Friday, June 23, 2006

Trending Behavior and Short-Term Stock Index Performance

I'm in the process of studying short-term market trending and its impact on prices in the near term. My Power Measure, charted each day on the Trading Psychology Weblog, is a metric that captures the propensity of any market to behave in a trending way. It is scaled in such as way as to vary between +100 (perfect upside trending) and -100 (perfect downside trending), with zero representing the complete absence of trending.

Readings of +90 or greater, interestingly, are associated with stronger prices in the near term. Since March of this year (N = 1894 15-minute periods), we've had 98 occasions in which the Power Measure has exceeded +90. The market (SPY) six hours later has averaged a gain of .10% (57 up, 41 down). When the Power Measure has been positive (greater than +20) but less than +90, the next six hours in SPY have averaged a loss of -.12% (291 up, 385 down).

What this suggests is that vigorous bull trends tend to continue in the near term, but moderately positive ones are subject to reversal.

What I'm also seeing is a relationship between intraday volume and trendiness. When the trend is very strong and positive (greater than +90), but volume is relatively light (N = 49), the next six hours in SPY average a solid gain of .17%. When trendiness is strong to the upside and volume is relatively high (N = 50), the next six hours in SPY only averages a gain of .02%.

Similarly, when trendiness is strong and negative (less than -90), but volume is relatively light (N = 59), the next six hours in SPY average a gain of .06%. When trendiness is strong to the downside and volume is relatively heavy (N = 58), the next six hours in SPY average a loss of -.11%. High volume during a strong bullish trend appears to lead to short-term underperformance; high volume during a strong bearish trend appears to lead to further short-term bearish performance.

These are very preliminary findings, with much more study to come. The goal is to see when trends tend to continue and when they tend to reverse.

No comments: