No registration required! (Why?)

Open to Close Change for the Day Trader


[Originally posted 23 June 2006]

In his blog TraderFeed Brett Steenbarger discusses Opportunity During the Day Trading Session. He states that in his calculations of the SPY over 834 trading days that we have the following open to close percent changes:

% of days Open to Close Change
40% < 0.3%
60% < 0.5%
85% < 1.0%

These figures give us the basis to make some assumptions. If we take the first two figures we notice that for each extra 10% of days the Open to Close Change increased by 0.1%. So we could extrapolate that out and assume that if we added another 25% to 60% and got to 85% that we would also add 0.25% to 0.5% and get to 0.75% Open to Close change.

As you can see, this is not the case. The Open to Close band needs to be widened at faster rate as we start including more days.

Do you know what this means for a day trading? Post a comment if you think you know the answer.
quote:
By omni72 on Thursday, July 13, 2006 9:46 AM
i'm not answering the question, but instead posing another question: how often does a market close at one end of the daily range after opening at the opposite end of the daily range? for example, how often, after opening in the lower 15-20% of the daily range, will a given market close in the upper 15-20% of the daily range? is there a percipitating scenario that produces a higher probability of this event? does a particular type of range produce these divergent closing? does a particular type of day follow such an extreme day? how many questions will i ask before i accept the fact that i have asked more than one question? :D

take care :)

omni

quote:
By day trading on Thursday, July 13, 2006 7:38 PM
:) The Trend% figure measures what you are talking about. The Trend% is (close - open) / (high - low) shown as a percentage. So, for example, if we opened at the low and closed at the high then the Trend% would show as 100% trend. If we opened and closed at the same price then the Trend% would show 0%. So by eyeballing a chart of the Trend% we can see what is most common and what follows what.

Here is a Trend% chart of the Emini S&P500:
http://www.deltat1.com/DailyNotes/trend.htm
id="quote">
quote:
By omni72 on Thursday, July 13, 2006 7:36
PM
very cool ;-)

thanks for pointing out the Trend%. comparing the Trend% to the Daily Ranges, it is interesting that the majority of the big trend days (>70%) are also big range days. in other words, if you are on the right side of a big range day, there is a high probability you will be able to capture the majority of the daily range. oh, sure, there's the whole process of getting on the right side ... but that's the fun, no?

on a tangentially related topic (uhm, in a way ... ) have you looked at typical price patterns/behaviors based on the day of the week or day of the month? for example, if Tuesdays are typically a range expansion (greater than an average daily range), does that mean that there is a high statistical probability that you will witness a high Trend% on Tuesdays?

again, thx for the Trend% heads up. very good stuff. as usual. :-D

take care :-)

omni

quote:

By day trading on Thursday, July 13, 2006 7:45 PM
I have never made a comparison of Trend% to Range so that is an interesting observation of yours. That is a good application for an XY chart which allows you to see if there is a linear relationship between two attributes as you have just suggested - which I suspect there is.

The output of my back testing system breaks down the results of each test in as many ways as I could possible imagine. For example, one of the breakdowns is by day of week, another by Trend%, and yet again by Special Days such as Full Moon or FOMC Fed Days. When I did the Gap Fade study I looked at the results of the probability of gaps filling on each day of the week to see if there was any day that significantly differed from the others.

Generally, from what I have seen, I have not been able to isolate any long-term patterns associated with a particular day. I have found that these "it happens on a Thursday" type of scenarios are often short-term phenomenon that when tested over several years usually even out on each day.

When doing that sort of measurement you also need to take into account that there may not be as many Mondays and Fridays in the trading week as other days so you need to compare your samples to total Tuesdays and total Mondays etc. so that the holidays that happen on days next to the weekend don't skew your results.