Formula for "Implied Volatility"

Discussion in 'Options' started by Derrenoption, Sep 26, 2016.

  1. <-- I referenced the ATM (not ITM) and OTM options, which is per the VIX White paper.
    My comments above. If you like, you can contact me privately, as we seem to be miss-communicating.
     
    #21     Sep 27, 2016
  2. MACD

    MACD

    WoW -- who knew that there were so many erudite, experienced, and are willing to share that knowledge. I rarely bother with reading ET forums and now I believe it does have merit. Thanks to these posters.

    Hoping, that my addition to this thread will Not be Considered "off topic"; How do you use the information that you gain from knowing (particularly with the specificity that some want) what the IV is? Is it for determining what strategy you will employ to initiate a trade? Do you use it to determine comparative vol to decide whether you wish to be a net seller of vol? So now that you have computed IV by using your method -- What are you doing with that info? Thanks for allowing this digression.
     
    #22     Sep 27, 2016
  3. IMHO: "Implied Volatility" may be the most misunderstood and misused topic in option trading/evaluation. I am NO expert, but am trying to lessen my ignorance.
    It is customary to refer to the IV of specific options, of specific expirations, and of collections of option expirations. Each may be used for varying purposes. The first "typically" relates to individual implied volatility, and the others are forms from the CBOE VIX white paper. (the specific expirations, don't require the extra time weighting, so are slightly easier to compute)
    TastyTrade recommends strategies which sell premium to take new positions when IV is above some IV rank or percentile, and either underweight or not trade if below some threshold. The idea has merit since IV tends to be mean-reverting. (good for strategies which improve with a decrease in IV while in the trade). When they refer to IV in this context it is typically the VIX or the 30DTE IV of the product. {Since they concentrate on 45DTE time frames, 30DTE is a good enough reference.}

    Most people will not need to compute IV, as they may merely reference it on their trading platform. It seems the OP may not have access to the data and wishes to compute it himself -- I am still guessing here. Bob's recommendation to consider LiveVol may be the easiest and best long-term solution, as the data may already be in a usable form.

    "How do you use the information that you gain from..." <-- It is a function of what you are looking for and are wishing to accomplish! The issue in this thread, I think, is some confusion with terminology and definitions, wherein the OP mixes the individual implied volatility with the pseudo-standardized IV metric. There are a number of derivations of the former, and a single well documented derivation of the latter (known to me). These are not to be intermixed or confused with each other (even though there is some unfortunate overlap). Derivations of the individual implied volatility are very important for those trading by the Greeks, since the Greeks are a function of the IV (individual implied volatility of all strikes making up the position). For some, smoothing algorithms are applied to the individual implied volatility, and for others, not so much. TOS can be configured to produce the individual implied volatility from 3 choices: individual implied volatility, Volatility Smile approximation, or fixed per expiration date(similar to the VIX white paper method): The proper choice depends on the desires/requirements of the trader.

    Regarding "that you have computed IV by using your method --" : this may be missing the point. I do not claim to have a magic method. My code merely produced the values I observed published, so I used in in some of my back-testing. IE: At the time, I was using the Bid/Asked prices of the options and the B&S model was matching the TOS individual implied volatility, I did that since I was unable to always obtain the TOS individual IV values. No magic, just apparently using algorithm that did not disagree with that of TOS at the time.

    Pardon the book, but I'm still learning.
     
    #23     Sep 27, 2016
    NaturallyNick likes this.
  4. >> For example if it is 15 days left in current months and 45 days left in next month contract: (15 + 45) / 2 = 30 days. Then take the average implied volatility for those to get a better picture for each days that pass?
    I was thinking again of a 30 DTE you mentioned. Would the above be a good idéa to do?


    >> Is it for determining what strategy you will employ to initiate a trade?
    Yes, the idéa is to have the implied volatility as one factor to know when to buy or sell.

    >> I referenced the ATM (not ITM) and OTM options, which is per the VIX White paper.
    Yes, I noticed that you mentioned the ATM. But I thought since ATM are very rare as the strikeprice has to be EXACTLY the same as the underlying, I thought to do some kind of median value of nearest ITM and OUT strike(s) to be as close as possible to ATM.

    >> It is customary to refer to the IV of specific options, of specific expirations, and of collections of option expiration
    Yes, that I beleive I have understood then. So that should be take into considiration for different purposes.

    >> It seems the OP may not have access to the data and wishes to compute it himself -- I am still guessing here.
    Yes, I really like to compute those values from the raw data myself as I integrate raw data into my own trading application so it is not nessecary to use data from one source. For example if backtesting from one source of data, it will be possible to trade a strategy on another source of live feed(if neccesary), knowing that same values are calculated live vs backtest.
     
    #24     Sep 27, 2016
  5. Since you appear to be going down a path of coding your own Back testing tool, I think it will be in your best longer term interest to study the CBOE VIX White paper and understand it. Then, you may chose NOT to implement its algorithms, but as long as you realize the error you are introducing, and are comfortable with that error, it is totally your choice. Some lights should begin to turn on for you as you digest this, which may be more significant than coding a specific function.
    Using VIX, or another 30DTE IV is a good metric for observing/quantifying IV Rank (NOT an individual option's IV). Using anything else, is NOT as stable.

    Me thinks your solution is for the wrong problem! ;-)
     
    #25     Sep 27, 2016
  6. sle

    sle

    The skew and term structure will really skew the results for some things if you always use a strikeless and fixed term IV measure.
     
    #26     Sep 27, 2016
  7. sle:
    I "THINK" the OP's quest in this thread is to quantify the IV with respect to the IV rank/percentile to use to make decision of placing or sizing a new trade entry. SKEW is likely beyond the scope of what he is currently seeking, and term structure is a slightly different animal. Both very important, depending on what one is trying to accomplish. If it not for the context of this thread, I totally agree with your statement!
     
    #27     Sep 28, 2016
  8. >> Some lights should begin to turn on for you as you digest this, which may be more significant than coding a specific function.
    Using VIX, or another 30DTE IV is a good metric for observing/quantifying IV Rank (NOT an individual option's IV). Using anything else, is NOT as stable.

    Your information is very valuable and I will really think about how I go about this. I will try to take a better look at the "CBOE VIX White paper" to really se what they are doing there. As you say I will probably stumble upon some problems/errors along the way. As I have tested the IV now for options at different strike prices, the IV are quite close to TOS and the delta are almost the same +-1% but I have not done many tests on that so I will need a lot of more tests later when I get more things in place. It seems as a good startingpoint and I am prepared for errors, they will most probably occur.

    >>I "THINK" the OP's quest in this thread is to quantify the IV with respect to the IV rank/percentile to use to make decision of placing or sizing a new trade entry.

    Yes, you are right ;) It will really be interesting to create this backtesting program, I beleive it will take a while with all the idéas I have in mind. But to see how IV affects different approches as one of the ingredients.
     
    #28     Sep 28, 2016
  9. I am very interested in what you are doing. I have done (am doing) similar, and find it difficult ( a huge amount of work) but very educational. The benefit of the things we learn by these endeavors, will likely outweigh the original goal of the task.
     
    #29     Sep 29, 2016
  10. ironchef

    ironchef

    Me too. When I back tested options, I could get historical dividend rates, risk free rates but historical IV in one value only. So, I ended up used a range of IV around the one value and some assumptions on skew, etc. for back test results to develop my strategies. It was very cumbersome and tedious and not always correct. However, it was good enough for me to see why some popular strategies, e.g. mechanically selling calls or puts (or straddle, which is a combination of selling call + put) were not as profitable as claimed once I extended the tests to multiple years.

    Regards,
     
    #30     Sep 29, 2016