direct statistical trading a "clearly" defined approach by NTW31

Discussion in 'Strategy Building' started by nukethewhales31, Dec 31, 2008.

  1. The NPR% from the tick data found that from open if price moved 2.1 pips from open price would continue to at least 5.4 pips from open with 90% accuracy on the EUR/USD
     
    #131     Jan 12, 2009
  2. any questions?
     
    #132     Jan 12, 2009
  3. This calculation with tick data is equivalent to the manual process you were going through with bar data right?

    How many hours of tick data was analyzed?
     
    #133     Jan 12, 2009
  4. thats correct same idea but this time we dont need to swap lsd's or sld's as tick data is very clear of its movements and only goes in one direction

    the idea of going through the other stuff was to give you the concept show you that even the beginning concept works..

    show you how the calcs are done because this doesnt only work for distances from open

    you can find

    -statistical distances when price breaks past high it will continue to moving in that direction or pull back

    -prices statistical action when it creates a low repasses the open how far it goes statistically.

    -etc.

    so you can apply this to anything and be successful

    to can find statistical reaction levels with what you trade lets say you trade a macd cross system you can run this to find better entrys/better stops/ better tp's
     
    #134     Jan 12, 2009
  5. from our data this graph
     
    #135     Jan 12, 2009
  6. check out this

    TRO's manual on buyzone is sorta what we've been testing statistically

    his is fixed tho note that the market is always changing its not as rigid rules with hard numbers.

    im not pushing this by all means so dont think that i am im simply making an observation that his is fixed remember to keep your variable test multiple samples.

    remember to run the stats ... as this is a sever process of curve fitting data remember to out-of-sample data test your stats once you are finished creating them.

    every vehicle is different across the board so stick to one vehicle and trade it well.
     
    #136     Jan 12, 2009
  7. I will now only be checking this thread in order to answer questions or to assist people who need a little help running this.
     
    #137     Jan 12, 2009
  8. Do you still need the bias from the 20 period MA? It seems like you'd be able to derive a bias from the tick data by seeing which direction the long sides have been poking out. i.e. have they mostly been poking up or down?
     
    #138     Jan 12, 2009
  9. if you record annotate bias in the excel you can see compared to bias how many times it moves with bias or against it or how far it moves with bias and how far against it.

    yup correct tick data can give you bias information as well.
     
    #139     Jan 12, 2009
  10. Interesting. Well, I don't have good forex tick data, but I *do* have good index futures data.

    This should be fun since I'm a programmer and it's right up my alley.
     
    #140     Jan 12, 2009