Concept for an Automated Strategy Development Process

Discussion in 'Automated Trading' started by subes, Mar 7, 2013.

  1. Zacks has a fundamentals database and backtester, it's not that expensive to use. I have not used it so can't comment on it.

    http://www.zacksrw.com/
     
    #21     Mar 10, 2013
  2. subes

    subes

    Regarding "It the reader supposed to fill in the gaps in the text with assumptions?":
    -> The purpose of this here is to detect gaps in the text so I can fill them in the thesis. If you have suggestions about what is missing in the text, I would be glad to hear about that so I can fill the gaps.

    Regarding "Do you have a set of illustrations that parallel the text?":
    -> I don't have a presentation ready for this yet, though I am going to create one for presenting this at my university in a few weeks. But this will most likely be in german then and will only be an excerpt of the thesis. Illustrations alone often only make sense when there is somebody presenting them and filling the gaps there. The thesis itself has the goal of being understandable without a presentator. I know reading a complete thesis is some work but a shorter version is not available right now. :/ Maybe it would be a good idea to create this in english aswell.

    Regarding "Do you have a device that tells you when the automation system has hit a stonewall and Why?":
    -> I don't know what you exactly mean by this, so I will aswer this in multiple viewpoints:
    1. The automation of the development process (or the platform) should not hit a stonewall itself because it is based on best practices from other successfull processes/platforms. It could only hit a stonewall if it turns out it does not save cost/time when developing strategies in the platform instead of the classical approach.
    2. If you mean the strategies itself by automation system, finding out if they hit a stonewall is in no way different than with any other strategy. During development, the reports should tell if you hit a stonewall, so you can adjust by trying other rules/combinations/parameters or by trying a completely different strategy.
    3. If it is during live-trading, the strategy risk building block should handle that as described in chapter 3.3 (which is explained in detail by the referenced book from Rober Pardo). Though this is up to the monitoring done by the developer/trader himself.
    4. If value investing hits a stonewall as an automated strategy, should become obvious early in the development process after the first testing results are available. Though since most initial shots at automated strategies are failures, it might still be possible to experiment with it to make it profitable afterall. This is what the strategy development process is about.

    The process will not tell why a strategy fails, instead the developer has to reverse engineer this information from the reports and his strategy design.
     
    #22     Mar 10, 2013
  3. MarkBrown

    MarkBrown

    exactly my point why the us has gone to crap. i give a+ why not just what the prof expected you to deliver how original.
     
    #23     Mar 10, 2013
  4. subes

    subes

    Thanks for your comment, though could you rewrite the second sentence or explain it?
     
    #24     Mar 10, 2013
  5. MarkBrown

    MarkBrown

    everything in that doc is 40 years ago i have gigs of docs from real phd's that probably taught your classes along with their resume. you think phd's are there because they want to be teaching? no they are there because they have failed to land a job on wallstreet.

    they expect to see what that document contains and for that the grade will be a+. it was an excellent job of cut and paste plagiary. not one original idea did i see? insulting i had to even view the word average. it may land a skin from the school but will land in the can at ren tech. send it there if you think i am brutal, the rocket launchers will come at you from 10 miles away.

    you should be proud though because you can dazzle just about everyone outside a quant firm and lower. please report back here and let us know your progress in the real world. it happened to me by accident but you might consider it. i had a heavy pipe destroy 75% of my brain when a child, it must have pissed off the remaining 25% - just saying.

    mark brown

    ps please take some hallucinogens that will induce paranoia so that you can come up with something desperately original - if for no one else but yourself. welcome to the real world it may actually be this tuff to get a job a micky's in the future.
     
    #25     Mar 10, 2013
  6. Thank you for your response.

    I did read the thesis and I asked the Q's because of: Q1 the gaps, Q2 omissions and Q3 errors in your approach.

    Your forward leaning forthright response is terrific.

    As you say in 4.1.4, it gets tough for the user (developer) to reverse engineer the automated development the quided computer is creating.

    For me, the feedback loops (See fig 4 missing, fig 5 okay, fig 6 poor and ommitted, and fig 17 (about as good as fig 4)) Are moe important than the forward flow and the API triple shell approach.

    The mechanism of market movement and movements ending are not opposites but instead are orthogonal. Running a portfolio comes down to "getting sunk capital free to exercise opportunities. It is an optimization of orthogonal variants.
     
    #26     Mar 10, 2013
  7. subes

    subes

    Thank you for the feedback.

    -- About the reverse engineering:
    It gets harder when you increase the degree of automation by not only generating combination of rules the developer wrote himself, but also generate individual rules. Genetic programming/neural networks produce rules that can not be easily understood by a developer. This is the degree of control that is revisited in the "5.1 Categorization" chapter. Here the developer again decides about where he wants to position his tradeoff.

    -- About the feedback loops:
    - Figure 4 is not an activity flow comparable to the other figures, thus not depicting a process with a feedback loop. It is more something between a use case and a sequence diagram.
    - Figure 6 is directly copied from Robert Pardo and not edited by me, instead only cited in the description of the figure. After the figure I comment on the incorrect feedback loop there in the paragraph starting with: "Also as a difference, the "Refine Strategy" step is modeled...".

    -- About the importance of the API:
    It is necessary to describe the API on which a developer works to define a concept for a platform he can develop on. The thesis has to balance the investors and developers view and thus describes both views. I would say they are equally important with the API being a lower level detailed view of parts of the feedback loop where the developer actively is required to do his work. Without this, the process would not be complete or not sufficiently described to be used.

    -- About the orthogonality:
    I do not understand where this refers to or how it affects what the text says. Can you elaborate this a bit further? Thanks!
     
    #27     Mar 11, 2013
  8. subes

    subes

    I am sorry for what happened to you in the past, but I don't think offending me here is the right basis for a discussion...

    I would say the original things in my thesis are among others:
    1. My research regarding automation of value investing for which I have found no literature yet.
    2. A new approach to an automated strategy development process.
    3. A modular strategy API that increases the comfort of developing strategies and allows to resuse complex rules. This saves development cost.
    4. A new method of applying variability models and finding an object oriented representation for them.

    The "plagiarized" ideas I use in the thesis are always referenced and marked as being derived from the specific authors. This is what is used in scientific research to build upon existing knowledge without getting the credit on oneself. Thus to respect the original works for their huge contribution.

    If you have a passage in mind that might be plagiarized, I would be very grateful to know which that is. I definitely did not plagiarize anything on purpose and my work would be a failure if it contained plagiarism. If there is literature that I did not find during my literature review and it contains knowledge I came up on my own, I would love to reference the existing work instead of giving myself credit for it. I would have spared myself some time if I had known the reference before and surely it will be very helpful to me when being able to read it now afterwards.

    Thanks in advance.
     
    #28     Mar 11, 2013
  9. MarkBrown

    MarkBrown

    wow i owe you a big apology - i just realized that your just to young to know that everything in that doc "everything" is old. it's so old i forgot the name of the software companies that have already not only done what you propose but been in business had customers for years and failed. but i did finally remember a few, LIM in austin texas, technalyzer in sarasota florida < still in biz, ptolemy project named "sss" by no other than the famous and brilliant http://en.wikipedia.org/wiki/Tom_Lane_(computer_scientist) and god kiddo i can think of many more your so like born yesterday.

    but have relief you will still pass and get an a+ cause that what our educational systems do ask ny who's 90% of hs grads cant even read and write.

    btw you misspelled resuse < not good since i am a dropout with no degree here. lol

    m

    ps hey when you get some time ask your prof if he and his grads are so smart why aren't there more than one ren tech? why isn't the financial world and the entire world run over by grads writing papers like this? instead of working at wendy's? wake up that was the blue pill man!
     
    #29     Mar 11, 2013
  10. subes

    subes

    I am sorry, but it is not sufficient to just name persons that might have done this, without me being able to verify that. Can you give me links to products I can look at or papers I can read? If you want you can also PM me links to PDFs.
    Or tell me more specifics about the works I supposedly plagiarize. Or you can send me contact information so I can send my thesis to the persons for a review to see if I accidentially plagiarize them. You could also send them the thesis yourself with my contact information if you like.

    Also I don't have the same view on teachers as you do. Sure there are ones that failed in the real world and thus went for a teaching position, though not everyone deserves that generalization. There are teachers who are successfull in companies and additionally teach lessons to give parts of their wisdom to the youth. Also there are very good researchers who do an excellent job on the front. Also there are ones that simply love teaching and find their profession there. Without these people our education system would not work and our society would loose a very important pillar.

    Anyway there are also people for whom academia is not the right learning environment, which can be criticized to produce mindless zombie slaves for the big company and government. Though this is also a too rigid generalization that does not hold true in all cases. You can be proud of yourself that you learned early that the school system is the wrong environment for you and you managed to become successfull without it. But this does not mean that your path is the right path for everybody.
     
    #30     Mar 11, 2013