February 28, 2008
Pattern Recognition Across Timeframes
TEMPORARY DRAFT
Two classical bullish patterns in daily bars.
Timeframe shifted to weekly.
Selector line moved to highlight the former Pivot Lo value.
Timeframe shifted to Monthly.
Selector line moved to highlight the former Pivot Lo value.
Note that the 2nd and third Pivot Lo’s (not marked on the charts) are not significant in the monthly timeframe i.e. the monthly chart does not truly represent the underlying waves (cycles) of the daily chart (which I claim is the natural rhythm of the market notwithstanding that the intra-day charts tell us as much, if not more, about the behaviour of market participants).
Bonus Commentary
Comparing perfect equity with a buy&hold strategy for three weeks (using an earlier section of the above chart).
The charts are linked and in different timeframes.
Perfect trades are an approximation of Pivot Hi to Pivot Lo marked in colour with corresponding % change obtained via the AB studies tooltips.
For the sake of the exercise I cheated a little on the perfect swings.
For convenience I ignored inside and outside days (not counted as a swing) -annoying little things those ID’s and OD’s - messing up my perfect theories like that.
Filed by brian_z at 1:50 am under Uncategorized
(No Ratings Yet)
Loading ...
Hi, this is highly impressive! I was trying to come up with something similar but wasn’t as successful as you. Would it be possible to post the code? Thanks!
Hello Stokkkid,
Aplogies for the late reply.
Your comments got lost in transit.
This is a temporary post with mixed content.
I might extend the post in the future.
Which of the three topics are you interested in:
pattern recognition across time frames, perfect equity, or expectancy (geometric mean)?
My creative ability runs a long way ahead of my code skills so I don’t know if I can provide the code or not. What do you want the code for?