Thursday, April 30, 2009

Gaps Up From 10-day Highs

The SPY is set to gap higher this morning after closing at a 10-day high yesterday. Below are the stats when buying $100k of SPY at 9:31 the morning when SPY closes at a 10-day high and then gaps up at least 0.75%. The trade is then exited at the 4pm close (2001-present):

Below are all the instances found:
(produced with Tradestation)

Looks to be a decent chance of a move down from the open today.

Wednesday, April 29, 2009

More Nasdaq Weekly Volme Spyx

Yesterday I looked at the extremely low reading the Quantifiable Edges Nasdaq Volume Spyx indicator put in during the week of April17th. After posting a reading below 0 did it again the week of the 24th. Below is an updated weekly chart of the Nasdaq Spyx ending on the 24th.

Going back to 2000 (as far as Nasdaq Spyx data has been calculated) there has only been one other time the market had back to back readings below 0. That was the week ending 11/29/02, which came 7 weeks after the October 2002 bottom. It marked the top of that rally. From there it pulled back for over 3 months before the March 2003 rally and new bull market began. Similarly, we are now about 7 weeks off of the March bottom.

Another instance that came close to a back to back sub-zero reading was December of 2008. What followed that was obviously quite bearish over the next few months.

There have only been 4 instances where the Nasdaq Spyx has posted a sub-zero weekly reading and then the Nasdaq has risen the next week. Those four instances along with their 8-week returns are listed below.

The “best” result above was the 12/04 – 1/05 period where the Nasdaq “only” lost 5%.

Sample sizes for these tests and yesterday’s are too small to read much into. It would be dangerous to draw conclusions from so few instances. Still, with such lopsided intermediate-term results its worth taking note. While the week is only 2 days old, the indicator has yet to rise as of Tuesday’s close.

Volume indicators aren’t the only ones reaching extremes. I noted last week breadth was more overdone by at least one measure than it’s been in decades. While price hasn’t yet begun to roll over the rate of ascension has certainly been slowed. The S&P hasn’t hit a new high in 7 trading days. This is the 1st consolidation of more than 4 days since the March lows. Are we setting up for a pullback or consolidation of several weeks? I don’t know. I am seeing some warning signs that would suggest caution. I’ll be keeping my eye out for more of them, as well.

Tuesday, April 28, 2009

Nasdaq Volume Spyx Weekly Chart Suggesting Trouble

I’ve discussed my Volume Spyx indicators on the blog a fair amount. What I haven’t shown is that Spyx readings can also be useful on weekly charts. The calculations for Spyx readings are proprietary but basically it looks at volume on a relative basis across multiple securities. For the Nasdaq chart it’s looking at Nasdaq securities. As the name would imply, it looks for “spikes” in the relative volume statistics.

Over the past 2 weeks the Nasdaq Volume Spyx weekly chart has been giving readings with potentially bearish ramifications. Below is a chart of the Weekly Nasdaq Spyx as it stood on the weekend of April 17th. The chart is followed by some research from that weekend’s Subscriber Letter.

The April 17th reading was the lowest in years. I looked at other times the indicator closed below 0 and found instances to be scarce. (Data goes back to 1/1/2000.) Below are the returns over the next ¼:

Note the above table used position sizes of $1,000,000 rather than the usual $100,000. Ten-eleven weeks out the Nasdaq was down every time. There was some slight overlap though. Below I’ve listed all occurrences along with their 10-week return.

The instances with overlap were 11/02 and 11/04

To increase the sample size, I also looked at instances where the Nasdaq Weekly Volume Spyx came in at less than 10. Those results are below.

Even with the loosened criteria, results are still quite bearish.

I have recently added the Weekly Nasdaq Volume Spyx chart to the charts page in the members section of the website. I will soon make all the weekly readings downloadable for Gold Subscribers as I do with the daily Nasdaq and S&P Spyx readings and the CBI. (Click here for a free 1-week trial.)

In my next post I’ll show the 4/24 weekly chart and some additional research associated with that. (Preview: Still below 0 and suggesting possible trouble ahead.)

Monday, April 27, 2009

Putting Large Gaps Down Into Context

As I am writing this late Sunday night the S&P futures are down close to 2%. Large gaps are often seen as fading opportunities by traders. This is due to the propensity of the market to reverse gaps. In a study I posted to the blog a few weeks ago, I showed that the propensity of the market to either reverse or follow through on gaps up of 1% up or greater depended largely upon the action leading up to the gap. I have found a similar dependency when looking at gaps down.

Below are the results of a system that looks to purchase $100k of SPY any time it gaps lower by 1% or more and HASN’T closed higher for 2 days in a row. The trade is exited at the close of the day. As you can see, reversals are slightly favored and the net expectation is for further upside. (1998-present.) Click any of the images below to enlarge.

But what about those times like now where the SPY has closed higher for 2 days in a row? Those results are below:

In this case results appear to go from somewhat bullish to strongly bearish. An overbought market that gaps down big tends to sell off further. Below is an equity curve of the system.

As you can see the downside tendency has been quite consistent.

Lastly, I also looked at gaps down following at least two down days in a row. Those results are below:

Here you see that although it’s a 50/50 proposition, the rewards outweigh the risks by a large degree. Of course since Monday’s potential gap down would be coming after two up days, the net expectation would favor more downside.

To best understand the meaning of a pattern, it often helps to take it in context. Gaps are no exception.

Friday, April 24, 2009

Weak Nasdaq Breadth On An Up Day

One of the most notable aspects of Thursday’s trading was the poor breadth in the Nasdaq, as only about 38% of issues closed higher. This is extremely rare on a day where the market rises. I looked at the possible implications a number of different ways in last night’s Subscriber Letter. The consensus whether looking at the effect on the Nasdaq or the S&P was bearish. Below is one study that looks at other times the Nasdaq Up Issues % (Advancers / (Advancers + Decliners)) came in lower than 40% while the S&P closed higher.

(click to enlarge)

Bearish results across the board over the following 2 weeks or so.

Subscriber Letter Trade Results for March

Between vacation the 1st week in April and taxes the 2nd week it’s taken me forever to finalize the March trade results, but they’re finally here. March was the 1st difficult month the Letter has had since last August. There were 2 main culprits – Catpults and Index trades.

The Catapult trades had their worst bout in a long time. The primary culprit was the single worst trade in the history of the system. As I’ll describe below, the results are just a scorecard and not a portfolio return. The Catapult trades are normally scaled into in 3 lots. I always just list the gain/loss of each lot rather than slicing the results of those that only had one lot on. Since I don’t suggest allocation sizes it hardly matters, but listing each lot can exaggerate results sometimes. In October for instance, the results were exaggerated upwards. This month, downwards. The net additive results of all the Catapult trades was a loss of about 17% on 21 lots – or about -0.8% per lot. The single worst trade I referred to above had 3 lots active and they accounted for a loss of 58%.

To be fair, it wasn’t just the Catapults that faltered. I was too early to enter some of the index trades as the market collapsed in late February / early March. I also took them off too early in the bounce. The net additive result of the 5 index-sized lots was a loss of 2.8%. These positions are typically scaled into as well and they were in March, with a maximum of 4 on at one point. What I did very right was I avoided trying to short the bounce on the initial thrust off the bottom. Many traders that use overbought/oversold methods mistakenly viewed the initial bounce as a simple “overbought in a downtrend” setup. I discussed extensively in the Subscriber Letter that shorting the initial bounce appeared to be a dangerous proposition. The only short index trade idea I took wasn’t until 3/26 and I exited it with a nice 2.9% profit on 3/30.

Other systems were quiet as I often defer to the Catapults in times of market stress. It didn’t work out this time, but it traditionally has (see last October and November for some outstanding examples). There was only 1 system trade idea tracked outside of the Catapults and it went for a decent gain.

April so far has been much more efficient. There have been only a handful of trade ideas that have received fills and results have been strong so far. I’ll get to those results next month, though. Below are some of the usual caveats and explanations followed by March’s results.

As mentioned above, I don’t suggest position sizes. The primary reason for this is I’m not acting as a financial advisor. I don’t feel it is appropriate to suggest allocation sizes without understanding someone’s financial situation and risk tolerance. Even for my own trading I run different portfolios with different levels of aggressiveness. For instance, my most aggressive portfolio is my IRA. Here I may use options to sometimes get 400-500% leveraged. Other portfolios on the other hand normally take much more conservative stances and some rarely reach or exceed 100% exposure.

Since I don’t suggest position sizes this is should not be considered a performance report, but rather a trade idea scorecard. Therefore, no matter how objective I try to be the reporting of the results is always going to be skewed depending on how you approach the trades. For instance, I always recommend scaling into the Catapult positions in 3 parts, whereas the “System” trades (whatever system I unveil other than Catapult) are normally one entry. The “Index” trades I normally recommend scaling into as well. For my own trading I trade much larger size with the index trades than any of the individuals. I also control my exposure by limiting the total amount invested per day. As I mentioned, this will vary depending on the account I’m trading. My most aggressive account I may put in up to 100%/day and get heavily leveraged using options. A more conservative account may max out at 15%-20% per day.

It’s unlikely anyone would have taken all of the trades with equal amounts, so personal results would vary greatly depending on the trader’s approach. Simply adding up the results of the individual triggers as I do below is an admittedly poor representation of returns. A net positive or negative does not necessarily mean a person following the ideas would have made or lost money during the period measured. And the sum total is certainly not representative of what a portfolio would return. All that aside, below are March’s results (click to enlarge):

Detailed trade by trade results will appear in this weekend’s Subscriber Letter. If you haven’t checked out the gold membership area yet, then click here to sign up for a free trial (only a name and email address required). It’s not just trade ideas. It contains research far beyond the blog as well as members-only charts, systems (with code included), and custom indicators.

Thursday, April 23, 2009

Late Day Reversal Flips S&P To Negative

We’ve seen before how strong end-of-day selloffs are often an overreaction. Frequently this means a bounce back over the next day or so. Tonight I looked at the below situation, which describes Wednesday's action:

Instances are low, but the results are interesting. The pattern is a sharp bounce followed quickly by another drop lower. Of the 9 instances, 8 of them closed higher than the entry trigger at some point in the next 3 days. Amazingly, 7 of 9 closed lower than the entry trigger within 4 days. Looking out 6 days would move the number to 8 of 9 and if you give it 6 days, then all 9 instances closed lower at some point. What I see is a propensity for violent chop over the next few days.

Wednesday, April 22, 2009

Fear vs Greed

There was some discussion of fear vs. greed the other day in the comments section of the blog. The discussion revolved around whether they were mirror images of each other. My contention was that they aren’t and I pointed to the study regarding Worden Bros. overbought indicator from Monday as evidence of that.

There are many ways to look at fear vs. greed. Today I’ll touch on the topic a bit more. For today’s study I first applied a zig-zag indicator to an S&P 500 chart to identify all places where the market rose or dropped at least 10% before swinging 10% in the opposite direction. Below is a copy of that chart.

While a reversal after a 10% move is an overly-simple and perhaps not terribly accurate definition of a “top” or “bottom”, it does well enough to make today’s point. For instance, I doubt many traders would refer to 10/13/08 as a “top”, but it qualified, so I used it.

I then took the dates of the last 12 tops and the last 12 bottoms going back to 1999 and ran a few simple tests.

The first simply looked at VIX levels. The VIX is often referred to as the “fear index”. It’s really more a measure of expected volatility based on options premiums, but when traders are fearful it tends to spike. I then took an average of the VIX for the tops and the bottoms. At the tops the average VIX reading was 27.41. At the bottoms the average VIX reading was 47.70.

This confirms the obvious point that traders are more fearful at bottoms than tops. It also suggests that expected volatility is much higher at bottoms. This volatility can present opportunity. But while the expected volatility is much higher, what about the actual volatility after the top or bottom has been made?

Below are 1-5 day returns for the 12 tops identified. (Based on $100,000 / trade)

Now look at the 1-5 day returns of the 12 bottoms.

People are too greedy at tops and too fearful at bottoms. Fear is a much stronger emotion, though. It therefore results in much stronger market moves. When trading overbought/oversold methods, traders should take this into account.

Monday, April 20, 2009

The Most Overbought Market At Least 23 Years?

I’ve seen it pointed out a few places that the number of stocks above intermediate-term moving averages (40 or 50-day) is now at an extreme level. I’ve done some testing in the past and found such indicators to be of limited value. Worden Bros. has several indicators that show the % of stocks trading relative to the 40-day moving average. In addition to the simple % above/below, they also show how many are at least 1 and 2 standard deviations above and below the 40ma. I believe these indicators are more telling. I’ve found this information to be especially useful in looking at extreme selloffs and have compared the % 1-standard deviation below the 40-ma indicator (T2114) to my Capitulative Breadth Indicator (CBI). (Click here for that post.)

Part of what gives the CBI and T2114 their effectiveness is the propensity for sharp and powerful short-covering rallies to emerge from such extreme conditions. In developing the Catapult method and CBI indicator I was unable to find an overbought equivalent. Of course the market is dealing with different emotions near tops than it is near bottoms. Fear, which is prevalent near bottoms, is much more powerful than greed, which is prevalent near tops.

So the question then becomes, since there is no CBI reading for extremely overbought, what might the Worden 1-standard deviation measure suggest when it gets extremely high? As of Friday it was certainly extremely high. Over 80% of stocks closed at least 1 standard deviation above their 40-day moving average. Worden Bros. maintains data back to 1986 and this is the 1st time the indicator has cracked the 80% level. I looked at other overbought levels to study the 1-month returns following some less-extreme readings.

As you can see above, returns have generally been positive following other times the indicator has reached extreme levels. On the low end the results are about the same as the long-term market drift. While not shown, periods leading up to 20-days are also all generally positive. As the indicator moves higher the results look even more bullish. But instances are incredibly low, so not much can be extrapolated. I see two points to take away from this exercise: 1) When the market gets extremely overbought that is not necessarily a bad thing, and on its’ own is certainly not a signal to sell short. 2) By this measure the market is now more overbought than it has been in at least 23 years.

The most overbought ever would seem to suggest the market is unlikely to continue to rise at anywhere near its recent pace. On the other hand, those expecting a sharp, sustained selloff from here had better be basing their expectation on evidence other than just overbought breadth.

Friday, April 17, 2009

How Time Stretches Can Provide An Edge

While it seems the S&P has hardly pulled back at all since the March lows, one market that has made it the entire time without a single close below the 10-day moving average is Singapore. EWS has now closed above its 10-day moving average for 25 days in a row. When a stock or ETF spends an extended amount of time on one side of a moving average it creates what I refer to as a “time stretch”. I’ve discussed time stretches on the blog before but not for quite a while.

When a time stretch gets large, such as the current 25-day 10ma time stretch for EWS, you can often expect a move back to the other side of the moving average in the near future. I have found time stretches to be useful in system building.

To demonstrate, looking at the current EWS situation I set up the following criteria among my list of about 115 liquid ETF’s. (This list excludes all inverse and leveraged ETF’s.)

1) ETF must have closed above its 10-day ma for at least 25 days.
2) Today it closes at the highest level of the upswing.

Selling short under those conditions and then covering on a close below the 10-ma provided the following results over the last 10 years across my list of ETF’s:

Trades: 211
Wins: 167
Losses: 43
% Wins: 79%
Avg Win: 1.8%
Avg Loss: -2.1%
Avg Trade: 1.0%

As you can see, although the concept is simple, the results can be quite powerful. Traders may want to consider including time stretches in their bag of tricks. I use them for a few systems tracked in the Gold members section of the website and in Quantifiable Edges Subscriber Letter.

Wednesday, April 15, 2009

Implications of an Extremely Low Put/Call

One indicator that has received some attention lately is the CBOE Equity Put-Call ratio. Many people believe that very low readings can be a sign of complacency and an impending top. I decided to test this out a little.

Below I look at the 3-day equity p/c ratio compared with the 100-day. To see why I normalize using a long-term moving average, click here. On Tuesday the 3-day average dropped over 25% below the 100 day. This has only happened a handful of times. To get a larger sample size I used a 20% trigger instead.

Based on the above test, the recent extremely low numbers in the CBOE Equity Put/Call ratio don’t appear to be predictive of a selloff. In fact, they could actually be interpreted as a short-term bullish indication.

Tuesday, April 14, 2009

Is Buying Drying Up...Again?

In the past I’ve discussed how extremely low SPY volume is often a bearish indication. Below is an updated version of a study I posted nearly a year ago on 4/22/08.

While the edge here is not the strongest in terms of % or magnitude it has been consistent over time. Below is an equity graph of the 5-day exit that was highlighted in yellow above.

Monday, April 13, 2009

Volume Spyx Hits Extreme Low - What That's Meant

While overall volume rose on Thursday, the Quantifiable Edges Volume Spyx indicator once again dropped precipitously. It came in at -12, which is an incredibly low number. I looked back at other times the S&P 500 volume Spyx indicator dropped this low. Going back to 1995, there were only 4 readings of -10 or lower. They occurred on 6/10/04, 4/28/08, 5/30/08, and 11/3/08. In every instance the market was trading lower 2 days later. The average drop for the 2 days was over 1%.

Loosening the requirements to look at instances with a close below 0 produced the following results:

As you’d expect we see bearish tendencies here. As I‘ve discussed in the past the volume Spyx can be especially effective when you also take the day’s direction into account. Low readings on up days tend to be especially bearish while high readings on down days tend to be especially bullish when looking out over the next few days. Below you can see the results following readings below 0 when the market also closed up on the day.

Wednesday, April 8, 2009

Is SPX Down Big & VXO Down Bearish?

Interesting about today is that while the S&P dropped well over 2%, the VIX (and VXO) also dropped. Often traders will interpret this as bearish. The thought process is that when the VXO doesn’t drop with the market it suggests a possible complacency among traders. The theory is that this complacency may lead to further selling until participants become somewhat fearful again. At that point a rally will become more probable.

I looked at this theory last month when considering action over the course of several days. At the time I found no evidence to support the bearish case. For testing I looked at a 5-day divergence to measure the return in the VXO and S&P prior to triggering. Tonight I looked at a 1-day divergence.

Not only is there a lack of bearish evidence but these results could be considered borderline bullish. I also looked at more extreme 2% SPX drops:

If the number of instances wasn’t so small then I’d certainly consider the results strongly bullish.
A sensible sounding theory is just that – theory. This is an example of why I make every effort to test all of my trading and market bias ideas.

Note: Apologies for the sporadic posts this week. Things should return to normal next week.

Friday, April 3, 2009

Some Simple Shorting Systems

When the market hits new highs it tends to excite the media. The S&P closed at its highest level since early February on Thursday. Such news may sound positive when delivered by an anchorperson. But since the beginning of 2002, when the market is trading below its 200-day moving average, there has been a strong tendency to pull back after hitting news highs. Below are results of some simple systems. These systems call for shorting the S&P 500 any time it hits an X day high while under the 200ma. The trade is covered when the S&P 500 next closes below its Y-day moving average.

All of the systems above have performed quite well on the short side. As you can see, hitting new highs really isn’t something bulls should get too excited about.

Thursday, April 2, 2009

Large Gaps Up After The Market Has Already Risen

The S&P futures are up over 2% as I type. Last week I took a look at large gaps higher since September 2008. Generally the result was the larger the gap up the better the market performed that day.

Not all gap formations are equal, though. It can also be important to consider the state of the market prior to the gap. When the market is already extended up the chances of a reversal down are much higher. One simple way to define an “extended” market would be to say if it has closed higher the last 2 days it is extended. Now let’s look at some test results. All tests were run on the SPY over the last 10 years. Results are based on $100,000 per trade.

The 1st study looks at buying at 1 minute past the open on a day where the SPY gapped up 1% and selling at the 4pm market close. In these cases the SPY had NOT risen the last 2 days in a row:

(click to enlarge)

As you can see the results are clearly bullish. Over 2/3 of the trades were profitable and winners were about the same size as the losers.
Now let’s look at the results of buying such 1%+ gaps when the market has already risen 2 days in a row:
(clck to enlarge)

Results are flipped here as we now have over 2/3 of the trades as losers. Again winners and losers are about the same size. The solid bullish tendency has switched to solidly bearish.

For those interested below are all of the instances of the 2nd (bearish) test.
(click to enlarge)

Wednesday, April 1, 2009

Good Days With Bad Finishes

Finishes like Tuesday’s often feel bearish to many traders. They interpret the inability of the market to hold on to its gains as a potential negative. In actuality, while the market may struggle over the next 1-2 days, over the course of the next 1-2 weeks implications appear bullish.

Below is a table showing the result of buying any time the S&P closes over 1% below its high for the day but still positive by at least 1%:

Between 5 an 9 days out you’ll notice some strongly bullish results. Not visible in the above table is that 19 of 24 instances (79%) posted a close higher than the trigger day within 3 days. Looking out 6 days that number increases to 23 of 24 instances (96%).