Monday, November 06, 2006

Testing The Market's Winds

Much of my morning routine consists of looking at a few core measures of market sentiment, participation, momentum, and strength and seeing what stands out in the recent market. Once I find something that stands out, I go back in time and see what happened following similar episodes. This gives me an initial idea of whether this historical pattern might be associated with a directional trading edge.

The historical pattern is just that. Such patterns do not last forever, and--at best--they can put probability on your side, not certainty.

I treat these patterns like a scientist treats a hypothesis: an idea that will have to be supported by further investigation before it is accepted. Each day in the market, in that sense, is a kind of laboratory experiment, either supporting or failing to support history's hypothesis.

This morning, one pattern that I'm looking at is four consecutive days of negative daily readings in the Adjusted NYSE TICK. The Adjusted TICK is quoted daily in the Weblog and reflects whether more stocks were trading at their offer prices or at their bids. This is an excellent short-term measure of sentiment, because it captures the willingness of buyers to pay up to acquire stocks and the willingness of sellers to bail out at market prices.

The first thing we see in the data is that four consecutive days of negative TICK (bearish sentiment) is relatively unusual. We've only seen 45 such occasions since the beginning of 2004 (N = 709). The next day in the S&P 500 Index (SPY), the market was up by an average of .29% (31 up, 14 down), much stronger than the average one-day gain of .03% (382 up, 327 down). Stated otherwise, the odds of an up day following a four-day period of persistent bearish sentiment have been better than 2:1.

Sometimes we'll see more than one historical pattern point toward the same general conclusion. That provides us with a bit more confidence in our hypothesis. Still, it is just a hypothesis. If today is going to be an up day from open to close, we need to see a net positive TICK. That means an abundance of high readings (above +500) and very few weak ones (below -500). If we see selling in the ES futures--traders hitting bids--but TICK staying relatively strong, I'll conclude that the selling is not spilling over into the broad market, and I'll be willing to go with my hypothesis on the long side.

Should selling in ES spill over to the broad market, I will entertain the notion that this market is not living up to its historical norms. That, too, is valuable information.

The good scientist is open minded. Carl Swenlin of the Decision Point service makes a fine point when he says that indicators are windsocks, not crystal balls. They tell you which way the market wind is blowing. All history can do is prepare you just a bit for those winds.