Advance decline skew bodes well for the market
On Wednesday, advancing issues beat declining issues by a 4.8-to-1 ratio on the NYSE. It’s worth noting that such a positive skew is very rare.
Tony Dwyer of FTN Midwest Research decided to see what that ratio meant over the history of the S&P 500. Dwyer, working with Jason Goepfert of SentimenTrader.com, they observed that there's been
"54 occurrences since 1965 of NYSE advancing issues beating declining issues by 4.5-to-1 or greater ratio. If you strip out those high ratio days that were within 3 months of the first instance, there were only 29 Initial occurrences. Taking it a step further, we looked for only those initial occurrences in the context of a market trading above the 200-day moving average (most similar to Thursday’s occurrence). There were only 11 instances of such positive breadth when the SPX was above it’s widely followed long-term moving average. The following are the results of all three ways to study it and what took place 90 days after"
Using a benchmark of NYSE 4.5-to-1 A/D (or greater), they broke the stats down as follows:
Total Occurrences: 54 instances, with the S&P 500 (SPX) up 3 months later 77% of the time; Average gain = 8.3%. Note this includes closely located significant breadth days.
Total Initial Occurrences: Out of 29 instances, 69% of the time, the SPX was up 3 months later with an average gain of 6.4%.
Total above 200-day moving average Initial Occurrences: Out of 11 instances, 10 were positive and 1 was negative with an average gain of 8.3%, a maximum gain of 19%, and the one loss of 4.4% in the signal of 11/26/79.
All told, a very significant set of correlations. This is not unlike simlar price/volume studies done by Ned Davis Research in the 80s, and the 90/10 volume/price day work by Paul Desmond of Lowry Reports. (Discussed in our contrary indicators piece).
TrackBack URL for this entry:
Listed below are links to weblogs that reference Advance decline skew bodes well for the market :
Great analysis. I've read the Lowry analysis several times and from a big picture perspective, this (90/10, 4.5 to 1 ratio)is the best predictor.
Posted by: Joe K | Jan 25, 2008 7:25:33 PM
The comments to this entry are closed.