I’m seeing some evidence suggesting the market has quickly become overdone to the upside and is likely to pull back. Below is one study from last night’s subscriber letter that looks at other times the SPX made short-term highs on strong breadth but weak volume while in a long-term downtrend.

Instances are very low and prior to 1992 there did not appear to be an edge. So I wouldn’t try and extract too much from these results but I do think the recent downward strength and consistency is notable.

September has a well-earned reputation as the worst month of the year for the stock market. But with August performing so poorly I wondered how that might affect September’s performance. The SPX finished August down 4.7%. Below I conducted a study showing how September has performed following August returns worse than X%.

While we’re not looking at a huge number of instances it does appear that going into September on a bad note in the past has not alleviated any risk. The worse August was the worse September was. Below I’ve listed the 7 instances where, like now, August finished down over 4%.

What stands out to me here is the size of the drawdowns. Four of seven instances saw September swoons of over 8%. While not statistically significant, I do think this is worth considering.

For those interested in reading more about September tendencies, I’d suggest checking out the below MarketSci post from yesterday.

https://marketsci.wordpress.com/2010/08/31/ugly-september/

One indicator I thought it was worth taking a closer look at this weekend was the AAII Investor Sentiment survey. In general the survey is viewed by technicians as a contrary indicator when it reaches extremes.

This past week the number of bears rose to 49.5% and the bulls dropped to 20.7% so it is now at fairly extreme levels. The last time the Bull-Bear Spread was this low was in early July as the July rally was just beginning. Prior to that is was the first week of November of ’09 just as that rally was kicking off, and the time before that was early March of ’09 just before that rally began.

I looked at the data a number of different ways this weekend. (For those interested you may get all the data using the AAII link I provided above. Just scroll down and you’ll find a link to a historical spreadsheet on the right hand side of the page.) Below is a sample if the kind of results I saw when conducting some studies.

So what we see here is that an extremely pessimistic outlook from investors has been followed by a rise in the market on pretty consistent basis. Six weeks later the market has been higher 81% of the time and the average gain was over 7%. The problem is that the failures have been very, very large. Using the same 6-week time frame the average loss was 10% and the max loss was 27%. The max loss occurred in September/October of 2008. August of 1990 also saw a sharp decline in the spread to levels similar to current levels. That was followed by further selling that maxed out around a 12% decline between then and October. Other instances that were followed by large selloffs included July of 2002 with a 12% decline and February 2009 with a 14% decline.

So it appears the AAII Investment Survey is suggesting there is a good chance of a rally emerging over the next several weeks. But if the market can’t manage to rally then the probable alternative is a substantial selloff.

Yesterday I showed a study related to Turnaround Tuesdays. I showed that when the market makes 3 down days going in to Tuesday there has been a decided upside edge. Well, that edge certainly did not play out Tuesday. So what happens when Turnaround Tuesday fails to ignite a turnaround? I went back to 1980 for this test to get a good number of instances.

This suggests Tuesday’s failure to turn around simply appears to be a delay rather than a bad omen.

In the January 13, 2009 blog study I looked at the concept of Turnaround Tuesdays. I found the old market adage really seemed to provide an edge. And while that edge had been prevalent since at least the 60’s, it had become even stronger over the last decade. Below I’ve rerun one of the tests from that post. In this case I looked for exactly 3 dwon days in a row and today being Monday.

These results are quite compelling and suggest a substantial upside edge.

Of course this morning’s huge gap down may throw a wrench into things. I examined similar gap situations on June 29, 2010.

Wednesday the market posted an inside day. I haven’t discussed inside days in a while. For those unfamiliar an inside day is simply a day that makes a lower high and a higher low than the day before. Over the last decade, when the market has been trading below the 200ma, inside days have suggested negative short-term implications. Below is a table that demonstrates this.

Of course there are other nuances and filters that could be applied that could increase or decrease this edge. But generally there has been a poor track record following inside days. It’s been fairly steady, too. This can be seen in the equity curve below which uses a 2-day holding period.

Last night both Cobra of Cobra’s Market View and Mr. Ice of ETF Prophet noted that Monday was the 5th day in a row that the market gapped down to start the day. Pre-market trading is dominated by the pros. Retail traders typically don’t trade until the official open at 9:30am EST. I’ve heard in the past that big institutions that want to buy will sometimes try and push futures lower in the morning when there isn’t as much liquidity. They do this with the intention of buying at the open (at cheaper prices) when liquidity comes into the market. Of course news and trading of foreign markets overnight will also have much to do with the open. In any event, 5 gaps down in a row is fairly rare. It raises the question, “Does a consistently cheap open provide an edge going forward?” Let’s take a look:

Instances are a little low but they sure offer some compelling evidence for the bull case.

Large gaps down rarely fail to bounce at all and sell off so persistently as happened yesterday. In the past this behavior has often been followed by a rebound the next day. The study below exemplifies this.

Unfortunately, it looks like the market is going to open today well below yesterday’s close. Of the instances listed above there were only two that gapped down the next day. They were the most recent two instances. This puts us in somewhat uncharted territory.