You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
"A massive aspect of calculating IVs, particularly in this interest rate environment, and if you’re considering American options that pay dividends or whose underlying security may be hard to borrow, is accurately calculating/estimating your forward price.
This isn’t trivial and from I can gather in your repo you aren’t implementing any thing to handle dividends (implied, discrete or continuous) or cost of borrowing. Correct me if I’m wrong but I’m also not seeing you de-Americanize the options anywhere, so you’re treating everything as European, which of course leads to another drawback which is that you’re using Black Scholes instead of a proper American pricer.
Further, I see you’re fitting the resulting Black Scholes vols using a spline fitter. How good are your fits across a wide set of securities’ surfaces? Are your surfaces free of vertical and horizontal arbitrage? There are models and methods account for that. This being one of the last steps in the journey of course, which starts with the correct forward.
In all, though, I do like the implementation and the thoughtfulness you’ve given certain things. These are just a few aspects that would improve your models.
EDIT: forgot to add one last but very important thing: option prices themselves. The choice between bid, ask, last, mid, or a model-free approximation is also critical."
The text was updated successfully, but these errors were encountered:
"A massive aspect of calculating IVs, particularly in this interest rate environment, and if you’re considering American options that pay dividends or whose underlying security may be hard to borrow, is accurately calculating/estimating your forward price.
This isn’t trivial and from I can gather in your repo you aren’t implementing any thing to handle dividends (implied, discrete or continuous) or cost of borrowing. Correct me if I’m wrong but I’m also not seeing you de-Americanize the options anywhere, so you’re treating everything as European, which of course leads to another drawback which is that you’re using Black Scholes instead of a proper American pricer.
Further, I see you’re fitting the resulting Black Scholes vols using a spline fitter. How good are your fits across a wide set of securities’ surfaces? Are your surfaces free of vertical and horizontal arbitrage? There are models and methods account for that. This being one of the last steps in the journey of course, which starts with the correct forward.
In all, though, I do like the implementation and the thoughtfulness you’ve given certain things. These are just a few aspects that would improve your models.
EDIT: forgot to add one last but very important thing: option prices themselves. The choice between bid, ask, last, mid, or a model-free approximation is also critical."
The text was updated successfully, but these errors were encountered: