Wednesday, January 03, 2007

Trading Opening Gaps to the Upside

A few notes before we start the first trading session of 2007. As of my writing (5 AM CT), we'll see a large upside gap in the ES futures. Below we'll take a look at what has happened recently following large gaps to the upside.

First, though, I want to thank readers for their interest in the selected 2006 posts. In case you missed those links, here are the trading and market psychology posts from the first half of the year, from the third quarter, and from the fourth quarter.

Also take a look at the Trader Performance page of my personal site. The latest entry links the Peak Performance Seminar that I participated in for the Chicago Mercantile Exchange and discusses trading goals for 2007.

Finally, I'll be using the Trading Psychology Weblog during 2007 to summarize my trading outlook at the start of each session. I'll review the research and indicators I've found to be useful and try to put them in a coherent picture for the next trading day.

So now let's look at large opening gaps to the upside. I went back to 2004 (N = 752 trading days) in the S&P 500 Index (SPY) and found 76 occasions in which the upside gap was more than 40% of the average daily trading range from the previous 20 sessions. Interestingly, there was no directional edge from the open to close. The average price change following the upward gap was -.07% (37 up, 39 down). In all, 25 of the 76 sessions completely or substantially closed that opening gap.

When we look beyond the day session (open to close), however, we see a more bullish pattern. Specifically, if we look at the market close two days following the large gap opening, the average gain has been .23% (47 up, 29 down), quite a bit larger than the average two-day change of .07% (414 up, 338 down) for the sample overall. Almost all of that gain, however, has occurred when the upside gap has followed a winning trading session (N = 27); the average gain has been a sizable .48% (20 up, 7 down). When the upside gap has followed a losing session (N = 49)--as is the case this AM--the average two-day gain is only .09% (27 up, 22 down). That represents no distinct edge.

The pattern suggests that upside gaps may be most valuable when they are part of a bullish market, contributing to an upside momentum effect. When the upside gap is making up ground lost in the prior session, near-term results have not shown a meaningful directional bias. It might be possible to screen for stocks that are bullish the previous day and that gap upward the following session to take advantage of upside momentum. Indeed, if I'm not mistaken, we see some trader-bloggers making good use of just such a pattern. More on this gap trading strategy to come!

10 comments:

Anonymous said...

Speaking of the devil, it just so happens Trading Markets just had an article about gaps,

http://www.tradingmarkets.com/.site/stocks/commentary/guestcom/GapsLaps.cfm

b hong said...

You had previously reported data from Mark Fisher showing that Gaps >0,5% tended not to fill and that Gaps <3.5% did tend to fill. Is your analysis, today, consistent with that or is this new data reflecting a diferrent market condition from the time that Mark generated his statistics?
Always glad to see your new analysis.

Brett Steenbarger, Ph.D. said...

Thanks for passing that on, Cucca!

Brett

Brett Steenbarger, Ph.D. said...

Hi Bruce,

In this post, I am not looking at the absolute size of gaps, but at their size relative to the average range of the last 20 days. So this is a new analysis. Thanks--

Brett

NO DooDahs said...

If you remember, I was extra-long in anticipation of this.

Was thinking about trimming the sails tomorrow, but your analysis has me thinking about waiting til Friday.

Brett Steenbarger, Ph.D. said...

Hi NO Doodahs,

It's been choppy, but as long as the large traders are leaning to the buy side, I have trouble letting go of those longs!

Brett

b hong said...

Very interesting to see how this day played out. Your post about the possibility of underlying weakness was right on the spot. Now to see if the Bulls have any strength left :)

Brett Steenbarger, Ph.D. said...

Hi Bruce,

I think you're right. When the Dow moved to a new high this AM and the ES could not break and hold above that important 1440 area mentioned on the Weblog, sellers gradually came in and then institutions began unloading. This is significant selling, with high volume hitting bids. It's not unusual to get short, sharp short-covering rallies after such concentrated selling, but that's different from sustained, fresh buying interest.

Brett

Anonymous said...

Hi Brett,

That downside volume was pretty significant during the drop. Considering the news today wasn't all that bad, it sure looked like the big boys were looking for an out. The reversal towards the end had some volume too. Looks like a tug of war going on.

I remember a column of yours a few weeks ago talking about when to sell. This morning I was staring at a really nice gain and although there wasn't any specific reason to get out, I did and was pretty happy when everything dropped. I tend to trust my gut sometimes, probably not enough but it usually pays when I do.

The next couple weeks will be interesting to say the least.

Marc

Brett Steenbarger, Ph.D. said...

Hi Marc,

You're absolutely correct. Volume corresponds closely with volatility, and the early high volume was a key that this could be a volatile day. I'll be watching intraday volume closely; the added volume represents institutions getting business done and can lead to significant swings.

Brett