Outside Gap Behaviour, Then and Now

By Lawrence

Many traders who daytrade emini S&P probably are familiar with the gap fill bias. What I am going to show here is how a particular type of gap fill, outside gap, has changed over time.

Gaps are created when a market open at a price different from the close of the previous session. Outside gaps are created when a trading day open outside of the range of the previous trading day.

Gaps are known to be closed very often within the same trading sessions they are created. Many daytraders lean on this behaviour as part of their bread and butter trades.

Over the past few years, especially since the QE has started, however, there are some changes to the outside gaps that traders should be aware of.

Following is a table of the outside gap statistics organized by every 2 years of data starting from 2001.

The column 2003 is the statistics collected from Jun 2001 to May 2003 and the column 2011 is the statistics collected from Jun 2009 to May 2011.

Notice the increase in number of days having outside gaps and that the consistent decline in the probability in complete gap fill in the same trading day.

The bias that survived the introduction of QE into financial markets, is the probability of half gap being filled in the same trading day. Although it is also declining, the bias is still strong enough (for now) to be considered as a valid trading setup.

Share

Comments
  • treed June 15, 2011 at 5:14 pm

    Lawrence, v. good info suggesting to at least light up at 1/2GF and be on the watch for resistance

    Have you ever looked at the probability of closing gaps of >10 pts vs <10? I haven't done the analysis myself, but I've heard a 10 pt gap referred to as a professional gap, likely leading to a gap 'n go day. It would be interesting to know if the 10 pt gap rule combined with the inside vs outside criteria provides significantly different odds of filling.

  • Lawrence Chan June 16, 2011 at 9:07 am

    The issue with 10 pt is that although it still works, it is too wide already as it cannot tell you the bias for those gaps that are way smaller (e.g. 5 pts, etc.) that are still not filled.

    Nowadays, we have many of these “gaps and go” setups with a smaller gap to start with.

  • You must be logged in to comment. Log in