Tuesday, December 15, 2009

Understanding How Markets Move

In my recent post, I emphasized that *what* you trade matters as much as *how* you trade. The same system can perform well or abysmally depending upon the asset class or stock that you're trading.

A corollary of this is that we can learn a great deal about how a stock or index trades based upon its historical performance across a variety of systems and system parameters.

Let's take the S&P 500 Index (SPY) over the past five years.

If we bought rises above the 100-day moving average and then went short on moves below the average, we would have made a total of 56.6 points or about 56%. That is despite the fact that only 8 out of 46 trades were winners.

If, however, we traded the system with a 50-day moving average, we would have generated 98 trades with 17 winners. That would have lost us 10.75 points or over 10%.

Suppose that we now trade SPY with the same system, only using a 20-day moving average. Now we generate 166 trades, with 44 winners, but the system loses 23.05 points or over 20%.

Moving even shorter term, we can define a system that would buy on moves above short-term envelopes and sells on moves below those. Using envelopes created by high prices over the past 10 sessions and low prices over the past 8 sessions, that system traded 249 times over the past five years with 67 winners. That system would have lost essentially all the trader's money over five years.

What that says is that, if we're trading in a trend following mode, the more actively we trade (using shorter-term definitions of trend), the more money we lose. This is an important reason why short-term traders who have bought on strength and sold on weakness have performed poorly in recent years.

A big shout out to the Barchart site for the performance data.