Search ForexCrunch

If you had been trading stocks or bonds back in the 1980s and 1990s in London or New York, chances are you would have been a rough-and-ready market maker with little or no training in economics or mathematics. Ever since the Stock Market Crash of October 19th, 1987, that all changed.

This was the green light for the emergence of the algorithm trader, or “quant”.

Quants are essentially computer hacks, more comfortable around hard disks and BIOS than with legging their opponents over for a fraction of a percent of a million dollars. Their qualifications are more likely to be MSc or PhD, than whether they could flog ice cream to an Eskimo.

So, what exactly is algo trading? Essentially, it means taking financial decision making out of the hands of a human operator and into the hands of a computer programmer. It is true, computers will always be able to work more efficiently than humans, but is that enough? And how would it work if you developed the best-ever program to buy and sell stocks or bonds at the “best price”? Wouldn’t that just drain all liquidity from the market? To date, this has never happened, and chances are that it never will.

There are other, more technical issues. For example, anyone who has traded a bull, bear, or trendless market will know, they can go up, down, or sideways. If markets are trending strongly up or down, as they did during the 1987 Crash (shares went down, bonds went up), an algo trader employing trend-following techniques such as moving average crossovers would have made a killing.

However, and this is a big one”¦how can you tell if a market is trending up or down before it starts to trend? Who knew that markets were going to bounce or dive on that fateful day in October 1987? An honest answer is – nobody. And until someone creates a crystal ball that can tell the future, the ultimate market direction will continue to be the bugbear of any computerized trading system. Of course, there are indicators that “tell you” whether markets are trending, and in which direction, but the downside is that the markets need to trend in that direction for some time until the signal is available. That’s a bit like telling you which horse won after the race is over.

Up or downtrend-following software employs technical analysis tools such as the Moving Average Convergence Divergence (MACD) oscillator, or double moving average crossovers. Markets trending sideways work better with indicators such as Relative Strength Index (RSI) or Stochastics. Use the wrong one in the wrong type of market, then start looking for a new job.

Until algo traders can call a market trend correctly, in advance, then no computerized system is going to outperform the markets. As the non-quant, rough-and-ready traders said back in the 80s, “the trend is your friend”.