Taking proprietary strategies from discretionary to systematic......

Discussion in 'Risk Management' started by Arcap, Sep 22, 2009.

  1. Arcap

    Arcap

    To all that have kicked in their input/thoughts both here and via PM, it is much appreciated! This has definitely given me a number of things to consider and rang a few bells.

    THX-
     
    #11     Sep 24, 2009
  2. It is remarkable, rather useful idea
     
    #12     Nov 12, 2009
  3. 1) Unless you can code it yourself you are going to have to pull down your pants!

    2) 15 years who shives a git! Sorry for being harsh here, but as somebody who has taken a couple of dozen trading systems from discretionary to systematic the traders I trusted the least were the ones who were super secretive with their strategies because they were usually the ones not making money. The ones that make money trust me since they know there is no gain for me to share this with the rest of the world...

    3) Find a developer who is not hungry. You want to find one who makes money themselves. For example I could help you take your system to discretionary using Excel and .NET (send me an email). Will I take your strategy? No, because I already make money and have no interest in taking somebody else's strategy. My background used to be book writing, and still speaking and if there is one I do not do as a rule is take IP. So why am doing these jobs if I make money? Simple I don't want to work for an investment bank (I have), but I want to make contacts to traders. I like having my own personal network of contacts of people who actually make money. As one trader has said to me, "The reason I like talking to you is because being a trader is a lonely thing and its kind of nice being able to talk to somebody who actually knows what the f**k they are talking about." BTW its not like we have daily chats, just every now and then questions, updates, etc...

    4) NDA's will not help you since we are a global world and hence are you going to go after a trader? Not likely.
     
    #13     Nov 12, 2009
  4. OP you could just partner up with a programmer. Give him a cut, when things go wrong he will be all over it...

    Also, is there a site where I can read up on this TT Autotrader? I know nothing about programming but I'm alright with excel.
     
    #14     Nov 20, 2009
  5. CyborgTrading

    CyborgTrading ET Sponsor

    Cyborg allows you to build your own strategies without learning to program or script. This may be something that could interest you. It is designed for traders that want to take a discretionary strategy and automate all or part of it.
     
    #15     Nov 20, 2009
  6. t1ck3r

    t1ck3r

    If the development firm you deal with is reputable and plans on staying in the financial services business a NDA will cover your risk. In my experience traders often discount their personal value added in running a system. Most systems we automate for traders have some 40 parameters that will vary based a products volatility and liquidity. When a trader has a successful idea, the risk of a NDA combined with the use of a robust fast architecture is well worth it.

    Respecfuly,

    greg
     
    #16     Jan 7, 2010
  7. you could always get 1guy to build the frame work minus a few important details. Tell him you want to plug in some other parts so he can keep it user friendly, Then get someone to do the other part.

    I did that with an ORB system. Got one guy to do the general ORB system and another to code the filter bit. each one pretty much useless without the other. But put em together, and presto.

    This is probably an oversimplfied example though, but it could still be done.

    Perhaps even get him to build in a filter that actually kills the performance and replace it with the correct one. He will build it, see that it sucks, think your crazy for wasting your time on it and then forget about it....
     
    #17     Jan 7, 2010
  8. As others stated, you will have to either learn how to code your rules or pull your pants down and let a programmer have access to your trading rules for them to code it for you. However, there's one more option for you (read further below).

    With that said, I've been having the exact same problem as you although I sure the reasons why are different...missed trades do have impact on our wallet and profit curve.

    You need to sit down an examine closely why you're missing trades...is it trading related (e.g. busy designing something new) or personal (e.g. late arriving to the computer because you're now in charge of taking the kids to day care). Next, after determining the cause of your miss trades...you need to find a solution if pulling down your pants doesn't sit well with you.

    The reason why I am missing trades is the same reason as the past...there's a new addition to my family and the late nights of helping the spouse has a toll on my sleep pattern. It's part of the game of being a trader for those raising a family and it will pass (kids get older). I've had two bad experiences with programmers and will never go that route again. Therefore, I feel better with just resolving the problems of why I'm missing trades instead of looking for automation.

    Mark
     
    #18     Jan 8, 2010
  9. LeeD

    LeeD

    This is probabbly best advice. If you are in a position to hire a full-time developer, it can work. The thing is most people aren't entrepreneurial and prefer a stable paycheck every month.

    Another thought is selling a system for a persom who is not in the business (of selling systems) is a massive effort with potentially zero pay-off.

    Regarding Do-It-Yourself, NinjaTrader is free if you have one of the supprted data feeds. You can construct lots of strategies yourself visually if you can break down the approach into a combination of indicators.

    If your strategy uses orderbook (market depth), no "retail" platform supporst backtesting (as far as I know) and the data will not come cheap.

    If I understand correctly, your system gives an entry signal "watch this stock, it may move". Good entry signal is not sufficient for a good trade. One needs right entry level and trade menagement (even if it's just a "close" level). So, disclosing the entry signal may not be much of a loss.
     
    #19     Jan 8, 2010
  10. Excellent advice and hope others remember it. :cool:

    Yet, what happens if the OP is interested in full automation that involves all of his rules instead of just the entry signal. :confused:

    Mark
     
    #20     Jan 8, 2010