Gap Fade Stats for All of 2008

Feb 5, 2009: 5:23 PM CST

2008 was a record-setting year in many ways.  Volatility reached extraordinarily high levels and the broader US Equity Indexes lost 40% in a single year as the Credit Crisis spread to the broader economy.  Let’s focus in on Gaps for the moment and form composite stats to see how the Gap-Fade strategy fared in 2008, looking at raw gap-fill percentage data.

For this simple study, we’ll be looking at raw overnight price changes and using the DIA (Dow Jones ETF) as our proxy as we have done all year.  For deeper analysis, it would be better to compare percentage gaps but from a trading perspective, it can be easier to look at raw price changes in terms of setting up trades, stops, and targets.

We will define a gap as an open that is $0.25 or greater up or down from the prior close (that’s roughly 25 Dow Points).  We’ll look at a few more parameters but $0.25 will get us started.  We will define a successful “Gap Fill” as occuring when price – at ANY point in the trading day – equals yesterday’s closing price.  This does not take into account stop-losses, trailing stops, or any other strategy.  We need to start at the raw data to see if the strategy has potential edge than can be developed further into a profitable strategy through using fixed stops, trailing stops, time stops, or any other strategy.  I’ll let you develop that on your own.

Here is the Excel data for all trading days in 2008 having a gap greater than $0.25:

2008 Gap Fill

Of the 252 trading days in 2008, 195 of them saw some sort of gap, meaning 77.38% (over 3/4) trading days had an overnight change of at least $0.25.

Of these 195 days, 116 resulted in an intraday gap fill, giving us an ultimate gap fill percentage of 59.49%, or roughly 3 out of every 5 gaps filled intraday.

The edge was slightly greater in up-side gaps that were filled, meaning price initially gapped up but then declined to trade equal or beneath yesterday’s close.  This is what we would expect in a down-market of 2008. By the same token, fewer down gaps (54%) had price rising to fill the gap intraday.

There were roughly an equal number of up (97) and down (98) gaps, so there was no significant difference there.

However, we see something interesting which is oddly characteristic of Bear Markets.  The average upside gap was $1.75 while the average downside gap was $0.95, meaning making money in 2008 was clearly not as easy as “get short, get rich.”  Bears (sellers) had to contend with extreme volatility both up and down.  There were times in 2008 when we’d make a record percent or point loss day… only to be followed right up with a record point or percentage UP day.  Though there were some strong down-moves, there were a good number of “shockers” which ripped the stops out from short-sellers.  Remember that Short-Selling on financial stocks was banned for a period in 2008, creating some massive upside gaps.

Let’s define our gap more stringently.  Let’s record the morning action as a gap ONLY if it’s greater than $0.50 from the prior close.  Let’s see if that gave us an ‘edge’ greater than 50% (random chance):

DIA Gap Fade over 50c

We still see the ‘edge,’ but it’s reduced.  If you do any sort of study on gaps, you’ll find logically that the larger the gap is, the less likely it is to be filled.

In this case, there were 138 (or 252) days which had a gap greater than $0.50, and of those, 73 gaps filled which returned a gap-fill percentage of 52.90%.

Strangely, there was a ‘flip’ in the percentage of up and down gaps.  There were 81 down gaps to 57 up gaps (remember they were equal at $0.25).  Now, more down-gaps than up-gaps filled which is interesting.

If we raise the criterion to $1.00…

The edge drops to less than random.  There were 72 gaps (up or down) greater than $1.00, and of these, 33 gaps filled intraday, giving us a gap-fill percentage of 45.83%, losing the edge of the strategy.

Remember, odds of a successful gap fill drop off as the size of the gap increases.  Think of it logically – which is more likely to fill… a gap of $0.10 or a gap of $2.00 (in the DIA)?  Just by random volatility, a $0.10 gap is more likely to be filled than a $2.00 one.  Often, large gaps (that go unfilled) are the precursors to “Trend Days” so keep that in mind.

It’s unlikely that 2009 will be as radically volatile as 2008 but there’s certainly no guarantees that it won’t.  Keep this in mind when developing your trade ideas and specifically when employing the gap-fade morning trade.

Corey Rosenbloom
Afraid to Trade.com

Keep up-to-date with the Afraid to Trade Feed.
Follow Afraid to Trade on Twitter:  http://twitter.com/afraidtotrade

3 Comments

3 Responses to “Gap Fade Stats for All of 2008”

  1. Jerry Says:

    Very informative information. Have you ever tried to calculate the percentage of gaps filled using a longer time period. (i.e. 5 trading days).

    Thanks for all the great info.

    Jerry

  2. Corey Rosenbloom Says:

    Jerry,

    No, but that’s a good idea and I can code that in Tradestation.

    I’ll try to do so and publish the results when ready.

  3. AtT Best of 2009 Part 1 | Penny Stock Trading System Blog Says:

    […] Gap Fade Stats for All of 2008 […]