I have no experience with the Hindenburg Omen, but from this chart, it hardly looks reliable.
Its had some terrific calls — 1987 in particular — but lots of false positives also.
Does anyone have anymore specific experience with this indicator than I? Please explain in the comments — good, bad or indifferent.
UPDATE II May 1 2006 3:14pm
Here’s an analysis of the Hindenburg Omen by Kennedy Gammage; He gives it more credit than I do . . .
UPDATE April 25 2006 4:14pm
John Hussman discusses the prior Hindenburg Omens:
Hindenburgs: “I’ve noted often that a great deal of the information conveyed by markets is contained in “divergences” between securities. While investors shouldn’t read too much into any indicator, there’s an interesting signal that has enough validity as a measure of divergence that it’s worth mentioning here. Think of it as slightly more than entertainment value but far less than a reliable guide to investment.
The signal is based on new highs and new lows, and is cheerfully called a Hindenburg (the actual name given to it by Kennedy Gammage is the “Hindenburg Omen” but that strikes me as far too, well, ominous, because it’s certainly not a sufficient condition for a market decline). It’s a relatively unusual event that has often preceded fairly substantial market declines with a fairly short lead time (usually within 30-60 days, including declines in 1987, 1990, 1998, 2000 and 2001), but has sometimes proved to be meaningless or insignificant as well (such as a cluster of signals in September 2005, among others).
The basic elements are:
1) the market is in a rising trend, defined as the NYSE Composite being above its 10-week average;
2) both daily new highs and new lows exceed 2.2% of issues traded, and
3) the McClellan Oscillator is negative – meaning that market breadth as measured by advances and declines is relatively weak (there’s some dispute, which I will not join, as to whether the Oscillator has to be negative that day or turn negative later). Peter Eliades added a couple of other conditions to eliminate signals occurring in clearly strong markets:
4) new highs can’t exceed new lows by more than 2-to-1, and
5) 2 or more signals occur within about a month (he uses 36 days) of each other.
As it happens, we observed a Hindenburg on April 7th (just 2 days after the market high) and another one on April 10, so those elements seem to be in place here.
The April 2006 Hindenburg Omen Has Now Been Confirmed
Monday, April 24, 2006 7:20 GMT
Main Line Investors