In my August 3 article, Market Oversold and Dangerous, I pointed out that certain indicators had reached very oversold levels from which rallies normally emerged. I also warned about trying to pick a bottom because oversold conditions can also set up more selling. As it turned out, the market did not bounce out of those oversold lows. Instead, it consolidated for a short period prior to initiating another severe down leg.

On Thursday the market experienced extreme panic selling on high volume, but it completed an upside reversal into the close. This set up some positive divergences -- on the chart below note the higher indicator lows compared to lower price lows -- and has gotten sentiment swinging quickly back to the bullish side.

The next thing that most people are expecting is a retest of the recent lows, and I happen to agree with that view; however, I am worried by the market's distinctly bearish behavior during the recent decline. By that I mean, instead of a rally, we got a consolidation followed by a decline after the very oversold readings of two weeks ago. Because of this, we should be alert to the possibility that the expected retest will actually be another down leg to much lower lows.

While there has been a lot of short-term and medium-term technical damage done, the long-term bullish picture remains intact. Note on the chart below that the S&P 500 remains inside the rising trend channel and above the rising trend line.

There are a lot of crazy ways that this situation can play out, and I plan to rely on our Thrust/Trend Model for most of my decision input. On the next chart all the components of the model are displayed. Note that a sell signal was generated when the 20-EMA crossed down through the 50-EMA. This put the model in neutral because the 50-EMA was above the 200-EMA at the time. A new buy signal will be generated when the Percent Buy Index (PBI) crosses above its 32-EMA, AND the PMO (Price Momentum Oscillator) crosses above its 10-EMA. As you can see, this will take a lot of work on the part of the market, but it should be worth the wait, considering the risk we currently face. Also, there is no guarantee that a buy signal will be profitable, but it does give us assurance that market internals have firmed sufficiently to justify optimism.

Changing the subject, I want to briefly discuss the recent change in the "up tick rule". Back in the 1930s the S.E.C. established the requirement that a short sale could only occur on an up tick. At the beginning of July this requirement was withdrawn. In my opinion, this rule change is primarily responsible for the increased volatility we have experienced, and probably accounts for the severity of the recent decline.

Keep in mind that this is a two-edged sword. While short-sellers can exacerbate a decline by selling into it, this will result in a larger number of short-sellers available to be squeezed when the buyers move back into the market. I think this explains the wild swings we have recently experienced. Whether or not you agree with the rule change, it is now a reality that must be assessed and dealt with.

Bottom Line: It is possible for the market to continue to rally without a retest taking place; however, "spike" or "V" bottoms are uncommon, and the most likely outcome will be another decline to test the recent lows. Our Thrust /Trend Model currently has us neutral in all major indexes and sectors we track. We will wait for the model to generate buy signals before re-entering the market.

Regardless of my personal opinion, we rely on the mechanical trend models to determine our market posture. Below is a recent snapshot of our primary trend-following timing model status for the major indexes and sectors we track. Note that we have added the nine Rydex Equal Weight ETF versions of the S&P Spider Sectors. This may seem redundant, but the equal weighted indexes most often do not perform the same as their cap-weighted counterparts.

Technical analysis is a windsock, not a crystal ball. Be prepared to adjust your tactics and strategy if conditions change.