Saturday, January 20, 2007

Swing Trading With Weekly Pivot Levels: Part One

In a recent posting, I explained how I utilize the pivot price levels originally developed by pit traders to establish potential price targets/exits for intraday trades. Basically, this method begins with historical patterns to anticipate next day price movement and then handicaps the odds of reaching target pivot levels by continually updating market demand and supply, as revealed by the number of stocks trading at bid vs. offer (Adjusted NYSE TICK) and the proportion of ES volume trading at bid vs. offer (Market Delta). At the request of readers, I am now publishing next-day pivot levels daily on the Trading Psychology Weblog. I'm also using morning comments on the stock indices to illustrate an updating of my thinking in real time; here's an example from Friday's trade. Note how I flipped my thinking from anticipating a downside test of the prior day's lows at 8:50 AM to leaning to the long side at 9:06 AM following the release of the Michigan sentiment numbers. After some choppy trade, the absence of selling finally emboldened the bulls to enter the market and we were able to hit profitable upside price targets.

Note that this kind of trading requires a very close monitoring of the market and an active processing of new information as it arrives. Many traders cannot follow the market this closely and may lack either the interest or the skill sets to process large volumes of incoming data. For those traders, might it be possible to develop swing-trading methods based upon weekly pivot levels? This post and the next one will explore this framework.

What this means is that, instead of tracking markets intensively and updating our views every few minutes, we would track the market at most a couple of times daily and update our odds of hitting price targets identified at the start of the trading week.

The most basic price target levels I track in my own trading are the high, low, and average trading price from the previous day. Let's take a look at how often we hit those price levels on a weekly basis.

Since 2004 (158 trading weeks) in the S&P 500 Index (SPY), we have exceeded the prior week's high 98 times and failed to hit that level 60 times. We have traded below the prior week's low 72 times and failed to take out the previous week's low 86 times. Clearly, these results are a reflection of the general bull market environment we've had over that time.

Interestingly, only 10 of the 158 trading weeks since 2004 have been inside weeks. That is, more than 90% of all weeks will either take out their previous week's high, low, or both. For all practical purposes, that means that we can assume at the start of a week that we will hit either the high or low. It is up to our historical patterns and market updating to help us handicap the odds of hitting one or another of those price targets.

One simple way to update demand vs. supply in the market is to track the proportion of stocks advancing vs. declining for the week. When we have had more weekly advancers than decliners since 2004 (N = 98 trading weeks), 74 of those occasions have taken out the prior week's high--about 75%. When we have had more than 2000 advancing NYSE stocks on the week (N = 75), 62 of those occasions--over 80%--have taken out the previous week's high.

On the other hand, when we've had more weekly declining stocks than advancers (N = 60), 43 of those occasions--about 70%--have taken out their prior week's lows. When we've had more than 2000 declining stocks in a trading week (N = 40), 34 of those occasions--almost 90%--have taken out their prior week's lows.

In short, by updating breadth numbers during the trading day and from day to day, we can update our odds of hitting either the prior week's high price or low price.

Here's another potential price target: if we compute the prior week's average trading price simply as the pivot level--the average of the week's high, low, and close prices--then we see that fully 147 of the 158 weeks touch their previous week's average price. This overlapping of prices from week to week is attributable to reversal effects on a daily basis. When markets move up, they tend to retrace some of that gain and vice versa. Thus, if we close away from the prior week's average price, we still tend to revisit it during the trading week.

As for historical patterns, here's a simple one: When we have exceeded the previous week's high (N = 98), the *next* week in SPY averages a gain of only .07% (55 up, 43 down). When we have failed to exceed the previous week's high (N = 60), the next week in SPY averages a gain of .32% (35 up, 25 down). And when we take out the previous week's low price (N = 70), the next week in SPY averages a whopping gain of .51% (44 up, 26 down). When we fail to take out the previous week's low (N = 88), the next week in SPY averages a loss of -.12% (46 up, 42 down). In short, simply knowing how the previous week traded can help you handicap the odds of taking out that prior week's highs or lows.

By focusing on different time frames, active traders of the stock indices can achieve a measure of diversification. Also, these methods can be easily applied to individual equities, updating their strength vs. weakness by such measures as on-balance volume or money flow. But how about pivot levels S1, S2, R1, and R2? Can these serve as meaningful weekly price targets? That will be the topic of my next post.