I'm currently doing some refactoring of the finite difference engines in
order to make them more generalizable to items such as equity options.
I'm pretty radically changing the internal structure of the engines.
This should be invisible to people who are using the finite difference
engines since I've put in compatiblity routines so that no external code
will break.
However, if anyone is subclassing the FD engines, let me know. To clean
things up, I haven't put in internal backward compatibility stubs, but
will do so if there is a need for it. The internal changes are fairly
straightforward (i.e. changing getProcess() to process() to make it fit
naming conventions)
Also, I was at a seminar where they presented this really cool paper by
Vadim Linetsky on unifying default calculations for equities and fixed
income.
http://www.ma.utexas.edu/Seminars/MathFin/One of the interesting things that the speaker mentioned is that the
largest open option position on General Motors is very out of the money
puts which makes no sense in conventional option calculation engines,
but it is a hedge against GM going bankrupt.
The interesting thing is that the PDE he showed was pretty much exactly
the same as the Alyache-Forstyh model for convertible bonds.
Once I've done cleaning up the FD engines, I'm doing to try to create a
process DefaultableBlackScholes which should go right into the FD engines.
-------
Joseph Wang
Interested in jobs involving Chinese quantitative finance
Locked
