Yesterday's selloff came with an Arms Index reading of 2.85. We've only had seven higher daily readings since March, 2003. The Arms Index (named after inventor Richard Arms and also known as TRIN) measures the average volume per declining stock vs. the average volume per advancing stock. When the Index is greater than 1, more volume is flowing into declining issues and vice versa. At 2.85 on Monday, volume was certainly skewed to the sell side.
We've had 40 occasions of daily Arms Index readings above 2.0 since March, 2003. Three days later, SPY has been up by an average of .75% (26 up, 14 down). That is much stronger than the average three day change in SPY of .17% for the entire sample. We have had three readings above 2.0 in 2006 thus far; all three led to higher prices the following trading session.
Interestingly, we've had two days in the last five trading sessions in which the Arms Index has been above 2.0. Similar clustering of extreme readings occurred in 3/2003; 3/2004; and 8/2004--all favorable buying opportunities. During the prior bear market, such clusterings were much more frequent.
UPDATE: As most of you know, I do not base my day's trading strategy on any single analysis. I need to see multiple historical patterns pointing in the same direction. Readers of the Weblog are familiar with the Demand/Supply measures, which are indices of the number of stocks that are showing unusually strong positive vs. negative price momentum. I will be posting more on this measure in the near future. On Monday, we hit a very high Supply figure above 200, which has only occurred five times previous since March, 2003. All five occasions were down three days later in SPY; four of five were down the next day. Very strong downside momentum does tend to spill over into further weakness in the near term.