Serving the Quantitative Finance Community

 
User avatar
emac
Posts: 1
Joined: July 7th, 2009, 7:15 pm

Finite difference and control variates

July 2nd, 2015, 12:01 pm

What is the spurious input?The standard error is usually quoted as a root mean square error, i.e.[$]\sqrt{E |X_M - \mu|^2} \leq \frac{std(X)}{\sqrt{M}}[$]But this is an average error. For a given realisation of [$]X_M[$] you don't know a priori how close to [$]\mu[$] it actually is, you only know that ON AVERAGE it is close. I was just telling you how to get a bound on the probability that [$]X_M[$] is actually close to [$]\mu[$].
 
User avatar
Cuchulainn
Posts: 20252
Joined: July 16th, 2004, 7:38 am
Location: 20, 000

Finite difference and control variates

July 3rd, 2015, 8:34 am

Details aside, this is a discussion between a pure mathematician and a numerical analyst. In the sense they have different views of formulae.
Last edited by Cuchulainn on July 2nd, 2015, 10:00 pm, edited 1 time in total.
 
User avatar
mutley
Topic Author
Posts: 20
Joined: February 9th, 2005, 3:51 pm

Finite difference and control variates

July 3rd, 2015, 11:53 am

emac, by spurious inputs I meant things like implied vol and some of the more heroic models. Wasn't referring to your work aboveAlas, I don't think I could claim to be either of those two professions. Am just a poor structurer with a healthy dose of cynicism for absolutes.