[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: An attempt to quantify patterns



PureBytes Links

Trading Reference Links

Hello Andrew and RTs:

Thank you for the warning advice, it helps me to be objective.  I think I
might be avoiding some of the pitfalls you mentioned due to my program's
design.  It's a program which acts more as a checklist and counter rather
than a predictive device.  

All I want is a program which automates and aggregates trading information.
This program would assimilate input from user's and present it as a whole.
The data would consist of identified patterns, support/resistance levels,
fib levels, reward to risk ratio, and suggested targets, etc. 

The main purpose is twofold, the primary one is to take some (most?) of the
subjectivity out of my trading decisions by standardizing (and automating)
my search and secondly, to provide an empirical method of assessing the
value of the various components.  However, this is still a tool within a
discretionary methodolgy.  As well, I see such a program as speeding up some
of my daily analysis.  Presently, I often stare at graphs with ambivalence,
because of the contrary indications.  Such a program is meant to quantify
the bullish/bearish power, and thus lessen my ambivalence.

As to what programming language I'm using, it's Perl.  This is my first
large scale program, although I have written other smaller programs which
calculated fib. levels, pivot points, etc.

Thanks again, for your input.  Any other comments/advice welcome.

Joe


At 02:24 PM 4/16/99 -0700, you wrote:
>This is truely an abitiouse project.
>
>I think you will find this is going to be very, very difficult.
>I ran a small sofware consultancy for sevral years and I know of two
>wallstreet
>firms that tried to do just what you are talking about, but they went much
>farther
>using nural nets in the process, (not as the main proccess, but as inputs
>and "number" crunching tools).
>They also used expert systems in the mix as well and they all failed.  both
>started small and kept adding
>new technology when the prier attempt failed.  I wish you the best of luck,
>but
>I beleave you are going to have a difficult time.
>
>Have you programed artificial intelegent systems before?
>If so, what language did you use?
>
>
>-----Original Message-----
>From: Joe Zelwietro <joe@xxxxxxxx>
>To: RealTraders Discussion Group <realtraders@xxxxxxxxxxxxxx>
>Date: Friday, April 16, 1999 12:32 PM
>Subject: An attempt to quantify patterns
>
>
>>Hello All:
>>
>> I'm working on a program which utilizes chart patterns to build an index
>>which will be one tool I use in taking trades.  Because of the dynamic
>>nature of patterns, rather then have the program search for the patterns,
>>the program accepts input from a user.  Therefore, the program's primary
>>function is to quantify the probability of a given direction.  I have found
>>the following patterns useful as indications of a probable price move.
>>However all patterns are NOT created equal.  Sometimes a price moves quick
>>and surely in the projected direction (eg. New Low) after a particular
>>pattern is elected, some patterns remain valid for quite awhile before the
>>price moves (eg. divergence between MACD or RSI and price).
>>
>>Using my experience and "Schwager on Futures" I have chosen several
>patterns
>>which have been useful. My question to all of you is:
>>
>>How would you rate the following in relation to each other?
>>Are there some that stand out for you as superlative?
>>
>>Could you assign a number (10 being VERY reliable ... 1 VERY Unreliable)
>>to the folowing:
>>
>>-Bearish Pattern Indicators
>>New Low
>>Downward Gap
>>Correction of Upward Move falls below assumed Support
>>2nd retracement of Downward move holding below .618 Resistance
>>Bearish Reversal
>>Downward Penetration of previous Trading Range Support
>>Neckline of H & S Top pierced
>>Downside breakout of rising wedge
>>Downside breakout of pennant (in a beartrend)
>>Spike High
>>Bearish Divergence with MACD
>>Bearish Divergence with RSI
>>C & H failing to penetrate resistance
>>Counter-Anticipated Breakout of Flag or Pennant downwards
>>Upside Three Tries Burst (Linda Rashcke calls this 3 little indians)
>>Bull Trap (return to mid-trend as confirmation)
>>Failed H & S Bottom (re-pierces neckline)
>>Upside Gap filled on the downside
>>
>>
>>-Bullish Pattern Indicators
>>New High
>>Upward Gap
>>2nd retracement of Upward move holding above .618 Support
>>Correction of Downward Move rises above assumed Resistance
>>Bullish Reversal
>>Upward penetration of previous Trading Range Support
>>Neckline of H & S Bottom pierced
>>Upside breakout of falling wedge
>>Upside breakout of pennant (in a bulltrend)
>>Spike Low
>>Bullish Divergence with MACD
>>Bullish Divergence with RSI
>>C & H Handle holding at Support
>>Counter-Anticipated Breakout of Flag or Pennant upwards
>>Bear Trap occurance (return to mid-trend as confirmation)
>>Failed H & S Top (re-pierces neckline)
>>Double Top pierced
>>Downside Gap filled on upside
>>
>>And of course if there are some patterns that you have found significant
>>that I haven't mentioned and want to share them, great.
>>
>>Thanks and Have a good weekend,
>>Joe
>>
>>
>>
>
>