Hindenberg Omen - In Effect

August 13, 2010 9:50 AM

Yesterday the conditions for the Hindenburg Omen, a combination of technical analysis indicators suggesting an increased probability of a market correction were all met (maybe) for the first time since 2008. Fortuitous timing for the soon to be released movie of the same name, if no one else. (Previously).
posted by BigSky (23 comments total)

chant down babylon
posted by kuatto at 9:53 AM on August 13, 2010 [3 favorites]

The previously tears this Omen a new one. Also, "sharpest decline since July 1"? These guys need to get a hobby and stop ruining lives.
posted by DU at 9:57 AM on August 13, 2010

These guys need to get a hobby and stop ruining lives.

Yeah, but then making money for themselves would be harder.
posted by ArgentCorvid at 10:12 AM on August 13, 2010

This is numerological nonsense.

Which is not to say that the market isn't going to suffer this fall; it very well might.

...soon to be released movie...

I don't think so. IMDB lists the 'release date' as Nov 2008, but there are no reviews, no media, no festival screenings, nothing. It looks like a low-budget vanity project that isn't going to get a distribution deal, ever. You might be able to buy a DVD some day.
posted by mr_roboto at 10:20 AM on August 13, 2010 [1 favorite]

The previously tears this Omen a new one.

In what sense? The "previously" post was made on July 7, 2008. In the thread, local finance expert Mutant said he hadn't seen any academic studies confirming its validity - which is not exactly a ringing endorsement, but is not a refutation either. In terms of false positives, it was observed that "Looking back at historical data, the probability of a move greater than 5% to the downside after a confirmed Hindenburg Omen was 77%, and usually takes place within the next forty-days. The probability of a panic sellout was 41% and the probability of a major stock market crash was 24%." (from Wikipedia, now as then)

And what happened in 2008? The market was flat for the next couple months and then dropped precipitously for two weeks (sorry if the table doesn't work - it's supposed to be weekly S&P quotes for late 2008). If anything, the omen was vindicated after the "previously" post!

Which, of course, is the problem with market prediction in general. Since you can't rewind history and change economic factors, it's pretty hard to sort out correlation and causation. But as far as correlation with a major stock downturn goes, the July 2008 sighting certainly did not "tear it a new one."
posted by rkent at 10:29 AM on August 13, 2010 [2 favorites]

I first learned about the Hindenburg Omen years ago here in the blue, and ever since I've been a strong follower of it. I think the name might scare some people off of taking it seriously, but it's a pretty basic model for market fundamentals, and it's really difficult for all the criteria to be met.

Long story short, if you're in the markets, now might be a good time to go short. Of course there are no guarantees in the marketplace, but there's never been a crash without a H.O.
posted by indiebass at 10:34 AM on August 13, 2010

From the Wikipedia entry:

The general rationale behind the indicator is that "under normal conditions" either

1.A substantial number of stocks set new annual highs
2.A substantial number of stocks set new annual lows
3.Conditions 1 & 2 cannot both take place at the same time, it is either one or the other—but not both
However, this indicator mainly tracks new lows and downside risk.

So... a declining market portends a declining market? Is this an oversimplification?
posted by Bummus at 10:39 AM on August 13, 2010

there's never been a crash without a H.O.

that's because all crashes lead to upticks in volatility, the criteria for the "Hindenburg Omen" are all volatility/dispersion driven.
posted by JPD at 10:40 AM on August 13, 2010

From the last post -

"The omen has appeared before all of the stock market crashes, or panic events, of the past 21 years", speaking about 1985 to 2006. Having a signal that can generate sharp market declines is appealing to all active traders, but this signal is not as common as most traders would hope. According to McHugh, the omen only created a signal on 160 separate days, or 3.2% of the approximate 5,000 days that he studied.


and that folks is all you need to know.
posted by JPD at 10:42 AM on August 13, 2010

There have not been 160 major events in the past 21 years. For most of those 21 years in that sample you would have been wrong to liquidate your positions based on this "omen"
posted by JPD at 10:43 AM on August 13, 2010

Bummus: it's not a declining market predicting a declining market... in fact, the third criterion requires the NYSE moving average to be rising, indicating rising prices.

The first criterion looks that there are BOTH very high and very low prices at the same time
("That the daily number of NYSE new 52 Week Highs and the daily number of new 52 Week Lows must both be greater than 2.2 percent of total NYSE issues traded that day.") which is looking for extreme volatility, which you extrapolate in a number of ways.

I'd say to JPD, that I take the H.O. to not be the sign that everyone should immediately liquidate their positions, but rather that it's time to really pay attention because "bad stuff is possible". Maybe more possible than it'd be without it. It's not a promise, but a warning. And honestly, people should be paying more attention more often, and if stuff like this helps focus people I think it's a good thing.
posted by indiebass at 10:55 AM on August 13, 2010

JPD: but doesn't anyone who credits the omen generally concur that you need to have a cluster of them to be meaningful? Even the Business Week article ("yesterday" link in the FPP) says:

The Hindenburg Omen must be confirmed with a second occurrence within 36 days, according to Riesner. He said the signal occurred seven times in 2008 as the S&P 500 posted its biggest annual drop since the Great Depression.

So presumably those 7 occurrences of the technical stats in 2008 (and I would further assume, concentrated in June-Aug 2008) count as a single "omen"? If they tend to cluster like that, then you could be down to 20-30 "events" in the last 21 years. And what counts as major is all in the eye of the beholder; I've always been somewhat confused by active traders' tendency to regard a 5% decline as a huge deal, but to some people, it is.

At any rate, I don't mean to really fight for this since I tend to agree that making major investment decisions based on the CNBC technical trading story du jour is not the wisest move.
posted by rkent at 11:03 AM on August 13, 2010

The problem with narrowing the criteria to fit the past better is that is all you are doing: predicting the past.
posted by smackfu at 11:19 AM on August 13, 2010 [2 favorites]

It's fire... and it's crashing! It's crashing terrible! Oh, my! Get out of the way, please! It's burning and bursting into flames
posted by sswiller at 11:36 AM on August 13, 2010 [2 favorites]

The NYSE is developing a head-and-shoulder reversal pattern, which is not yet fully verified, according to Guppy.

"It’s a rare double-right shoulder, but the same pattern as 1930," he said.

What? Is this some kind of Iron Cheffery of butchers? "While Claude here has prepared baby-back ribs with mango-lemon barbeque sauce, Bob presents a rare double right shoulder, so rare it's still mooing. Folks, we haven't seen anything like this since the great depression!"
posted by redfisch at 12:02 PM on August 13, 2010 [1 favorite]

but it's a pretty basic model for market fundamentals

Not to nitpick, but the articles indicate this "Omen" is based on the technical indicators, not market fundamentals. These two concepts couldn't be further apart, and it's a mistake to elide them in any way. Here's one non-technical discussion of how they differ.

Analysis of the technical indicators means only analysis of trends in pricing actions, positions held, etc., not analysis of market fundamentals. The fundamentals are factors more closely related to reality, while the technicals are concerned only with the perception of reality as captured indirectly through pricing actions that (according to the magical theory of technical analysis anyway) are somehow aligned to the underlying reality in a reliable and predictable way.

I still maintain that technical analysis is little more than the financial counterpart to phrenology, and most studies of the predictive power of these kinds of analysis seem to back that up, finding no better long term performance among traders who trade on technical indicators than one would expect to see from those trading on random luck alone.
posted by saulgoodman at 12:07 PM on August 13, 2010 [2 favorites]

Well if the market does crash in the next couple months, then that's all she wrote for the Obama administration. We'll be looking at two years of no legislation passing, Senate "investigations" of Obama's citizenship, and one impeachment proceeding after another. And then a Republican president in 2012 that will make GWB look like a wild-eyed Lefty. And that's even leaving out the plague, methane clathrates sublimating, the Yellowstone Supervolcano blowing, and the mega-tsunami from the Azores landslip wiping out the East Coast.

So yeah, given that we are all completely doomed and will be unlikely to even be alive two years from now, it may well be time to sell short.
posted by happyroach at 12:29 PM on August 13, 2010 [2 favorites]

Saulgoodman: you are correct, and I misspoke/typed, re: fundamentals/indicators. Thank you for clarifying!
posted by indiebass at 12:35 PM on August 13, 2010

I've already hidden a Coke way up in the mall vending machine for my as-yet unborn son to drink after the crash.
posted by benzenedream at 12:47 PM on August 13, 2010 [1 favorite]

There have not been 160 major events in the past 21 years. For most of those 21 years in that sample you would have been wrong to liquidate your positions based on this "omen"

Also, you can fit a pattern to historical data without actually discovering a model with real predictive value. If you gave me enough information about conditions of previous Kentucky Derby races, I could probably come up with a pattern that could predict the winner for the last 21 years 77% of the time based on some complex calculations involving the weather or some other random factor, but that doesn't mean that anyone should use that model to try to predict next year's winner. Or in an actual real-world example, look at the Washington Redskins presidential prediction myth, things like that only exist because there are enough data points to make up an illusory connection between to independent events. If you backtest a ton of random models, one is bound to match up pretty well just due to overfitting. Every model is going to do well on the backtests, that's why they are being proposed in the first place, the question is whether it will work tomorrow. And for most technical models, there is no valid reason for why they should work better than similar random models other than voodoo logic and historical performance.
posted by burnmp3s at 12:53 PM on August 13, 2010 [3 favorites]

People still put their money in the stock market?

Haven't we all figured out by now that it's Evil?
posted by mrgrimm at 1:19 PM on August 13, 2010

If you backtest a ton of random models, one is bound to match up pretty well just due to overfitting

Doing a whole lot of coin flip series also results in the multiple comparisons problem.
posted by benzenedream at 1:20 PM on August 13, 2010 [2 favorites]

That trailer looked awful!!  The producers need $100M and Pierce Brosnan as the lead guy explaining the casino/insurance industry to make this thing look like a feature.

Who is the writer? Andrew Kevin Walker or David Koepp? Who is the DP? They need to get some professional lighting going on and a steadicam, because it looks like something someone shot for their high school project back in 1985. Is the plot going to go supernatural or corporate? With a title like Hindenberg, it's hard to tell which way they're headed.

I won't hold my breath.
posted by vhsiv at 7:41 PM on August 13, 2010