Sunday, January 21, 2007

Swing Trading With Weekly Pivot Levels: Part Two

In the first posting of this series, I built upon the framework of trading daily pivot levels to explain how a swing trading method could be devised from weekly pivot numbers. This would enable those traders who approach trading part-time, as a recreational supplement to income, to develop a systematic approach to trading, while still freeing them for other work and family obligations. My recent article showed how the previous week's high, low, and average prices are reasonable price targets for such a swing approach. Now let's take a look at the pivot-derived resistance and support levels.

Recall that we are defining two levels of upside resistance and two of downside support. The basic pivot level (P) is the average of the prior week's high, low, and closing prices. R1, then, is computed as (P*2)-L, where L is the prior week's low price. R2 is defined as P+(H-L), where H is the prior week's high price. Similarly, S1 is computed as (P*2)-H and S2 is P-(H-L).

Going back to 2004 (N = 158 trading weeks) in the S&P 500 Index (SPY), we find that exactly half of those occasions (N = 79) surpass the R1 first level of resistance. When the weekly number of advancing stocks exceeds the number of decliners (N = 98), 68 of those occasions--about 70%-- surpass R1. When the number of weekly advancing stocks exceeds 2300 (N = 48)--a level at which advancers lead decliners by roughly 2:1 or better--fully 45 of those occasions exceed R1. Conversely, when we have fewer than 2000 advancing stocks on the week (N = 83), only 22 of those occasions--a little less than 30%--exceed R1. It thus makes sense to use the ongoing count of weekly advancing vs. declining stocks as one yardstick for handicapping the odds of hitting the R1 target.

Interestingly, only 31 weeks of the 158 since 2004 have surpassed the R2, second level of resistance. Even when weekly advancing stocks have led decliners (N = 98), only 30--or about 30%--of those occasions exceeded R2. When the weekly advancers have exceeded 2300 (N =48), exactly half of those occasions (N = 24) have risen beyond R2. My observation of occasions when we've hit R2 suggests that good market breadth (advancers vs. declines) is not enough to guarantee solid odds of reaching the target. In addition, there must be above average volatility and strong market momentum (many stocks trading above their short-term moving averages). It thus makes sense to rely on prior week highs and R1 as price targets, only holding positions for R2 in markets that are showing very bullish momentum and breadth.

Now let's take a look at the S1 and S2 levels of support. Overall, only 63 of the 158 trading weeks trade below the S1 level. When advancing stocks lead decliners on a weekly basis (N = 98), only 15 of those occasions--about 15%--ever trade below S1. On the other hand, when decliners lead advancers on a weekly basis (N = 60), 48 of those occasions--80%--trade below S1. Clearly, the strength of advancers vs. decliners is displaying some promise in identifying the likelihood of hitting our first downside pivot price target.

As you might guess, not too many of the weekly occasions trade below S2: only 26 of the 158 weeks. When advancers have led decliners (N = 98), only 4 of those occasions traded below S2--about 4%. When declines have led advancers (N = 60), 22 of those occasions surpassed S2 to the downside--a bit less than 40%. When declining stocks have exceeded 2300 on a weekly basis, meaning that they outnumber advances by about 2:1 or more (N = 19), 12 of those occasions--a little more than 60%--hit S2. It thus appears, as with R2, that it takes more than negative breadth to hit S2; we also need above average volatility and downside momentum.

Finally, does hitting a pivot level tell us anything about the *next* week's trade? When we trade below S1 (N = 63), the next week in SPY averages a gain of .43% (40 up, 23 down). When we don't hit S1 (N = 95), the next week in SPY averages a loss of -.02% (50 up, 45 down). Such tendencies toward reversal are useful in anticipating the next week's likelihood of hitting initial price targets. Additional indicators of volatility and momentum are needed to reliably handicap the odds of hitting the more distant R2 and S2 targets. I used advancing vs. declining stocks simply because it is an easily available measure of market strength. It is very likely that other measures--such as the Demand/Supply momentum measure tracked by the Weblog--will perform even better. Clearly this is work in progress.


MidKnight said...

Hi Brett,

Excellent findings and I'm inspired to do some work in the European markets around these ideas for the DAX and DJ Euro Stoxx 50.

This is interesting, but I am confused how you are doing the weekly advancers vs. decliners. Is this merely a daily summation of the end of day advancers vs. decliners? Sadly, in the European markets I haven't found these values all summed up like you can get with USA markets. But I can forward test it out day by day and start to build my own databank in excel.

When you say "When the weekly number of advancing stocks exceeds the number of decliners" is this based on the end weekly value or is this based on the intra-week value? I'm trying to understand if this information is timely enough to make a decision upon or if it is close of weekly bar type information.

Thanks in advance.

Best regards,

Brett Steenbarger, Ph.D. said...

Hi Matt,

I'm using the number of advancers vs. decliners for the entire week, not the composite of the daily advances and declines. Tracking the latter is definitely helpful, but it's how the week as a whole is shaping up that is critical. I track a basket of stocks from the Friday close to the present for this purpose, and it's quite helpful.