The logic goes something like this: everyone is bearish and there is money on the sidelines to provide fuel for any rally. Therefore, it is time to get bullish. Wrong!

I actually heard something like this on CNBC - several times - on Thursday. In general, I would agree with the analyst's logic: we should be bullish when sentiment is bearish. This is a basic tenant of my own philosophy, and this notion is a key component of my Market Bias Timing System. I have written about these dynamics for over 3 years now.

However, I would argue that now is not the time to get bullish. Now is the time to play defense. So why am I whistling another tune?

Let me explain.

Based upon increasing bearish investor sentiment, the Market Bias Timing System, which is my market timing model that keeps me on the right side of the trend most of the time, gave the bullish "call" on January 11, 2008. Since that time, we have given the market the opportunity to bounce and it hasn't. In fact, it is breaking down. This is a failed signal, and failed signals can lead to accelerated price moves lower.

It is the old market adage: when the market doesn't do what you expect, look out.

Despite the bear market (which I had been writing about the last 6 months of 2007 and made official in January, 2008), I expected a bounce based upon increasing bearish sentiment. This was the high probability play 7 weeks ago. It hasn't materialized, so it is likely look out below.

Let me show you graphically my concern, and then I will present some data. Figure 1 (next page) is a weekly chart of the S&P500 with the Investors Intelligence Bull/ Bear data in the lower panel.

Focus on the gray oval in the lower panel. Here we see that the numbers of bearish (red line) advisers exceeds the bullish (green line) ones. This dynamic, which we would normally consider bullish for prices because of the excessive bearish sentiment, existed from April, 1974 to November, 1974.

Figure 1. Price Failure

Now focus on the green up arrow on the price graph. This shows a pivot low point. Prices came down. The market put in a very short bottom as prices pivoted upward. The bounce didn't last long, and after 3 short weeks, prices closed below the pivot low point. The low was eventually seen 14 weeks later and 29.5% lower despite the extremes in bearish sentiment. Ouch!!

So this is how I define a failed signal. Sentiment is bearish. The bounce, which makes the pivot low point, is anemic; while sentiment remains bearish, prices close below the pivot.

Now for some data. See the table in figure 2.

Figure 2. Table 1.

This table shows the 14 times since 1970 when a failed signal occurred in the S&P500. From 1990 to the present, I used the Investor Sentiment Composite Indicator to define extremes in bearish sentiment; from 1970 to 1990, I used the Investors Intelligence data solely, and I only considered those times when the bears exceeded the bulls in the data.

Columns A and B in the table tell you the starting and ending times that sentiment was bearish. Column C tells you what price the pivot low point occurred; column D tells where the low eventually happened (after the breakdown) and column E tells you how many weeks it took to get to that low after the break below the pivot point. Column F is the percent change between the pivot point and the eventual low.

The orange section of the table is from 1970 to 1981 and encompasses the last half of a secular bear market; the aqua section is from 1984 to 1998 and this period essentially encompasses a bull market; the pink section is post 2002.

What can we gleam from the 14 failed signals?

Pre bull market (the orange section), the failed signals produced greater losses and carried on for longer duration then the bull market period. The average difference from the pivot low to the actual low for the 1970 to 1981 period was 14.6%. In the 1984 to 1998 period, the average excursion to the low (following a failed signal) was 4.05%. All 14 failed signals traveled on average another 10% below the pivot point to the eventual lows.

Now it should be obvious that the secular bear market period produced potential losses over 3 times that seen during the bull market. As we contemplate what course to take over the next couple of weeks, we should be mindful that we are in bear market.

Now you could argue that the failed signals from the bull market period where just tremendous fake outs. After all, the failed signals of 1994 and 1998 eventually led to monstrous market runs.

While true, I think it is best to play defense here. I don't have any knowledge to know if this price failure will eventually be a fake out. Back in January, sentiment turned bearish and we expected the market to bounce. Then that was the high odds play. Fast forward 8 weeks and we have a price failure, and based upon this, I believe there is potential risk for downside acceleration in prices. While I am the first to be bullish when others are bearish, I think that time has passed and we should be entering capital preservation mode.

Our CNBC analysts are right. We should be bullish when the masses are bearish. However, I believe their "call" is too risky in light of the current price failure.

Our current pivot low on the S&P500 is at 1323.87 and on the NASDAQ it is at 2318.74. These now become resistance areas and weekly closes above these levels would suggest that the price failure is really a fake out.

To learn more about our quantitative and disciplined investment approach please visit www.thetechnicaltake.com and sign up for our free weekly newsletter and downloads.

Guy M. Lerner may be reached at guy@thetechnicaltake.com.