Probability of Touch calculation

Discussion in 'OptionVue Forum' started by Gasman, Aug 6, 2017.

  1. Gasman

    Gasman Member

    Can anyone explain why the POT in OV gives very different values to options than what TOS generates. Is this not a fairly consistent calculation? Thanks in advance.
     
  2. PK

    PK Well-Known Member

    Thanks for bringing up this interesting point! I noticed that the POT in OV even changes, sometimes by up to 5%, when you repeat the calculations several times without any change in market data, which indicates that OV calculates this number not with a simple formula but rather with some kind of simulation, maybe Monte Carlo, with a limited number of events so that the obtained results are still quite far from converging on a single value. If anyone knows more about the mathematics behind OVs POT calculation, please let us know. Thanks in advance.
     
  3. Anil

    Anil Member

    As a rough guide, I believe probability of touch is roughly = 2 * Delta. This is a takeaway that I have from watching Tastytrade. However, I'm not sure how it compares to OV or TOS, but believe it may be similar to what's in TOS as the quants at Tastytrade were the ones who built the models in TOS.
     
  4. Len Yates

    Len Yates Active Member

    Yes, OV uses a Monte Carlo simulation of the various possible moves the underlying might make day by day from now until expiration. I seem to remember McMillan publishing an article about this years ago and I believe we picked up his Monte Carlo approach at that time.
     
  5. stevegee58

    stevegee58 Well-Known Member

    I have a probability of touch algo in C++ that's based on Black-Scholes. It took some Google-foo to find but seems to work correctly.
     
  6. PK

    PK Well-Known Member

    Len, thanks for the clarification. I find the probability of touch calculation very useful when it comes to estimate risk of open posiitions and, as suggestedby Tom Nunamaker, as an "indicator" to take into account for the adjustment of Road Trip Trade positions. Though a rough approximation may be taking twice the %ITM probability or the delta value, I appreciate the added value of providing a Monte Carlo simulation. Do you refer to the attached publication?
     

    Attached Files:

  7. Gasman

    Gasman Member

    Called OV and spoke with James. Here is his written response. Inline with above comments, but definitely different than TOS 2x Delta.

    "When you click the PT button the Matrix performs a 500 iteration Monte Carlo computation for every option in the Matrix that is out of the money. The resulting number for each option represents the probability of the underlying ever moving to touch the strike price during the remaining life of the option. An hourglass cursor is displayed during the computations, which may take anywhere from a few seconds to a minute or more, depending on the speed of your computer and how many options are defined in the Matrix. 500 iterations is enough to get reasonably good numbers but there will still be some degree of error involved and from one round of computations to the next you can see the numbers change a bit just due to this. Once you click the button and see the numbers filled in, the program will remember the computed values for the remainder of the day (or until you click the PT button again, which you had better do if the underlying price has moved appreciably in the meantime). The next day the PT numbers are automatically blanked out."
     
    PK likes this.
  8. Jim N

    Jim N New Member

    Isn't the POT a specious stat? As someone said, (I think it was Tom) "It is just another instrument on the cockpit dashboard". However if the instrument is indicating "dive"/"ascend"/"level off"/"ascend"/"dive" in a matter of 'seconds' as PK has pointed out, then there is a bumpy ride ahead.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice