Blog | Quantifiable Edges | Assessing Market Action With Indicators & History

Equity Put/Call Ratio Spikes Down

Posted by Rob Hanna in Quantitative Study May 7, 2008

Dr. Brett Steenbarger over at Traderfeed has done some excellent work with put/call ratios over the last few years. One way he measures it is by comparing a short-term moving average of put/call volume to a long-term put/call moving average. He has demonstrated numerous times how spikes in the short-term averages over the long-term averages tend occur near market bottoms.

The options market is now experiencing a spike down, rather than a spike up, in the put/call ratio. An example of this could be seen by considering the 5 and 65 day moving averages of the CBOE Equity put/call ratio (two averages that Dr. Steenbarger sometimes uses). As of tonight’s close the 5-day stood at 0.62 and the 65-day at 0.80. This is over a 20% gap from the 65 to the 5.

If spikes up in the put/call can predict a bottom, could a spike down predict a top?

Not based on the limited data available. The equity put/call data has only been tracked back to 2003. Below are the results:

The dates for the 4 instances were 11/10/04, 12/16/04, 5/24/05, and 7/14/05. Basically it happened at times where the market rallied strongly following a pullback or correction. I suspect the sharp drop in the average put/call ratio represents a shift in attitude of market participants. Coming out of a corrective period it appears this buying enthusiasm is a good thing and not representative of complacency.

I also ran the same test using the Total Volume Put/Call Ratio back to 1996. Results there leaned to the bullish side as well with between 57% and 67% of cases showing continued gains over the next 2-3 weeks.

The low put/call ratios may become an issue if they remain low for an extended period. I’d be careful of trying to call a top based on any spikes lower in the ratio, though. Spikes may represent enthusiasm rather than complacency. Now if we could just get some volume we could really make a case for enthusiasm.

Quantifiable Edges Subscriber Letter April Results

Posted by Rob Hanna in Subscriber Features/Letter May 6, 2008

I’ve presented below the summary results of the Quantifiable Edges Subscriber Letter trade ideas for April and since inception. April was a decent month. Although there were fewer trade ideas and rewards were a bit smaller than March, it was still strongly positive.

In addition to the trade ideas, the Subscriber Letter provides additional research beyond the blog and shows CBI analysis down to the sector level. More features will be announced soon. Should anyone wish to receive a free three-day trial to the Quantifiable Edges Subscriber Letter, just send an email to QuantEdges@HannaCapital.com and include your name and email address.

A Look At Light Volume

Posted by Rob Hanna in Quantitative Study, volume May 6, 2008

The market pulled back today on light volume. Often heralded as a good thing, readers of this blog understand over the short-term it isn’t.

So I spent some time looking at volume tonight. I’ve been noticing how it has just fallen off a cliff since Easter. I decided to compare the volume the last 30 days to the volume of the previous 50. The last 30 days have posted average daily volume less than 75% of the average of the previous 50. I thought this unusual and did a study to find other times this may have occurred.

Looking back to 1960 I found 10 other instances. Over the next 2-6 months prices rose 7 of ten times. There was also a general decrease in historical volatility 8 of the 10 times. Gains were generally in line with random and 7 for 10 isn’t mathematically significant, so I’m not sure there’s a lot to be learned here. I did find it interesting that all prior instances happened between 1960 and 1980. Dates are listed below:

6/26/61, 8/10/62, 7/26/63, 6/29/66, 2/25/74, 8/25/75, 4/19/76, 7/27/78, 12/13/78, 4/21/80.

If someone sees something I don’t I’d love to hear about it. I did find it interesting that many of the occurrences happened in a generally difficult period for the market but returns were pretty good. For instance, the 70’s had 5 instances – 4 of which led to higher prices over the next two months.

The Gap Band

Posted by Rob Hanna in gaps, Quantitative Study May 5, 2008

I received an interesting note from Dr. Steven F. over the weekend who observed the SPY gap open put it above it’s upper Bollinger Band on the gap. For an index to gap open outside of its Bollinger Band, price will almost always need to be somewhat stretched already. It would seem logical that a gap up into an already short-term overbought condition would be a likely candidate for a reversal. I decided to test.

The criteria was simple. A gap up over the upper Bollinger Band would signal a short entry. The trade would be exited near the close of the day. Looking back to 1998 in the SPY I was able to identify 79 such instances. There were 43 (54%) winners and 36 losers. The average winner made 0.56% and the average loser lost 0.50%. The profit factor was a modest 1.34.

I then broke it down by instances above and below the 200-day moving average. Above the 200-day moving average there have been 62 instances of a gap up above the Bollinger Band. Shorting these and covering on the close resulted in 53% winners. Winners outsized losers by 0.57% to 0.38%. The profit factor was a decent 1.7.

The trouble occurred with this less than “Outstanding” Gap Band strategy when it was attempted below the 200-day moving average. There were 17 instances. Ten winners, but the average loss was 1.0% vs. an average gain of 0.5%. Overall a losing strategy below the 200ma.

As with previous gap studies, it appears gaps up in long-term downtrends are dangerous to try and short. While it would’ve worked out on Friday, you always need to be wary of a short-covering rally or trend day up.

Initial results of buying a gap down below the lower Bollinger Band appear better. I will look at them in more detail at a later time.

The 200ma Cross

Posted by Rob Hanna in Quantitative Study May 1, 2008

The Nasdaq 100 led the way today as it rose over 3% and closed decisively above its 200-day moving average. It is the 1st major index to retake the 200-day (the S&P mid-cap 400 also did it today if you consider that one major). Some people believe the 200-day moving average is an important technical measure. Some suggest it’s a somewhat meaningful psychological level. Others see little value in looking at it. Rather than discuss and postulate on the merits of an indicator, I prefer to test it. Let’s see what a cross of the 200-day moving average in the NDX has led to in the past. (Test period is mid 1986 – present.)

Over the period tested the average gain per day in the Nasdaq 100 was just under 0.09% ($90). As you can see, the market seemed to gain steam after crossing the 200-day moving average and strongly outperformed the typical period.

What if we add a filter to eliminate those times where it barely peeked across? I filtered to only look at crosses that finished at least 0.5% above the 200ma. Results below:

These results are even stronger.

How about if we also require that the NDX makes a good-sized move on the day of the cross? I used a 2% gain to test this. More results:

Whatever the reason, a move through the 200-day moving average has provided the NDX some extra fuel in the past. When the move was strong and decisive like Thursday, that made for even better results. A cross above the 200-day moving average isn’t a magic buy signal – but there are worse ones.

Image courtesy of DialBforBlog.

Fed Reaction Not As Ugly As It May Seem

Posted by Rob Hanna in Fed Study, Quantitative Study May 1, 2008

Last night I showed that when the Fed disappoints and the market drops by 1% or more in reaction to it, then the market has generally recovered and worked its way higher over the next couple of weeks.

The Fed disappointed today, although not to the degree we looked at last night. The S&P finished down a relatively mild 0.4%. While the end result wasn’t that poor, the fact that it was up 1% shortly after the announcement and then faded late left many traders with a bad taste. Measuring where in its range the market closed the day is one way to measure the mood near the end of the day. Today the market closed very close to its lows. I ran a test to see how the market performed after closing near its lows on a Fed day. Results below:

Looks fairly positive. Since I’d already established moves of 1% down or more could be a positive, I decided to exclude those instances from the results and re-run it:

Not quite as strong, but still a slight upside edge seems apparent.

Also notable about today was the fact that the market made a recent high. I ran a test to see the following 1) Fed day 2) Made 20-day high. 2) Closed in bottom 10% of range. I was only able to find 3 instances. Therefore I removed the Fed day requirement. Below are those results:

The first few days were choppy, but certainly not bearish. After that the market generally rose.

Some technicians may suggest today was an ugly late-day reversal. I’m having trouble finding much ugly about it.

What A Strong Reaction To The Fed Could Lead To

Posted by Rob Hanna in Fed Study, Quantitative Study April 29, 2008

Last month I presented historical returns following days when there was a strong reaction (+1%) to a Fed (pictured at right) announcement. The results showed that the positive reaction was typically short-lived. Another highly anticipated Fed announcement is due tomorrow. Tonight I thought I’d present historical returns following a disappointment (just in case).

As you can see, a strongly negative reaction to a Fed announcement has typically been followed by a very positive next two weeks. So…if the Fed does something the market “likes” it will go up tomorrow, but over the next two weeks returns will likely be disappointing. If the Fed disappoints the market tomorrow we’ll see afternoon weakness. This disappointment could lead to a nice rise over the next two weeks. So the bulls should want the Fed to upset them tomorrow, and the bears should hope for a short-covering rally.

I find these tendencies to be quite interesting. My untested (as of yet) theory on why this occurs is that “good” news normally will come in the form of a rate cut or other stimulus. This kind of good news happens during times of economic and market weakness and the improvement for both can take time. “Bad” or “disappointing” news many times will come in the form of tightening. The Fed tightens normally when the economy and markets are strong. Just as they can’t fix it in a day, they can’t break it in a day either. Their “bad” decision won’t derail a rally right away and the market will typically continue to trudge higher for a while.

Whatever the reason, the point to remember is this: Don’t get too caught up in the reaction to the Fed tomorrow. The move likely won’t last longer than a day or two before reversing itself.

Edit note: Test was run back to 1982.

Bearish Bar But Fed Looming

Posted by Rob Hanna in Quantitative Study April 28, 2008

Last week I showed how very low volume in a short-term uptrend is a negative. It happened again today along with some other action that hasn’t been constructive historically. Here’s a quick look at what happens when several negative all come together:

For those who would like to review the previous instances they were 6/2/94, 11/6/00, 11/13/01, 10/28/02, 11/29/02.

You don’t want to read too much into just 5 instances, but the formation has been quite bearish in the past. Of course as I mentioned last night, the Fed announcement Wednesday will likely have a larger influence on short-term direction than historical precedents.

Ten High Straight Up

Posted by Rob Hanna in Quantitative Study April 27, 2008

I’ve previously discussed some of the findings in Larry Connors’ book “How Markets Really Work”. In the book Larry looks at certain market situations and determines whether the market has historically outperformed or underperformed when those situations arose. One I previously discussed was a 3-day rise in the market when it is trading under its 200-day moving average. Historically, the market has struggled to add further gains after this has occurred.

Another edge Larry discusses in his book is performance following a 10-day high. In the book he shows there has been a negative expectancy over the next 1-5 days following a 10-day high. I have personally examined 10-day closing highs and found a negative expectancy 3-5 days out when the market is under its 200-day moving average. When it is over the 200-day the expectancy is no longer negative.

On Friday the S&P 500 closed higher for the third day in a row. It also made a 10-day high and closed at a 10-day closing high. I ran some tests to see what happened when you combined some of these 10-high criteria with 3-straight up days. Results of the different combinations I looked at were similar. Below is one example:

A negative expectancy persisted up through 12-days out. The greatest part of it appeared in the 1st three days. Of course during the next three days there is going to be a Fed announcement. The reaction to that may have a larger affect on market movement than my little test. Still, it’s worthwhile noting the negative expectancy in these situations. Below is a chart showing all the recent instances with a 3-day exit strategy.

Are The Leaders Suggesting A Market Meltdown Or A Sector Rotation?

Posted by Rob Hanna in Quantitative Study April 25, 2008

While the major indices faired well on Thursday, the IBD 100 got whacked. Only 28 stocks rose and the Index declined over 2%. (Hat tip to IBDIndex.) I noticed this has been an especially bad week for the IBD 100. While the S&P 500 is only down 2 points the IBD 100 is lower by almost 3.5%.

The IBD 100 is a group of 100 stocks compiled by Investors Business Daily that represent market leadership. Unlike more traditional indexes, it is updated weekly and turnover on the list is quite high. Since it is supposed to represent current market leadership, I was curious to see if the strong underperformance so far this week would be significant if the major indices actually managed to close positive. In other words, in a generally rising market, does a breakdown in leadership signify sector rotation, or is it a sign that the indices will soon follow the leaders south?

I looked back at weekly data to 3/12/2004, which is all I had available for the IBD 100. There have only been 3 times when the IBD has lost as much as 3% while the S&P has managed to finish the week positive. They were 3/10/06, 7/21/06, and 3/21/08. Below is a chart of all the instances where the S&P has had a positive week and the IBD 100 has dropped at least 1%.

I’ll let you draw your own conclusions, but I’m having trouble finding anything that would suggest a meltdown is imminent. It appears the next 3-4 weeks have often been positive following these instances.

Note: If anyone has or knows where I could find the daily IBD 100 Index values going back to its inception in 2003, please let me know. Thanks.

Intraday Extremes

Posted by Rob Hanna in Intraday, Quantitative Study April 24, 2008

The price action on Wednesday was quite interesting from an intraday perspective. I’ve posted a chart of SPY below. What sticks out to me it the fact that there were two extreme price moves in close proximity of one another. The chart is a five minute chart, which is the intraday chart I look at most often. Notice the two arrows with notes attached.

When you get a strong and steady move like was seen from 10:10 to 10:55 or in the opposite direction from 11:55 to 12:25, one indication that it may be nearing its end is if a large range bar is posted.

To help illustrate this concept I ran some historical studies. The first one looks to sell short any time there have been at least 6 up closes and the most recent bar makes the largest rise of any bar in the move. It then sells “X” bars later or at 4:00 – whichever comes sooner. No trades are taken before 9:50.

As you can see, selling into this extreme move has a positive expectancy from 5-50 minutes out.

The second study looks at exactly the opposite formation. It buys the SPY any time there have been at least 6 down closes and the most recent bar makes the largest decline of any bar in the move. It then covers “X” bars later or at 4:00 – whichever comes sooner. No trades are taken before 9:50.

Again you can see that the edge is for a counter-move rather than a continuation for at least the next 5-50 minutes.

The large bar after the steady trend many times signals a blowoff. It can be a good place to take profits if you are in a trade, or perhaps begin to look for a reversal. This is not a daytrading system by any stretch, but it does illustrate a concept that daytraders may want to keep in mind.

Mid-Sized Gaps Down

Posted by Rob Hanna in gaps, Quantitative Study April 22, 2008

With little notable action today I thought I’d write up the next part of my study on gaps during uptrends vs. downtrends. Last week I looked at mid-sized gaps up. Tonight I will look at mid-sized gaps down. As a refresher, a mid-sized gap is defined as a an opening between o.25% and 0.75% away from yesterdays close. A long-term uptrned is defined as a close above the 200-day moving average and a long-term downtrend is defined as a close below the 200-day moving average.

I looked at 3631 trading days going back to 11/17/93. Of those there were 368 mid-sized gaps down in uptrends and 222 mid-sized gaps down in downtrends.

Buying at the open and selling at the close when the market was in an uptrend would have resulted in 177(48%) winners and in total gained 8.5% over the 368 trades. On a per-trade basis that’s 0.02% – basically break even. Of those gaps down 213 (57%) filled at some point during the day. (A fill in this case is defined as a move back up to the previous day’s close.)

Buying at the open and selling at the close when the market was in a downtrend would have been profitable 98 (44%) times and LOST you almost 32% over 222 trades. Per trade that’s about a 0.14% loss on average from open to close. Of those mid-sized gaps down, 152 (68.5%) filled at some point during the day.

I previously showed that large gaps down in downtrends typically represented an intraday buying opportunity as they had a large positive expectancy. Mid-sized gaps do not act the same at all. They contain a negative expectancy. Interestingly, although a good percentage of them fill, they also generally fail. Fading the open could be one play. Another could be looking for a short entry after the gap fills.

As I said last week, make sure you take the following two things into account when deciding how to approach a gap opening: 1) Long-term trend of the market. 2) The size of the gap. They both matter.

Edit note: SPY was used for the above test.

Is Buying Drying Up?

Posted by Rob Hanna in Quantitative Study, volume April 21, 2008

When looking at the market statistics today, the one thing that really stood out was the complete lack of volume. The exchange market volume was nearly the lowest of the year. The SPY and QQQQ volume WAS the lowest of the year.

Back in March I busted the old “light volume pullback after a short-term runup signals a healthy consolidation” myth. Today I’ll look at it a little bit differently.

Only 2 conditions – 1) Today’s volume is the lightest in at least 20 days, and 2) the market is trading above its 10-day moving average. I don’t care whether the market is pulling back or not for this test. I only care if we see exceptionally light volume during a short-term uptrend.

Using these conditions I ran tests on both SPY (back to 1994) and QQQQ (back to 1999). Results below:

It appears buying interest is drying up. In the past the market has not fared well under these conditions.

Some may point out that these results differ greatly from the breadth study I showed yesterday. It appears breadth and volume are currently giving opposite indications. Studies that conflict make analysis more difficult. In the Subscriber Letter each night I discuss my take on all the recent studies. In the blog in the near future I will write some detailed thoughts on how I go about doing this.

Strong Breadth To Finish The Week

Posted by Rob Hanna in Breadth, Quantitative Study April 21, 2008

A quick breadth study to share with you this morning. Wednesday and Friday’s rallies were both accompanied by strong breadth figures as advancers outnumbered decliners by more than 3 to 1. I looked back to 1969 to see how the market performed in the past when breadth readings were at least 3 to 1 positive in two of three days. Results below:

These results appear strongly favorable. The average week over the sample period returned 0.175%. The results above beat the average by more than 3 to 1 from 1 to 4 weeks out. Last week’s action appears quite bullish.

Morning Breadth

Posted by Rob Hanna in Breadth, Quantitative Study April 17, 2008

Tonight…some indicator analysis. I think many times traders focus too much on pattern recognition and therefore only look at price. This is especially true of intraday traders. Looking at price and price-based indicators is like looking at the Grand Canyon on a black and white TV. You can see it, but throw in volume and maybe you’ll get some color. Breadth could get you HD. Sentiment gives you surround sound. Add some intermarket analysis like Dr. Steenbarger employs and some historical studies like I prefer and you might actually get to view it for real. Tonight I’m going to look at breadth.

Many traders look at breadth measures like the Advance/Decline line or the McClellan Oscillator on an end-of-day basis. But let’s examine whether intraday breadth analysis can lend some value.

I ran a test back to April of 2001 looking at 15 minute bars in the SPY, the NYSE Advancing Issues and the NYSE Declining Issues. I wanted to see if strong or weak breadth in the morning led to price strength or weakness throughout the rest of the day. The rules for this test were fairly simple.

If advancers led decliners by 4-1 at 9:45, then I would buy the SPY and sell it at 4pm.

If decliners led advancers by 4-1 at 9:45 then I would short the SPY and cover it at 4pm.

Results:

4-1 Advancers:
Trades: 33
Winners: 17
Losers: 16
Avg Win: 0.8%
Avg Loss: 0.5%
Avg Trade: 0.2%
Profit Factor: 1.72

4-1 Decliners:
Trades: 36
Winners: 23
Losers: 13
Avg Win: 1.0%
Avg Loss: 1.3%
Avg Trade: 0.2%
Profit Factor: 1.36

In general, strong advance/decline numbers in the early morning show a mildly positive expectancy in the SPY for the rest of the day while weak a/d numbers show a mildly negative expectancy for the rest of the day.

Traders may benefit from tracking early morning breadth statistics. While this is certainly not something you would want to trade on its own, it may help to confirm or filter trades you are considering. Using breadth in conjunction with price and other kinds of analysis can make for a potent combination.
——————————————————
For those who would like to explore this concept more it is available for purchase and download in Tradestation format here. The study comes with flexible inputs to adjust times and breadth requirements. A pre-set workspace is also included.