Machine designed strategies. Do they work?

Discussion in 'Strategy Building' started by v75z52, Mar 20, 2012.

  1. alexvnew

    alexvnew

    Thanks I will try the demo. This software is expensive but I will do it just out of curiocity. I think price patetrns may have some potential. I know indicators do not have any chance for real performance.
     
    #131     Apr 17, 2012
  2. Indeed. Registered March 2012, 12 posts and all posted in this thread. :D
     
    #132     Apr 17, 2012
  3. How true is that. The product touted here is snake oil. Complete overfitting and I repeat myself in saying that such "optimizations" can be had free of charge in R and similar statistical packages. Heck, even Python offers many free libraries that provide access to any optimization routine you could ever ask for (stuff that goes most likely way beyond the sophistication level of this pitiful software pushed here.

    Amateurs stay far away I can only say.

     
    #133     Apr 17, 2012
  4. you are wrong: Those algorithms curve fit the heck out of any given time series, NOTHING MORE. There is not a single hedge fund under the sun that would every consider using such amateurish software.

     
    #134     Apr 17, 2012
  5. Lol, you turn the poster's words in his own mouth. He claimed SUCCESSFUL traders WONT SELL SOFTWARE, I bet there are many software vendors that sell trading applications who never traded before. Focus and getting the point straight I consider a basic requirement for success, including programming and marketing software...enough said.


     
    #135     Apr 17, 2012
  6. You sound excessively dogmatic. The highest principle in life is not to generalize if you want to succeed. Bloomberg was a successful trader in ML but he created one of the best trading software and became very rich. Actually that was the first vendor to provide analytics online for traders of bonds and options. Why do you think he gave up his analytics edge and created software? Is there maybe something you do not understand? Do you know young dogmatic man that the first Bloomberg terminal which I used to trade Eurodollars ran on FORTRAN 77? Not R, not C++, not anything of the sort you cocking around?
    Imagine if Bloomberg thought like you do.
     
    #136     Apr 17, 2012
  7. I'm glad someone has finally told me, and with such good arguments, after me posting dozens of posts explaining my opinions and findings. Thanks for that! ; )
     
    #137     Apr 17, 2012
  8. ssrrkk

    ssrrkk

    I think selling tools (not trading algorithms) makes some sense if your trading profits are in the low 6 figures. It lets one diversify higher risk trading with a low risk income stream. But I guess if your trading brings in upper 6 figures or 7 figures, then it probably would not make sense to waste time writing, marketing, and selling software.
     
    #138     Apr 17, 2012
  9. *** Bloomberg NEVER worked at ML, he worked at Salomon Brothers which was later acquired by Citi
    *** Bloomberg NEVER was the the first vendor to provide pricing and market analytics. Reuters was earlier among several others that were later on taken over by the likes of Thompson Reuters & Co
    *** I spelled out FACTS, nothing more, nothing less, not sure what "cocking around" you refer to. YOU AT LEAST GOT YOUR FACTS PLAIN WRONG
    *** a snake oil salesmen is someone selling an expensive piece of arse without trial, without any groups of users who have extensively used it, other than 1-12 post self-created users that are perused for testimonial purposes. This is what this thread is mostly about unfortunately.



     
    #139     Apr 17, 2012
  10. lol, which findings? That you ended up with an overfitted model? Hmm, did I not exactly tell you that this is what this software is all about? And you are one of the poor souls willing to shell out 60 grand for that? You must be stupid or in the same bed with the snake oil salesman.

     
    #140     Apr 17, 2012