Developing a Trading Entry using Artificial Intelligence
By Ron Jaenisch
There are various forms Artificial Intelligence. One of
them is called Engineered Natural Intelligence.
If you ask Professor Silva at UC San Diego to describe it
to you, you will receive a rather complex answer, which focuses upon how the
brain works along with a discussion of geometry.
What he leaves to others is the all-important collection and use of
data, that is specific to the application.
In this case study we will look at the development of
trading system entry signals and the type of data that is used. The reader needs
to understand that this case study was not hypothetical and resulted in an
actual trade transaction. It is written in a format that could be understood by
those that are not familiar with machine learning.
Levels of data
There is the basic data……..tick by tick with time………..very
level 1.
At level 1 we have Big Data. In this category of data we
can input pretty much anything in the most granular form. As a market technician
I am excluding 99.9999% of data available. I am focusing upon the time and price
of ticks for level 1 data.
This data is then turned into level 2 data which has
different forms of the level 1 data.
The level 2 data that is being used is the daily format of
the tick data, in the chart below.
Level 3 data consists of various tools to generate higher
level data from the level 2 data. The tools used for this demonstration are the
Median Line, Parallel lines and Warning Lines. These tools are applied to the
data in a specific manner.
The relationship of level 2 data is combined with these
level 3 tools to generate level 4 data.
In the example this would be a level 4 data pattern, called
the double trouble. This is a pattern that often occurs prior to a price
reversal that is 5-25%.
Also in this case the level 4 data is the relationship of
price to the Median Line and the lines that are
Parallel.
An action is generated by level 5 data. This uses all prior
levels of data to determine the action, which in this case is the entry into the
trade.
Developing a Trading Entry using Artificial Intelligence
By Ron Jaenisch
There are various forms Artificial Intelligence. One of
them is called Engineered Natural Intelligence.
If you ask Professor Silva at UC San Diego to describe it
to you, you will receive a rather complex answer, which focuses upon how the
brain works along with a discussion of geometry.
What he leaves to others is the all-important collection and use of
data, that is specific to the application.
In this case study we will look at the development of
trading system entry signals and the type of data that is used. The reader needs
to understand that this case study was not hypothetical and resulted in an
actual trade transaction. It is written in a format that could be understood by
those that are not familiar with machine learning.
Levels of data
There is the basic data……..tick by tick with time………..very
level 1.
At level 1 we have Big Data. In this category of data we
can input pretty much anything in the most granular form. As a market technician
I am excluding 99.9999% of data available. I am focusing upon the time and price
of ticks for level 1 data.
This data is then turned into level 2 data which has
different forms of the level 1 data.
The level 2 data that is being used is the daily format of
the tick data, in the chart below.
Level 3 data consists of various tools to generate higher
level data from the level 2 data. The tools used for this demonstration are the
Median Line, Parallel lines and Warning Lines. These tools are applied to the
data in a specific manner.
The relationship of level 2 data is combined with these
level 3 tools to generate level 4 data.
In the example this would be a level 4 data pattern, called
the double trouble. This is a pattern that often occurs prior to a price
reversal that is 5-25%.
Also in this case the level 4 data is the relationship of
price to the Median Line and the lines that are
Parallel.
An action is generated by level 5 data. This uses all prior
levels of data to determine the action, which in this case is the entry into the
trade.
Profits with Andrews Geometry
By Ron Jaenisch
My friend Professor Alan Andrews is best known for what
traders refer to as the Andrews Pitchfork.
It is simply three lines that are parallel, with the outer
two lines being equidistant to the Median line in the center. Each of the three
lines start at pivot points, as is seen in the above silver chart (chart #1).
Professor Andrews, who taught engineering at the University
of Miami, contended that price will make it to the median line 80% of the time.
If price does not make it to the Median line, then it will make up for it when
it reverses and goes in the opposite direction. Computer studies show that his
concept can be used to build systems that are worth trading.
This brings up the question……….Is there a way to predict
when price will not make it to the Median Line and quickly go in the opposite
direction, to profit from that?
Markets tend to go in a sideways pattern prior to a decisive move that is quick
and strong. What does a trader usually see occurring prior to a decisive strong
move, that brings home the bacon? If there is such a pattern, how can it be
described so that a trader can quickly understand it?
As I think back, there were concepts that he taught in his
60 page manual for the public and many other concepts he taught privately at the
kitchen table. A concept that was worth remembering is how to know when prices
are likely to make it past the median line and then strongly past the far
parallel.
An example is his sliding parallel concept.
In some cases, after drawing the pitchfork, price will go outside of the
pitchfork. As long as it does not go past the pivot point where the pitchfork
was drawn from, a sliding parallel line (SH Line) is drawn from the extreme of
that small move.
The NY Harbor heating Oil (chart #2), has two examples. One
is upsloping and one is down sloping. Note that in each case price stayed within
the SH line on a closing basis for a considerable length of time. In each case
the trader had the opportunity to generate handsome profits from the move.
Chart #2 is the same as Chart #3, with a few minor
exceptions. The first pitchforks
were removed and the next pitchfork in each case was added. This helps the
trader to know how far the move might go prior to making a reversal from which
the SH line will be drawn. Note that in each case price did not make it to the
red Median Line. This is the concept he taught privately.
What this means is, if the trader places a trade near where
the point where he thinks he will draw a future SH Line, he can place a stop
past the appropriate Median Line…and later past the SH line. Since the markets
are fractal in nature, Andrews’ concepts are then used on a smaller time frame
to determine the most likely point where price will reverse, which would be
before the Median Line.
For traders that use hourly charts, an example is Chart #4
and Chart #5, the profits are fast and furious and the risk is manageable. To
see the actual real profits chart that was used. See Chart #1.
After some thought, these insights will
bring the trader to other questions:
what patterns occur prior to this type of
trade?, what is the target for the trade?, what are the results of
computer studies that have been done on this strategy? what markets does it work
best in? …..and of course what is the best way to implement this strategy?
………together all good questions for a webinar.
This month new advanced andrewscourse.com course members
will see the in depth video that covers the technique and answers many
questions.
Above is a snapshot of an actual account. It shows that the
Soymeal short trade was taken near the SH point.
Can Advanced Andrews techniques help your trading?
-----------------------------------------------------------------------------------------------------------------------------------------
Ron Jaenisch spent time at the kitchen table with Professor
Andrews. He teaches for Andrewscourse.com, manages his family office accounts
and designs Andrews technique software.