Quantcast
  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

OptionVue Settings

Discussion in 'OptionVue Forum' started by AKJ, Oct 1, 2016.

  1. AKJ

    AKJ Well-Known Member

    For people who trade broken-wing butterflies, what are the general opinions on using EIOIO versus Variable Vol models for the greeks and T+n line P&L profiles? I have been testing some things over the past few weeks, and in the current environment, it appears the EIOIO model is a far better and reliable predictor of P&L changes (both in terms of changes in underlying and passage of time) than Variable Vol.

    For some of the veteran users, has anyone developed an intuition about which type of volatility environments are conducive for different OV volatility settings? I know John Locke's preferred model is Variable Vol, and he states his belief that this model is the most reliable across ALL market environments. I'm wondering if I can do better, and pick and choose the most predictive model for different environments, and then switch between models as appropriate as vol regimes change.
     
  2. DavidF

    DavidF Well-Known Member

    Hi Andrew, not a veteran trader by any means but hope it´s ok to chime in. I looked at this a lot during the period when OV was struggling with an updated model and found EIOIO to be much more accurate. My method of checking was to retrospectively see where EIOIO vs. variable would have predicted the P/L to be versus where it actually landed. I believe EIOIO eliminated the problems with the call skew model which were obivously way off in the variable model.

    Now however, if I use the same method to check the current environment I find the variable to be more accurate (but still way off as I´ll post elsewhere in down moves). For example if I place an Oct 2175/2125/2050 BWB on SEP 30, EIOIO predicts it would have been worth about $9000 at yesterdays cllose, varable about $6000. The correct value was $6300. OF course one woiuld need to run a larger sample over mutliple dates so just a quick check. I understand you´re a meticulous person so I assume you´ve already done this and found EIOIO more accurate?
     
  3. Philip

    Philip Member

    How did you conduct your test? Just a feeling or can you back it up with empirical data?

    I am asking as I am very interested in understanding how we can get a more accurate model. In my own testing, I have found The Variable model to be better.



    Sent from my iPhone using Tapatalk
     
  4. AKJ

    AKJ Well-Known Member

    one possible test that I have done at times - for each t in T, compare the analyze graph's prediction of the t+1 P&L versus the actual P&L across multiple models. The model that results in the lowest standard error (or some other measurement of dispersion of your choice) would be the more accurate model imho.
     
  5. Steve S

    Steve S Well-Known Member

    Have you guys ever checked out the CEV model? It's bananas, so far as I can tell - much too flat. To test, buy 1 put 50 or 100 points below the money and check out how the volatility of that put on the T+0 line changes on a move to that strike. It will be way too low.
     
  6. Rtb

    Rtb Well-Known Member

    Hi Andrew, After reading your reply to the thread I started the otherday on Optionvue, I decided to phone Optionvue tech support on Friday to discuss how to set up Optionvue because I want to do a lot of backtesting. They advised me not to use EIOIO and to use variable and Yates, true delta gamma and true theta and dont select the combined call/put skews. I was told that if you select EIOIO you are making Optionvue alot more basic. Just want to point out I am new to Options Trading and using Optionvue. It might be good to get some feedback from some veteran OV Users.
     
  7. AKJ

    AKJ Well-Known Member

    Hi RTB,

    Yes, I understand the different settings that OV themselves advocate for in using their software. What I am suggesting is that some empirical research should be warranted to validate the accuracy of their suggested model versus the more straight-forward EIOIO model. My recent observations are only anecdotal, but suggest that the CEV model is not as accurate as EIOIO in the current volatility regime. This does not necessarily mean that it isn't the more reliable model across ALL vol regimes or even necessarily the current regime, just that I've experienced some fairly inaccurate P&L projections from OV's Variable Vol model over the past few weeks.
     
  8. Steve S

    Steve S Well-Known Member

    Just to clarify, the CEV part of OVue's model is (I believe) just the model for atm vol dynamics ... that's just one chunk of OVue's entire skew dynamics model.
     

Share This Page