Anyone still using Stratasearch for backtesting/strategy development?

Discussion in 'Strategy Development' started by SteveM, Jul 9, 2019.

  1. SteveM

    SteveM

    About six months ago, Stratasearch's developers decided to abandon the platform and are no longer providing updates/further improvements.

    The platform still works perfectly fine, but I'm a bit concerned that over time there will obviously be fewer and fewer people using it for their strategy development/backtesting needs, and therefore will be harder to find other users to bounce ideas/issues off of.

    Just curious if there is anyone here on ET that is currently using Statasearch? I'd like to connect with others to share ideas, and think it would be good to have some other brains around in case software bugs come up, or any other issues with the platform arise.

    If anyone is using Stratasearch, and would like to connect with other users to collaborate and possibly develop strategies together, feel free to respond to this thread, or just shoot me a message.

    Edit: Also, if anyone used Stratasearch in the past but moved onto a platform with similar capabilities that you would recommend, feel free to mention those programs as well. Thanks.
     
    Last edited: Jul 9, 2019
  2. RedDuke

    RedDuke

    Take a look at NinjaTrader. Fantastic in all respects for backtesting, optimization and live trading. Great support. Also free until you need to trade live.
     
    SimpleMeLike and SteveM like this.
  3. MattZ

    MattZ Sponsor

    I would recommend using Multicharts if you are looking to automate your trade entries. It was one of the most stable platforms that we ran automation on. They have a lifetime license, and periodically run specials. so we can always ask for a discount if necessary.
     
  4. MotiveWave

    MotiveWave Sponsor

    Take a look at MotiveWave. Full backtesting, optimization, walk forward testing, replay mode. Supports 30+ brokers and data service providers. Leases and Lifetime licenses. Very stable. We also have a Java SDK for custom study and strategy development. We're running a Summer Sale right now with 20% off.
     
  5. TomHam

    TomHam

    I have not used StrataSeach for many years. After spending many hours learning it, and testing many different ideas I came to a simple conclusion. The system developed performed reasonably good for about a week into the future, then quickly degraded to the point of losing money. IMHO, even with all the IS & IS and walk forward techniques, it still basically over optimized a system.
    I could never get it to provide profitable models any more than a week or so into the future. Too bad because it is an interesting program. Maybe you just have to start a new model run every weekend. But that just didn't sit well with me.
    Other users comments?
    Cheers and Good Trading.
     
  6. RedDuke

    RedDuke

    What do you mean about new model every
    Weekend? What was the dataset for backtest?
     
  7. TomHam

    TomHam

    Hey RedDuke -
    "New Model Every Weekend" what I mean is that the results of each "search" seemed to be profitable for only a short period of time (i.e the next 5 days or so). After that perform dropped off, thus the work around would be to renew the search frequently (very frequently). IMHO, this is a sure sign of over optimization. Your results make vary. :)
    "Data Set": liquid ETF's, typically the major sectors and limited to 10-15 (from what I can remember. Nothing special.
    You get some great results, they just don't last very long in real time. Hopefully others have found out something different. I could not.
     
  8. RedDuke

    RedDuke

    How far back does your data go to build a model?
     
  9. TomHam

    TomHam

    Ah Ha, you hit upon an important point in back testing. Does one select back data representative of the current market regime or over a broader period encompassing both up and down markets. And then, which index does one choose for a benchmark to evaluate the period?

    My philosophy is to use the S&P 500 as a benchmark, and evaluate over multiple regimes so that a model would performance well (i.e. get in / out) in most environments. Now, if you have the philosophy to "tune" a model to perform well in "up markets" and another in "down markets", then you can get outstanding (short term) results. But then again I return to the idea of over optimizing.
    It's good to have everyone have a different philosophy so we all don't do the same thing. That would be disastrous. :) Cheers
     
  10. I used it but systems were sort of random combinations of indicators fitted on data and most would fail immediately unless you were lucky to pick the right one . I then moved to PAL now called DLPAL. Fitting was still potential problem but due to working only with price patterns and some specific methodology it was not as serious and some systems worked well. I still use it with Ninjatrader 8 since it generates code for the platform. I have tried some other programs but mostly operated like Strata.
     
    #10     Nov 4, 2019