Monday, October 03, 2005

The Hindenburg Omen

I noticed there is a bit of chatter on the web about something called the "Hindenburg Omen". The Hindenburg Omen is a technical indicator that supposedly predicts crashes. Or more specifically, when the Hindenburg Omen is triggered, supposedly there is a high probability of a market crash in the next 30 days. This website explains how to compute the Hindenburg Omen, but it doesn't give any specific information for us to know what the past performance of this indicator has been. Specifically I wanted to know how many times this has happened in the past, and what was the average return of the S&P 30 days later. I Googled "Hindenburg Omen" and found lots of sites talking about this omen. Most seem to agree this is a "very bad sign", but amazingly almost none of the people who've discussed this indicator seem to have bothered to check if it has actually been a bad sign in the past. I would backtest it myself but I don't have the necessary data (or time). I was just about to give up, when I found this site which has a backtest of the "Hindenburg Omen" back to 1965.

The backtest has 35 instances of the Hindenburg Omen. The average return is -0.7%, so that may indeed indicate a negative tendacy after a Hindenburg Omen has been triggered. The Omen has indeed been triggered before several major market downturns, but it has also preceeded several major upturns as well. The author of the website seems to be saying that although the Hindenburg Omen doesn't necessarily portend a market crash, it might be predictive of volatility since the standard deviation is 8.5%:

...thought it would be instructive to include the standard deviations of returns, which are large. What does that tell us? It shows that while there may not be a distinct bias to S&P performance after these signals, the market does tend to move well in one direction or the other.

I pulled his numbers for the 30 day return after a Hindenburg Omen into Excel and I took the absolute value of the returns. The average 30 day move after a Hindenburg Omen has been 13%. This suggests that maybe it might be a good idea to buy some out of the money calls and puts when the Omen is triggered.

Here's the thing though. I kind of suspect there may be a little bit of curve-fitting happening here. Here's what I think may happen. Naturally the time periods prior to a crashes get heavily scrutinized by technicians after the fact. Just by shear coincidence there are bound to be some set of indicators that taken match up in some way across many of the pre-crash time periods scrutinized ( This is curve fitting). Since the Hindenburg Omen was most likely created after the fact to explain crashes that had allready happened, the sample space is biased because the time period being analyzed includes all of those crashes that we were looking at in the first place. Or more simply, the Hindenburg Omen probably has preceded volatility because the person who "discovered" the omen was looking at volatile times when he made the "discovery".

The most recent Hindenburg Omen was triggered 2 days a couple of weeks ago on September 20th and then the next day on the 21st. October options expire 30 days later on October 21. SPY options 6% out of the money are selling for around 10 to 15 cents. I think the Hindenburg Omen is probably bullshit, but I might just buy a few of the 116/127 combination just for th hell of it anyways.


Post a Comment

Links to this post:

Create a Link

<< Home