on the basis of patterns they find in the charts of stocks, bonds, currencies, and any number of other securities.

While some technical-analysis tools, such as a stock’s trading volume, price trend, or moving average, have become part of the CNBC lexicon, the trading strategies derived from these can seem as strange as their names: Bullish Homing Pigeon, Death Cross, Fibonacci Time Zones, Sushi Roll, and Jennifer Lopez, to list just a few.

Nevertheless, John Brooks, a senior analyst with Lowry’s Reports, which bills itself as the oldest continuously published technical-investment advisory in the U.S., estimates that about 500 technical analysts currently work at major American investment banks and that thousands of money managers use technical analysis as part of their investment strategies.

The craft has also been incorporated into algorithmic trading where it has prominent practitioners including hedge fund king James Simons and futures trader and Boston Red Sox owner John W. Henry.

Brooks argues that technical indicators can reveal changes in the supply-and-demand relationship for an asset before the data used in fundamental analyses are released. “One of the kinks with fundamental analysis is that you’re looking at a trend going up and making projections based on past information,” Brooks says. “You’re not going to know that there is any change in that flow of information until a report comes out. With technical analysis, you see a break in the trend line and a red flag goes up.”

Technical analysis has lately become more widespread, as computer programs have allowed individual investors to try their hand at it. It’s particularly popular in currency-trading markets. Reuters, for example, publishes daily notes from technical analysts. (Full disclosure: I was formerly a Reuters employee.) Here’s a recent jargon-filled sample on trading between the U.S. dollar and the British pound from an analyst at J.P. Morgan: “Yesterday’s rally on the back of the BoE minutes developed a clear impulsive wave structure, while suggesting additional upside is likely.”

To test the technical analysts’ assumptions behind the gobbledygook, Reich created Stock or Not, an online game that challenges players to decide which of two stock charts is real. The real charts are taken from Yahoo Finance; the fake ones are generated by a model Reich has devised.

From a historical perspective, Reich’s game pits Charles Dow, the co-founder of Dow Jones and the grandfather of modern technical analysis, against Eugene Fama, the architect of the efficient-market hypothesis—the belief that the price of an asset always already reflects the available information, leaving no room for arbitrage.

Fama’s hypothesis comes in three flavors: strong, semistrong, and weak. In the strong form, no amount of insight or insider information will help an investor beat the market, because all assets are accurately priced. In the semistrong form, prices reflect all available public information, but insiders still have an advantage. And in the weak form, careful review of public information—fundamental analysis—can uncover undervalued assets and lead to higher returns. No version of the hypothesis, however, allows for the legitimacy of technical analysis. 

Reich, who has a background in mathematics and medicine, generated charts in line with the weak form of the efficient-market hypothesis and launched his game on his personal website. He says he chose the test for its pure simplicity: If technical analysis is legitimate, it should be able to uncover the absence of trends in fake charts relying entirely on an efficient-market model. A more obvious test, perhaps, would be to have users predict the future price moves of charts, since technical analysts can supposedly discern future price movements by studying past patterns. But Reich says he wanted to ram his point home with something simpler. Unfortunately for him, it didn’t quite work out that way.

Buoyed by blogosphere word of mouth, more than 120,000 games of Stock or Not have been played so far. If technical analysis is mere guesswork, then that would suggest that the probability of a player choosing correctly should be the same as the likelihood of a coin toss landing heads: 50 percent. But surprisingly, players so far have chosen the real chart a statistically significant 53 percent of the time. That might not seem like a great victory for technical analysis, but imagine if your stock picks were right that often. With enough money invested over a long-enough time period, you could make a killing.

Reich says that people who have done well on Stock or Not have had trouble explaining how they do it. While he believes that his fake charts are consistent with the efficient-market hypothesis, he suspects there might be tells in the way he designs them. “I’ve tried looking into it,” he says, “but my sense is that it’s my calibration process.”

Incidentally, academic studies of technical analysis have yielded mixed results: Many have found that its techniques can provide market-beating results, but they’ve also shown that once the added risk of using technical analysis is accounted for, the gains are wiped away. And other studies have shown that even if you don’t account for the additional risk, gains offered by technical trading techniques have largely disappeared since the 1990s.

And they have one main explanation for that: the rise of the machines.