Monday, October 31, 2011

When Stock Prices and Bond Yields Are Both Hitting New 50-Day Highs

The fact that the 10-year bond rates hit new highs on Thursday along with the SPX is notable. The study below is one I’ve published in the subscriber letter before.

Generally it seems that higher interest rates have often made bonds an attractive investment. This may lead people to forsake stocks in favor of lower risk returns with improved yield. (Not sure this will be the case this time with yields still so low.) Implications of this study appear to be longer-term in nature than we usually see. To help visualize how this edge has played out over time I have pasted the equity curve using a 50-day exit strategy.

This one looks very similar to the 20-day exit strategy. In this case the downside edge didn't begin to exert itself until the 1970s but it too has persisted lower for a long time.

Tuesday, October 25, 2011

Very Low SPY Volume At A New Intermediate-Term High

SPY volume came in at the lowest level in over a month on Monday. Very low SPY volume when the market is at or near highs is often a bearish sign. A few studies related to this appeared in the Quantifinder this evening. I decided to examine the combination of a 20-day low in volume combined with a 50-day high in price.

Over the next 2 to 3 days there appears to be a solid downside edge based on the numbers. While I expected this to be the case, I was somewhat surprised to see that the edge persisted well beyond that. While I frequently show profit curves in the Subscriber Letter, I rarely do so on the blog. Today I decided to show it. So here it is using a 2-day holding period.

The consistently down sloping curve appears as impressive as the numbers.

Thursday, October 20, 2011

Why I Still Look At FTDs & What Happens When A 1% Drop Follows One

I received a note the other day from a reader who asked why I have so many studies related to Follow Through Days (FTD) on the blog. The reader mentioned that the edges often provided by FTDs are not as compelling as many of my other studies. But one man’s trash is another man’s treasure. And while FTDs may not work as advertised and accurately predict new stock market rallies, they do a very nice job of defining the environment. A FTD tells us the market has undergone a correction. It tells us the market has made a multi-day move off the bottom. And it tells us that strong volume and price action have come into the market. Enthusiasm is picking up.

We understand that the rally is only going to succeed about 40%-50% of the time based on the FTD, but we don’t need to know right away whether it is going to succeed or not to make good use of the information. While many of the past FTD studies on the blog and in the Subscriber Letter have focused on action on and around FTDs and what that might mean for the intermediate-term, it can also be useful to simply put the current day’s action in proper context so that we may better understand what that action may imply over the next few days and weeks. I use context in many other ways and people hardly notice anymore. Studies are always framed by where the market is. Is it above or below the 200ma? At a 20-day high? At a 20-day low? These are all helpful, but recent work has led me to believe that FTDs can be just as useful in defining context, if not more so. So I’ll continue to incorporate them and am optimistic that doing so may uncover some real gems. Anyway…we had a FTD Tuesday and then Wednesday the market sold off strongly. Strong enthusiasm has quickly turned. Let’s look at other instances and what has followed.

The number of trades is a bit low, but the early indications appear to strongly favor another day of selling. Two things really strike me here. 1) There hasn’t been an instance in over 10 years. 2) Run-up/drawdown is heavily skewed in favor of the bears. Overall I find these results compelling enough to take under consideration.

Wednesday, October 19, 2011

When FTDs Occur In Conjunction With 20-Day Highs

I discussed the other day that there has never been a Follow Through Day (FTD) that occurred AFTER a new 50-day high. There has also never been a FTD that occurred in conjunction with a new 50-day high. These things changed on Tuesday since the move up was also accompanied by an increase in volume. But there have been some FTDs that occurred in conjunction with 20-day highs. Below is a new study that shows how they fared.

Results here are impressive over both the short and intermediate-term. To get a better feel for the short-term returns I have listed the instances below.

The run-up to drawdown ratio here is quite impressive. I’ll also note that 7 of the 10 instances went on to have “successful” rallies. (“Success” means it either hit a new 200-day high or at least rose 2x as much as it had already risen off the bottom.) The 3 instances whose rallies did not succeed (circled in red) all saw run-ups of at least 2% before they eventually rolled over and made new lows.

More information on FTDs may be found here.
Positive aspects to this one include the strong breadth and the fact that it came after day 10.
Some obstacles to success include the fact that it is occurring under the 200ma and it is occurring after a substantial market decline.

Monday, October 17, 2011

This is the 1st Time SPX Has Rallied to a 50-day High Without One of These

One of the more amazing things I’ve noticed about the rally over the last 2 weeks is that it has come without any 1% Follow Through Day (FTD) on rising volume. Investors’ Business Daily first published and popularized the concept of the Follow Through Day (FTD). Though they have changed the definition slightly over the years, I have found their original definition to be useful in several studies. My tests go back to 1971, which was the inception of the Nasdaq, and also as far as some of my volume data goes. Since that time there has never been a rally that has taken the SPX from a drawdown of at least 8% to a new 50-day high that was not inclusive of a FTD – until Friday.

This puts this rally in uncharted territory, which is always a little bit of an uncomfortable place for me. A FTD could still occur, and just because we have had a strong 9-day rally does not mean a bull market has already been missed. But one purpose of the FTD concept is to help in identifying market bottoms. If we are already at a 50-day high, then I would say this is one case where the FTD has let traders down in try to identify that bottom.

Note: There was a 1% FTD in the Russell 2000 last week.  I do not look at the Russell 2000 for FTD purposes.  My studies have always looked at the Dow, Nasdaq, and SPX.  The Nasdaq goes back to 1971, and I wanted to be sure to include that index initially.  The Russell only has history back to the mid-80s.  I feel consistency is important when testing and therefore I only look at those 3.  IBD and others may sometimes look at additional indices.  For consistency in testing, I don't.  And this is the 1st rally where none of those 3 have registered the FTD before hitting a 50-day high.

Friday, October 14, 2011

The Incredible Shrinking VXO

The VXO has dropped very strongly over the last week and a half. On Thursday for the 2nd day in a row it closed more than 20% below its 10ma. Looking back to 1986 I was only able to find 3 other instances where this occurred. While it’s dangerous to draw solid conclusions from just 3 instances I decided to show them below:

It hasn’t happened in about 21 years, which makes the setup even more questionable, but the run-up / drawdown stats were so lopsided I thought it was worth pointing out. Over the next 2 days the instances all saw a drawdown between 2.1% - 3.8%. Only 1 instance saw any run-up, and it was just 1.6%. (We’ll be testing that at the open.)

Wednesday, October 12, 2011

What Tuesday's Tight Range Implies

I’m starting to see a number of indications that the market is ripe for a pullback. One indication I noticed yesterday that will often suggest a pullback was the extremely tight range. Tight range can be a sign of indecision. When it occurs with the market extended upwards it can also imply the bulls are running out of gas and likely to step back for at least a short time. The study below demonstrates this concept.

Much of the edge here is realized within the 1st 2 days. If the market is going to act on this signal and flounder  it will often do so rather quickly.

Monday, October 10, 2011

A Columbus Day Tendency

While the stock market is open on Monday, banks, schools, government offices, and the bond market are closed. In past years with the bond market closed, the stock market has done quite well on Columbus Day. Of course the most famous Columbus Day rally was in 2008 when the market gained over 11% after having crashed the week before. Last year I showed that positive momentum leading up to Columbus Day has generally led to a positive Columbus Day. Columbus Day has been celebrated on the 2nd Monday of October since 1971. Below is an updated version of last year’s study.

I’ve circled some of the more impressive stats here. With 75% of trades profitable and winners nearly twice the size of losers risk/reward has been very favorable.

Friday, October 7, 2011

Why Shorting In This Environment Can Be Dangerous

One of the most important and largely ignored aspects of market analysis is putting the pattern or indicator readings you are examining into proper context. Is the market in an uptrend? A downtrend? At new highs? At new lows? It can make a big difference in how the market will react to setups.

Currently the SPX is bouncing strongly off of an intermediate-term bottom. Initial thrusts from bottoms are often much riskier to short than similar thrusts in downtrends that don’t originate off a bottom. I’ve shown many examples of this in the Quantifiable Edges Subscriber Letter. Below is an example that simply looks at 3-day rallies in downtrends. The 1st set of results looks at times where the rally originated from somewhere other than a 50-day low. The 2nd set require a 50-day low.

Here’s the 1st set:

Results here appear solidly bearish.

Now let’s look a times like the present where the market is emerging from an intermediate-term low.

These results appear to be a polar opposite of the previous set. It has been quite a hot streak since 2001 with these setups. Prior to that no edge was evident in either direction, but it still wasn’t bearish like the 1st setup.

Always keep context in mind, and don’t get too aggressive shorting thrusts off bottoms.

Tuesday, October 4, 2011

A Turnaround Tuesday Scenario

Tuesday happens to be the most likely day of the week for a turnaround.  I first examined this in the 1/13/09 blog.  With the last 2 days being heavily lopsided to the downside the odds of a turnaround on Tuesday are even higher.  This is demonstrated in the study below.

All the stats heavily favor a rally on Tuesday, and if not Tuesday, then Wednesday.  Perhaps the bulls can win a day or two here. 

Of course the current environment is a bit more extreme than most of the sample...

Monday, October 3, 2011

Beginning of Month Effect In Uptrends Vs. Downtrends

Monday is the 1st trading day of the month. The 1st trading day of the month is renowned for having a bullish tendency, and this has been the case since the late 80s. But this tendency has primarily played out during uptrending markets. During down times the 1st of the month has not been particularly bullish. This was illustrated quite nicely by my friend and fellow market analyst, Tom McClellan, in his “Chart in Focus” column on Friday. Using the 200ma as our measure of uptrend versus downtrend, I will demonstrate this concept numerically below.

This first study shows results since 1988 of committing $100k per trade on the 1st of the month if the SPX is trading above its 200ma.

The numbers all look very solid. About 2/3 of the trades were winners, and the winners were about 1.5 times the size of the losers, making for a profit factor of over 2.5.

Next let’s look at times like now when the SPX is below its 200ma.

The numbers here net out to a positive number, but just barely. It certainly doesn’t appear to be anything you would want to base a trade on.