Daily Commentary - Posted on Wednesday, March 25, 2009, 10:59 PM GMT +1

5 Comments


Mar Wednesday 25

Rollercoaster ride w/the S&P 500

Rollercoaster ride w/the S&P 500

Following the S&P 500 during Wednesday’s session, one may feel like being on a rollercoster ride on a folk festival. The market opened almost unchanged, posted an intraday high +2.54% above yesterday’s close (and a higher high above yesterday’s high as well), only to loose all of it’s gains and some to post a low -1.85% below yesterday’s close (and a lower lower below yesterday’s low as well), to finally close up +0.94% for the day. Wilder’s True Range exceeded 4% today, 4.39% to be exact.

Wilder’s True Range (TR) is defined as the greatest of the following:

  • today’s high less today’s low;
  • the absolute value of today’s high less the previous close;
  • the absolute value of today’s low less the previous close.

In order to evaluate if this pattern may provide an edge for tomorrow’s session (Thursday), I checked for all occurrences since 01/02/2001 when the S&P 500 posted a higher high, a lower low (than the previous session’s high/low) with a True Range of 4%+ (greater than 4%). There were only 10 occurrences, but the sessions following a session which fulfilled the setup mentioned above share something in common:

  1. They ALL posted a (significantly) higher high (would be at least 826.78 for Thursday’s session) ,
  2. They ALL posted a (regularly extraordinary) higher low (the minimum of those 10 occurrences was 3.65% above the previous session’s low), and
  3. The True Range for those sessions regularly exceeded again the then current average True Range (in the majority to a greater extent).

Unfortunately the sample size is very small (and hopefully there is no data issue), so I wouldn’t read too much into these figures, but it is definitely something to keep in mind for Thursday’ session, because in the event tomorrow’s session would queue up with those sessions listed below not breaking ranks, we could expect at least an intraday run up above Wednesday high of 826.78.

# Date Higher High Lower Low Close True Range
1 12/08/2008 +4,45% +7,86% +3,84% +4,85%
2 11/14/2008 +0,42% +6,25% -4,17% +5,16%
3 10/29/2008 +3,13% +9,11% -1,11% +5,07%
4 09/19/2008 +4,46% +7,02% +4,03% +4,86%
5 01/24/2008 +1,20% +5,06% +1,01% +1,56%
6 03/18/2003 +0,48% +3,65% +0,42% +1,11%
7 10/11/2002 +4,56% +4,87% +3,91% +4,90%
8 08/15/2002 +1,42% +4,79% +1,15% +1,64%
9 07/25/2002 +1,13% +5,21% -0,56% +4,47%
10 01/04/2001 +0,18% +4,28% -1,06% +1,57%

(‘Date’ represents the date of the session AFTER the respective session which fulfilled the setup mentioned above, and ‘Higher High‘ and ‘Lower Low‘ represent the percentage change compared to the previous session’s high/low, and a positive prefix concerning the ‘Lower Low‘ means there in fact was no lower low at all, only a ‘higher’ low)

Successful trading
Frank

Comments (5)

 

  1. Renee Ann says:

    Great post. This is my first day to read your blog and I look forward to more. : ) I subscribed to your RSS.

  2. Aesus says:

    First of all, great blog. I am a numbers nerd and love crunching but
    On this data set, the lack of samples, as you mentioned, is troubling. Truly robust samples are well over 100 no?
    Also, 50% of the available data points happened in one year which could point to a shift in how this data set behaves.
    Finally, as the closes were almost evenly split, looks like flipping a coin might be just as predictive?
    Perhaps it might be useful to look at the standard deviation of the changes on the day after the ATR spikes and compare them to the previous 20 days or so? Then, look at the same for the day the ATR spiked and see if there is anything correlative there? Just a thought.
    With such a small data set, I would find the results to be statistically not correlative.

  3. moneyfriend says:

    Thanks so much for sharing! Congrats on your new blog!

Leave a Reply

Your email address will not be published. Required fields are marked *