Saturday, January 27, 2007

Will The Market Trend And How Far Might It Go?

In recent posts, I have described how I handicap the odds of directional moves in the equity indices. By defining pivot levels of support and resistance, we can then calculate the odds of hitting those levels based upon the unfolding real-time sentiment of large traders, as reflected in the number of stocks trading at bid vs. offer compared to a lookback average (Adjusted NYSE TICK) and the proportion of the day's volume trading at bid vs. offer.

While these measures tell us a good deal about the directionality of moves, they are not the best predictors of the *extent* of market movement. In my research, the standard deviation of morning NYSE TICK values correlates .48 with the size of the morning's trading range. That's a meaningful correlation, but we can do better.

Above, we see a scatterplot of the ES market from December 1st to the present (N = 36 trading days). Specifically, we're looking at the average number of trades placed in ES per minute during the morning and the average size of the morning's trading range as a percentage of the opening price. The upward slope of the chart is apparent: when we have more activity in the market (more trades being placed), we get a wider trading range. The correlation is .75.

Of course, most traders don't track, minute by minute, how many trades are being placed in their instruments. More often, they track total contract or share volume. As it turns out, the number of trades and the number of contracts traded in ES are very highly correlated: about .92. As a result, tracking real-time volume in the morning ES provides us with a .68 correlation with the morning's trading range, going back to the beginning of December. Indeed, if we look over the past 200 trading days in SPY, daily volume correlates with daily trading range by the same .75 level. In short, volume reveals volatility.

By updating volume every five minutes and comparing it to average volume as of that time of day, we can estimate the amount of movement that the market is likely to give us. Many times, if we get a breakout from a trading range during the morning, volume will expand, as trend followers jump on board the move and traders leaning the wrong way have to exit their positions. As volume picks up, we also can raise our estimates of likely movement for the day.

The two questions to be asking throughout the trading day are:

1) How directionally biased is the market compared with normal?
2) How active is the market compared with normal?

The first we gauge by comparing the current day's NYSE TICK readings with average readings over the past 20 trading sessions. The second we assess by comparing the current day's volume with average volume for that time of day over the past 20 trading sessions. What we're really asking is: Will this market be likely to trend, and how far is it likely to go? Question One helps us address whether a market will be range bound or trending; Question Two helps us establish price targets and trade exits. Much of the skill of trading consists of remaining flexible enough to update the relevant market data through the day and revise one's expectations.