Subject: sharad ' s houston visit
sharad
as expected we didn ' t get time to discuss in detail what to do in houston .
we have discussed the subject informally , but i ' m laying it out here for
clarity , and for vince / stinson ' s information . here are my thoughts :
1 . exotica . this is the top priority , as you ' d rightly identified yourself .
a . learn how to build xlls .
b . catalogue functions in houston exotica not yet available in london .
c . catalogue differences between inputs to london and houston exotica
functions .
d . produce london xll that allows current function calls to be used where
possible , and incorporates new functions .
e . update documentation as appropriate .
if you do nothing else in your two weeks , i ' ll be very happy with this .
main contacts are paulo and zimin .
2 . understand american monte carlo ( amc ) . the aim is to get a feel for amc ' s
characteristics . before we embark on the use of amc for real option
valuation , i ' d like to understand how it behaves for financial options . i
suggest something like the following :
a . produce amc code for single gbm for american option . compare greeks
against those from appropriate tree methods .
b . two gbm spread option model , to compare against 1 - d numerical integration
method . although the " american " bit of amc isn ' t going to be used here , it
will be interesting to think about the bucketing issue in price / payoff space
and , again , the greeks .
c . some form of mean reverting model , preferably two factor .
d . hjm for pure financial option valuation .
if you get somewhere on a - d , and amc behaves sensibly on pure financial
models , then i ' ll be very keen to roll it out for real option valuations .
vince / stinson - since i ' m expecting the worst from it i . t . o . email support ,
would you be good enough to print this email out and hand it to sharad ?
many thanks ,
steve