The Big Picture

« Fall On Me | Main | Martin Feldstein on the Housing/Credit/Economic Mess »

One Way Days: Are the rules different this time ?

Wednesday, September 12, 2007 | 07:08 AM

That's the intriguing question asked in a recent WSJ article, New Rules For Picking A Bottom?

Don't leap to the wrong conclusion. The phrase "Its different this time," in its most expensive permutation, refers to rationales why an unsustainable trend will continue, despite obvious risks.

The subject presently at hand is somewhat different: Why are usually reliable technical/quantitative signals failing to have much forecasting prescience?

One of the attractions of quantitative decision making is its attempt to bypass our inherent weaknesses. The way Humans developed has led to a great deal of fallible interaction with capital markets. Error prone decision making is hard-wired into our wetware. (See You just ain't built for it). We bring a lot of baggage to investing, courtesy of a few million years of evolution.

So it is always fascinating when decision making processes designed to bypass this weakness suddenly stops working. And we have seen several examples of this as of late:

90/90 Days: When 90% of volume is in the same direction, and 90% of price moves are also, you get a "One Way Day." These usually are very bullish signals.

"When stocks are approaching the end of a decline, investors tend to be in a panic, and their sell orders dominate trading. Then, once the selling runs its course, bullish investors step in with heavy buy orders that dominate trading and, in turn, signal the beginning of a rally.

Lately, that combination of heavy selling followed by heavy buying is exactly what the market has seen -- on steroids.

"We have been getting these days at the rate of one every 3½ days, and that's just crazy," says Paul Desmond, president of research service Lowry's Reports in North Palm Beach, Fla., who has done extensive research on the subject. "We don't have anything like that anywhere in our history" of data, going back to 1933, he says."

Other independent research shops have had related problems. Ned Davis Research tracks a variation which they call nine-to-one days (trading volume only). The problem is that the huge uptick in volatility has wreaked havoc with these signals. Because there were too many nine-to-one days, Ned Davis simply raised their 9-to-one threshhold to 10-to-one days.

This isn't the first time I've seen this: From 2001 thru 2003, the usually reliable Arms Index simply stopped being a good timing signal for buys. Dick Arms re-jiggered it, placing the basic index into an oscillating framework. Like Ned Davis' approach, this eliminated the previously rare but suddenly all too common signals. The weaker "false" signals were eliminated.

What is different this time is that 2 trillion dollars worth of fast money is in the hands of active hedge funds. Failing to adapt to that could be quite expensive for traders . . .

Sources:
New Rules For Picking A Bottom?
E.S. BROWNING
WSJ, September 10, 2007; Page C1
http://online.wsj.com/article/SB118937309413321829.html

Fear the Roller Coaster? Embrace It
DENNIS K. BERMAN
WSJ, September 11, 2007; Page C1 (THE GAME)
http://online.wsj.com/article/SB118947349416123314.html

Investors' View Of Risk Returns To Normal
Justin Lahart
WSJ, September 10, 2007; Page C1
http://online.wsj.com/article/SB118938618650822150.html

You just ain't built for it
Apprenticed Investor: Know Thyself
Barry Ritholtz
RealMoney.com, 5/3/2005 10:20 AM EDT
http://www.thestreet.com/_tscs/comment/barryritholtz/10221284.html