Experimenting With Probabilities

Ever since I listened to the Avery Horton interview, I've been looking more into probabilities.  Scouring the web brings up some interesting results, everything from advanced financial probability concepts to forex signal software.  I took statistics, probability, and advanced mathematical courses in college but some of the information was way over my head and I just don't have the desire or time to explore the advanced information.  

I started a simple experiment with historical GBP/USD data since January 1st, 2007.  I wanted to find out whether the price went up or down based on previous price action.   It's easiest to give an example of what I tried to do.

First, find out whether price action was BEARISH or BULLISH between 2:00 am and 8:00 am EST.  Then find out whether price was BEARISH or BULLISH after this period of time between 8:00 am EST and 5:00 pm EST.  So I'm splitting the day into 2 time periods.  The first period (between 2-8 am EST) is a non-tradable time period.  The second period is when an actual trade might take place. 

What direction will the price go after this first BULLISH or BEARISH period? 

This may be a total B.S. experiment but I wanted to start somewhere.  Based on initial testing, in 2007, here are the results:

BULLISH 2am-8am period resulted in the price closing UP from 8am-5pm 66% of the time.

BEARISH 2am-8am period resulted in the price closing DOWN from 8am-5pm 38% of the time.

Looking at the results don't show any correlation between past and future price action in this experiment because when the first period was BEARISH, the second period actually turned out BULLISH 62% of the time.  On the contrary, when the first period was BULLISH, the second period turned out BULLISH 66% of the time.  I guess this shows that the trend in the GBP/USD was BULLISH throughout 2007 and trading in this direction would have been most desirable.   

Popularity: 1%