Concept for an Automated Strategy Development Process

[jack hershey]

Quote from subes:

you provided this picture depicting volume as the pillar:
"the independent variable volume.png" http://www.elitetrader.com/vb/attac...&postid=3758646

traitor786 also mentioned using volume as a filter with his comment:
"P.S don't forget looking at volume during your screening. If you can program maybe you can get some ways to compare price and volume without having to give up a "seat in your living room". "

-- I guess you both talk about the same thing here. You suggest a filter that looks for a buy opportunity by measuring volume. From what I know of volume, it can be used to detect trend changes, because volatility of volume increases during these market phases. Thus the "suppression" you mean is a filter that prevents a false entry signal to take place if there is no volume to back up that trend change (which you describe as a new step in the cycle). Volume could also be used as a signal in that regard, not just as a filter.

I am not sure if you base your pillar on volume alone or on volatility of volume though. Anyhow, this is surely some interesting filter/signal to experiment with. Though as long as I have not validated this, I won't add it to the thesis. The explanations about value investing are based on Phil Towns books. I say there that more experimentation with rules is advisable and I guess this includes a potential filter/signal based on volume or volatility of volume.

Thanks for putting this in front of me. Correct me if I am wrong.

Yes volume is THE filter/signal. You saw that trends are processes aand go through a set of event in a specific order. As Braithwaite would have put it in bandwidth theory.

By surrounding the trading fractal (Mandelbrot) with two other fractals (slower and sub fractal faster), you can "SEE" cycles and have the trading fractal establish its "container" in the two defining first pieces of the independent variable. All fractals are interlocking in a fixed ratio of 3:1. One extreme is assigned immediately as expected the other extreme establishes the container. Not coincidently, this is where the "overlap" of trends ends and the new trend dominance (the opposite direction) is intiated.

Frenchfry (search for him if you wish) did this in a trend using just a few bars. It does not matter if he is precisely correct beecause his applied method is self correcting. (this facet of logic is deeply interesting). At this point, the trend continues or the container is violated or continued. So I took him through enough bars to "SEE" two more trading fractal trends (using an illustration).

Reason dictates that a trend overlap must end for a new trend to be legitiment (The proof is entitled "Clean page 4").

Market participants form markets and the heritage and nurture of those persons gives the market its precision. No randomness is ever measurable as a consequence. A book by an uninformed person explains a differnt level of getting fooled. Getting fooled by insignificant things is very common for the financial industry. This creates an always avaialble loser pool that is huge to say the least.

Our vision is closely coupled with our minds. Thus, it is not possible for the vast majority of people to ever "SEE" markets, ever. Sorting these people out is simple. They all are learning failure until they leave the scene.

In recent years a new name has appeared. It is PA which for those users stands for Price Action. It is a lagging indicator with signals. The dependent variable by its nature lags.

Suppression originates with Price. Measurements occur on volume when permission is granted. Containers are generated and used to determine continuation or violations. Violations end trends.

By completing the logic for the systematic operation of the market, a practicioner has a complete template and uses it to just take the full offer of the market by being a participant.

If you do that, you create the "insider profile". The Regulators then cite you and your lawyers educate the transgressors (SEC). In my case there was no incentive in the SEC to understand market operation (today it may be the same, still). I will include this in my book for laughs.

 

[subes]

Quote from jack hershey:
Go get a 30 day trial and learn all there is to know on the platinum v6.4 level of TN of Genesis. There are several problems you cannot overcome with the data as displayed. You will not realize this so just do the reports. Click on reports and then do some experiments. [/B]

I found out this refers to: http://www.genesisft.com/tn_platinum.php

After having a look at it I, it made me understand where you come from with your choice of words. It seems TradeNavigator has a more mathematical/physical sciences approach to describing strategies. My approach comes from the direction of object oriented design instead.
I also looked at the reports and it seems I already got most of the statistics they have in my reports already. I also posted a sample report before which I now repost:
http://invesdwin.de/public/Report1Week.pdf
I explicitly decided against looking at yearly, monthly, daily statistics which TradeNavigator provides. Instead I only look for depot (multiple instruments), instrument (stocks, currency pairs) and order specific (individual per instrument) reports to get various levels of detail. I guess this meets what you mean with me already thinking in Events instead of Time-Series.

The reports I was talking about I saw in Trading Blox were not ones for simple backtests like TradeNavigator has and which my reports already show. Instead I meant I am more impressed by the reports about walk-forward-analysis which I would like to incorporate into my platform.

 

[jack hershey]

Quote from subes:

I found out this refers to: http://www.genesisft.com/tn_platinum.php

After having a look at it I, it made me understand where you come from with your choice of words. It seems TradeNavigator has a more mathematical/physical sciences approach to describing strategies. My approach comes from the direction of object oriented design instead.
I also looked at the reports and it seems I already got most of the statistics they have in my reports already. I also posted a sample report before which I now repost:
http://invesdwin.de/public/Report1Week.pdf
I explicitly decided against looking at yearly, monthly, daily statistics which TradeNavigator provides. Instead I only look for depot (multiple instruments), instrument (stocks, currency pairs) and order specific (individual per instrument) reports to get various levels of detail. I guess this meets what you mean with me already thinking in Events instead of Time-Series.

The reports I was talking about I saw in Trading Blox were not ones for simple backtests like TradeNavigator has and which my reports already show. Instead I meant I am more impressed by the reports about walk-forward-analysis which I would like to incorporate into my platform.

Yes.

Walk forward will take the full offer of the market.

Events follow a specific order.

Then continuation has a failsafe termination (the presence of a violation of the continuation defined context).

The case of the End Effect can only occur in 1 of 10 fully defined contexts. The contexts follow a parallel order to the trend continuation order.

Finally, the overall mathematical basis is a RDBMS orientation which is dictated by the market and not determined by the user.

For stock trading, the front end is the filtering of all stocks to get a high money velocity Universe.

For one instrument markets, choosing the optimum fractal is key.

In all pool extraction, deductive science and the market dictated mathematics provides the invariant system for taking the full offer in the absence of noise, anomalies and flaws.

Not many people follow a critical path the the solution. The operation of the brain provides a suitable parallel.

My history, was perfectly serendipitous. My teaching department head in college explained the WSJ in detail and offered me a job (automatic tenure, I would guess). He explaind salary was insignificant for any life style. Trading provided all.

Secondly, at my only formal intial full time employment (IBM, 1957), I became a trader in 40 days, in half of which I began to use an account. I have only used one approach all my life (subject to technological support devices).

Lastly, when compters began to be used in the Financial Industry, my trading profile as a trader was confirmed. I am what is called by regulators (SEC) in citations an "insider trader". Well their evaluation is a correct "output" evaluation; it is an incorrect "input" evaluation. Fortunately, they did come up with a caveat: "I couldn't know that many people".

All this adds up to the fact that there is no entry bar to trading. The financial industry does not understand how to apply science and mathematics to markets. So people begin anyway at the base fo a tree; go to a branch; break off the branch and quit. FI corporations play the role of LEVI during the Californai gold rush; they sell pants (accounts to participate in markets for fees).

This was a fun exchange.

QED

 

[subes]

Thanks, for me too.

 

[subes]

I've added a new chapter "5.3 Comparison" where I compare other platforms to the one suggested by the thesis.

If anyone has suggestions on more platforms that should be explicitly compared there, I am open for that.

 

[traitor786]

Quote from subes:

PCs make things so much easier by making knowledge available en masse. But that knowledge is also a huge hindrance in ones development. You are bombarded with useless information and don't spend enough time with the important information to fully grasp it. Thus the PC also becomes the greatest hindrance.

I follow the advice of "The 4-Hour Workweek and Timothy Ferriss" to have an information diet. I don't watch TV, I don't follow news and I try to spend more time with knowledge that most other people do not consume every day. This is because I figured that knowing what the masses know makes you like the masses. Knowing what few others know, gives you an advantage. Today you can make important information publicly available for free, but it might still be forgotten because no one takes the time to worship it. So even the best system you publish will be disregarded by the masses and only be filtered and used by selected few individuals. That is why I think with more freely available knowledge, the world will still not change much. It will only make it easier for those people who realize they are put into a rat race to escape from it. There will always be rats running in the rat race, because it is still a lot of effort to escape it. Not everyone wants to put that much work into it. I am still a member of it though. :/

You, Jack Hershey, have given many of your pupils a way out of that rat race by teaching them position trading I perceive.

The issue that arises is the amount of information and the quality of information. Simply said there is too much, especially in fields like this. Actually the amount of relevant information is probably less then 1 %. with such a big data base, one can spend a life time in a sea of ideas. Each idea needs to be tested in theory. Ideas are presented for a reason. They must be catered to you. Ideas that follow a certain system appeal to a certain person. So ideas that you may find relevant or of importance appeals to you not because of its content (as it has not been tested) but because of how it is presented. This post may or may not appeal to you.

Now you have a dead man helping you, congratulation. If this appeals to you I wonder why you did not go to the cemetery for your answers.

In his defence his notion of age may be skewed. Age is used as a measure of maturity that is why different animals have different days in a year. also he has mentioned that his name is not exactly that of which you may assume.

You mentioned a program for back testing I was wondering if it gave you any insight in to things.

I have few questions about it. Feel no pressure to answer.

1.
how much does it cost after the 30 day system (no need geting wrapped up in to a software that will take more then 30 days to see the true benefit of if one is not planning to pay for it )

2.
Does it have data, how far back ,for what instruments, what is the quality of the data and what type of data is given :

Does it give :
price
volume
time and sales (every order)
colour of time and sales ( who pressed what button)
any thing else?

3. did you learn anything new with the data you uploaded. that you did not know before because of back testing limits

4.
Do you see any difference in the back testing.

Please do not feel you need to answer. or wast time to answer. It probably will not effect my life unless it stands out like a sore thumb.
I find that it may be a bit limiting my potential as it may need some focus and I have a list of outlets for such work.

You quoted me, I am glad I may have helped. but it should be known that I do not know of the relevance of anything I speak of. This is something that only testing could help you with. If I had a team of testers life would be much easier. Instead I work on a team of mathematicians and programmers.

just a reminder. for a while you were following a dead person. Im curious to know what is your criteria of how you filter information. Is it filtered by what you want to hear? or curiosity? popularity ? spelling punctuation use of the English language.

Nothing against any one on the thread, I even went out of my way to defend the dead by letting you know that he may have known he was once dead and indirectly alerted us of that fact.

 

[traitor786]

Quote from subes:

I've added a new chapter "5.3 Comparison" where I compare other platforms to the one suggested by the thesis.

If anyone has suggestions on more platforms that should be explicitly compared there, I am open for that.

Damn sorry , this post may have answered my questions.

 

[subes]

Quote from traitor786:

The issue that arises is the amount of information and the quality of information. Simply said there is too much, especially in fields like this. Actually the amount of relevant information is probably less then 1 %. with such a big data base, one can spend a life time in a sea of ideas. Each idea needs to be tested in theory. Ideas are presented for a reason. They must be catered to you. Ideas that follow a certain system appeal to a certain person. So ideas that you may find relevant or of importance appeals to you not because of its content (as it has not been tested) but because of how it is presented. This post may or may not appeal to you.

Now you have a dead man helping you, congratulation. If this appeals to you I wonder why you did not go to the cemetery for your answers.

In his defence his notion of age may be skewed. Age is used as a measure of maturity that is why different animals have different days in a year. also he has mentioned that his name is not exactly that of which you may assume.

You mentioned a program for back testing I was wondering if it gave you any insight in to things.

I have few questions about it. Feel no pressure to answer.

1.
how much does it cost after the 30 day system (no need geting wrapped up in to a software that will take more then 30 days to see the true benefit of if one is not planning to pay for it )

I am not currently thinking whether or how much this software should cost. Currently I am working to get it running.

2.
Does it have data, how far back ,for what instruments, what is the quality of the data and what type of data is given :

Does it give :
price
volume
time and sales (every order)
colour of time and sales ( who pressed what button)
any thing else?

For stocks everything you see on MSN Money (10 years reports summary and 5 years detailed) and Yahoo Finance (end-of-day price quotes). For Forex the tick data provided by dukascopy JForex. I guess you can investigate there. As soon as IB is connected, their data aswell.

3. did you learn anything new with the data you uploaded. that you did not know before because of back testing limits

I have not backtested anything yet regarding automation of value investing. The data I posted is from a milestone I had a few years ago. Getting the data and creating a value investing screener was the proof of concept there. Took about 3 months of coding...
I need to implement the concept first before trying the value investing stuff. On the other hand I am currently looking deeper into AmiBroker. It might allow me to do this faster without the generated strategy stuff. I will consider your suggestions about the data and value investing improvements. Thanks for that. Though they do not fit into the thesis because I only describe the stuff rudimentary there. I try to avoid subscriptions to data I have to pay for currently.

4.
Do you see any difference in the back testing.

See answer 3,

Please do not feel you need to answer. or wast time to answer. It probably will not effect my life unless it stands out like a sore thumb.
I find that it may be a bit limiting my potential as it may need some focus and I have a list of outlets for such work.

You quoted me, I am glad I may have helped. but it should be known that I do not know of the relevance of anything I speak of. This is something that only testing could help you with. If I had a team of testers life would be much easier. Instead I work on a team of mathematicians and programmers.

just a reminder. for a while you were following a dead person. Im curious to know what is your criteria of how you filter information. Is it filtered by what you want to hear? or curiosity? popularity ? spelling punctuation use of the English language.

I value when somebody takes the time to read my work. Just as I value your input. I guess the length of the conversation came from too many questions popping up while I was reading the answers.
This can speak for or against the conversation itself.
About how I filter, in fact I take in as much as I can and work out if it helps or not later.
Listening to lots of audio books and reading a lot of books also helps to figure out what information is valid or not. The process trains the perception I think. Also I had the same questions you have, though I mostly answer those to myself. Even if it is not a good answer I make, it helps me to focus on the important stuff. Less meta-work and more actual work. The answers to the meta-questions come along while doing the work. Everyone optimizes himself by trying various things. One does not have to think about it that much. Intuition should lead the way.

Nothing against any one on the thread, I even went out of my way to defend the dead by letting you know that he may have known he was once dead and indirectly alerted us of that fact.

 

[traitor786]

Quote from subes:

Yes sorry about not reading the complete theses. It did not pop up as something that may help me in where my current problems lie. No doubt I will love and need the info you supply in the theses. If I were to read books and let me intuition lead me the results would not be as good. I find the books close one off to other ideas. If you are reading about probability and maths and numbers you may miss a key component or even not know it exists. You came here with a goal and maybe now you have deviated form the goal. This is learning. but understand your path. reconfirm that it is right every now and then. Stick to the other posters here they seem to be know more and are doing better.

My work is worse then my writing.

 

[subes]

Quote from traitor786:

Yes sorry about not reading the complete theses. It did not pop up as something that may help me in where my current problems lie. No doubt I will love and need the info you supply in the theses. If I were to read books and let me intuition lead me the results would not be as good. I find the books close one off to other ideas. If you are reading about probability and maths and numbers you may miss a key component or even not know it exists. You came here with a goal and maybe now you have deviated form the goal. This is learning. but understand your path. reconfirm that it is right every now and then. Stick to the other posters here they seem to be know more and are doing better.

My work is worse then my writing.

I would not say I am off course. It is more like I always oscillate around my path and adjust course everytime I detect I am going the wrong direction. This also counts into the self optimization I do.