Large Gaps Up After Very Large Down Days

In my last post I looked at how SPY has performed from open to close when a 2% gain was followed by a gap up of at least 1%.  Today we are seeing another 1% gap up, but this time it is on the heels of a 2% loss.  So I thought it would be interesting to examine this scenario.  Results of the last test suggested the open to close action favored a move lower, which is what played out on Tuesday.  Below is the general stats for setups like today.

Not a ton of instances.  At this point we can see what appears to be a moderate upside edge.  Let's take a look at the list of instances to see if we can learn anything more.

Most striking to me here is that there was not a single instance that closed within 1.5% of its open.  Volatility after the bell was huge.  Traders looking for intraday moves should have plenty of action today.

Large Gaps After Large Up Days

SPY is gapping up large this morning after rising 2.3% yesterday.  Below is a study showing how SPY has performed intraday under similar circumstances over the last 8 years.

Based on other gap studies we have seen, the negative results are no surprise.  Traders may want to be careful of chasing long entries at the open.

Some Potentially Bearish Inside (Day) Information

I’ve shown before that inside days in long-term downtrends are often short-term bearish. (An inside day is a day like Friday where the market makes a higher low and a lower high than the day before.) So what if that inside day closes higher and comes immediately after a 20-day low like we are seeing now?

Instances are a little low but initial results here appear quite bearish. As I note below the chart I also included a column showing the max losing trade. While it is very large, it was not just one outlier, but 3 extra large decliners that cause the average trade to look so weak. Note though that even 2 days out, prior to the extra large declines taking place, the edge still appeared fairly bearish.

Of course the market appears ready to gap up big this morning. There is certainly the possibility that a large gap up could trigger a short-selling rally which would run overrun the inclinations of this study.

Fed Days During the Weakest Week

Yesterday I mentioned that the week post op-ex in September has historically been the worst week of the year for stocks. I also showed a chart. Later a reader sent the following comment, “Thanks for this Rob, I'm enjoying going through your FED Days book. Seeing as we're in a week with an FOMC meeting, which you show has historically strong performance prior to the meeting does this study include the FED days effect?”

For those who may not be aware, Fed Days have had a strong bullish tendency over the years, but as I mentioned last night, this has historically been the most bearish week of the year. Below I have listed all Fed Days that have fallen during this week in September.

There have only been 11 instances, and I don't see a substantial edge in either direction. Returns are lower than with a typical Fed Day, but not so much that it will change my approach. So rather than worrying about the week, I will apply some of the other filters I’ve used before when considering Fed Day edges.

Several Fed Day edges may be found using the Fed Study label on the blog.

A more comprehensive look is available in The Quantifiable Edges Guide to Fed Days.

The Weakest Week

The week following September options expiration has historically been the most bearish week of the year. I showed some detailed stats in my weekend report, but below is a graphic that shows how consistently weak this week has been.

Seasonality is not normally a reason to sell on its own. When combined with a long-term downtrend and a short-term overbought market, it can often provide a pretty good signal.

Long-Term % Change Charts of XIV and VXX

I saw the other day that UBS is launching 12 new VIX-based ETFs. Today I thought I’d show some interesting charts of VXX (IPath short-term VIX) and XIV (VelocityShares Inverse VIX short-term). These are not run by UBS, but they do use a similar approach to the UBS ETFs that are coming out.

Trading volatility can be exciting and there are some terrific swings available for short-term traders. Just Friday VXX rose 9.5%. But if you think these ETFs may be suitable buy-and-hold vehicles, the charts below may change your mind.

The 1st chart is a % change chart of XIV, the inverse VIX vehicle. It goes from the 11/30/10 inception through 9/9/11. After 6 months or so it began garnering some major attention and there were lots of articles written about it. And rightly so, since it posted gains of over 100%...

…but if you had bought at the inception and held through 9/9 you would have seen your investment rise by over 100% and then lose all its profits and now be down over 30%. That isn’t a pleasant ride. (Also interesting is that XIV did a 10-1 split in June when shares were near $170.)

Now let’s look at VXX. VXX has risen over 125% since hitting its low in early July. But let’s look at a % change chart over the same time frame as XIV to see how it has done during the last 10 months or so.

Even with the massive rally since July, an investment made on 11/30/10 would be down 7%. And if we go back to VXX inception?

The massive 2-month rally doesn’t look like much on this chart. Since VXX’s inception in 2/2009 it is down about 89% despite more than doubling in the last 2 months.

Some of the reason for the long-term failure of these ETNs is due to expenses, but much of it is thanks to the impact of backwardation and contango when they do daily rolls, as well as the mathematical difficulties in overcoming large drawdowns. (A 50% loss needs a 100% gain to break even. A 75% loss needs a 300% gain…)

I won’t get into backwardation/contango issues here, but below I have listed a few places you could learn more. The bottom line is that these products can be terrific trading vehicles, but 1) you need to understand them and 2) as I hope was clearly demonstrated in the charts above, buy and hold is typically not a good strategy.

Other sources of information:
I did a series of webinars for Quantifiable Edges subscribers on our members site a few months ago that explain backwardation and contango in detail and how they relate to these products. They are archived on the "Videos" page of the members site and the first one of these is even available to trial members. (Click here for trial subscription.)

Another good source of information is Bill Luby’s VIX and More site. Below are a couple of links to posts on his site that will get you started.

http://vixandmore.blogspot.com/2009/05/vxx-calculations-vix-futures-and-time.html
http://vixandmore.blogspot.com/2011/08/vix-term-structure-evolution-over-last.html

Performance After SPX and VIX Both Close Higher on a Monday

The SPX and the VIX typically trade counter to each other, so it's a little unusual to see them both close higher as they did yesterday. It does happen from time to time, and Monday is the most common day of the week for it to occur. This is because the VIX has a natural tendency to rise on Mondays. Even so, when it has occurred on Mondays it has typically suggested a short-term downside edge. The edge is more pronounced and lasts longer when the SPX is trading below its 200ma. Below are the numbers.




As you can see, the statistics suggest a downside edge over the next few days.

When Big Drops Come After Bigger Gains

When the market experiences a strong pullback as it did yesterday then there is often a tendency for it to bounce the next day.  This tendency varies greatly though, and depends on a large number of factors.  One factor to consider is how the market moved the day before.  Let’s look at a couple of equity curves that will demonstrate my point.  This first one shows results since 2003 of buying all 1% drops at the close and selling at the close the next day.

I ran this back to 2003 because that is about the time this equity curve began to show a consistent upside tendency.  Now let’s look at all times the big drop failed to close the SPX below the close of 2 days ago.

There’s different ways to look at this, but here is a simple one.  If today’s big drop did not wipe out yesterday’s gains, then it may have further to go.

What Yesterday's Partial Gap Fill Suggested

The market is gapping up large this morning so I feel a little silly discussing how yesterday's action suggests good things for the next few days.  Still, I did not note anything bearish  about yesterday so I thought I would show one of the studies that I found compelling that appeared in the Quantifinder as we approached the close.

There have typically been 1-2 days of buying following such setups.  It will be interesting to see if the upside edge exhausts itself early here or whether bulls can follow through and add to this mornings gains over the next day or so.

September Performance After Bad Augusts

September has a well-earned reputation as the worst month of the year for the stock market. Last night both Woodshedder and Michael Stokes provided nice historical breakdowns of September. For the 2nd year in a row the market has suffered a difficult August with the SPX closing down about 5.6%. This left me to wonder how September has performed following bad Augusts. Below I have compiled a list of all Septembers after August lost 4% or more.

There doesn’t appear to be a strong directional edge, but one thing that is evident in all of these Septembers is that there was high volatility. The 1966 instance saw the smallest range with the market moving a little over 5% from high to low. Six of the eight instances saw ranges of 9.5%+ in September. So I would not look for the action to dull this month.