[R-SIG-Finance] R and FinCad

[Ryan Sheftel]

First thanks for everyone's responses. I will try to hit all of the
questions in one email:

1) FinCAD is a commercially available library of financial functions that is
available in Excel form
(http://www.fincad.com/default.asp?id=2000&s=Products&n=FINCAD%20XL) and a
developer library
(http://www.fincad.com/default.asp?id=2100&s=Products&n=Fincad%20Developer).

2) Online they have a list of functions that they offer and the User Guide:
http://www.fincad.com/default.asp?id=17300&s=Support&n=References

3) They have competitors in the space, (http://www.techhackers.com/) FinCAD
is just the one I have used in the past and have found to be the best for my
uses.

4) The cost is about $3,500 for the Excel license, and the developer code is
about $15k for the base and then I think you pay per user. Certainly not
cheap, but in the world of trading  technology the numbers are small.

5) What does FinCAD have that is good, in my mind the following:
	a) Extensive product coverage (interest rates, options, and credit)
	b) their functions conform to market standards
	c) as a firm I found them very responsive
	d) documentation is clearly written and usable to non-technologists

6) Price matching. Generally speaking any pricing formula that is not closed
form will mean that any two sources (FinCAD v. dealers) will have different
prices, because the price is model based and require assumptions. For
example Bermudian options, no 2 dealers will agree on the price. Also having
worked at a dealer, I can say that there is no reason to assume that dealer
pricing is "correct". I have seen plenty of bad code, and in some ways when
you only have one captive customer (your in-house desk), the quality control
is going to be less than a FinCAD or QuantLib with many eyes. Either way, I
think FinCAD has enough users that any bad pricing is corrected.

OK, so where does that take us. Several people have expressed similar
interest in R<>FinCAD connection, so perhaps Eric and myself can co-ordinate
something where we share the effort.

All that said, I definitely think something like QuantLib is a better long
term solution. The blockage for me is:

	a) There is no modeling for credit, which is 1/2 of what we do
	b) The documentation is written for programmers from what I can see,
so deploying it to end users means writing user docs. (if I am wrong please
direct me, I tried to search for it). Difficult to tell from the docs what
the financial math behind the functions is.

We would like to assist in making QuantLib the best solution so that it has
the models, and the R connectivity. Can someone recommend the lists or sites
we should connect with to get involved? Thanks.


-----Original Message-----
From: r-sig-finance-bounces at stat.math.ethz.ch
[mailto:r-sig-finance-bounces at stat.math.ethz.ch] On Behalf Of Dirk
Eddelbuettel
Sent: Thursday, April 12, 2007 10:17 AM
To: Robert Sams
Cc: r-sig-finance at stat.math.ethz.ch; Yong Xiao
Subject: Re: [R-SIG-Finance] R and FinCad

On Thu, Apr 12, 2007 at 01:08:12PM +0100, Robert Sams wrote:
> I've used FinCadXL for over a year. So far, all calculations 
> approximate reasonably well to bloomberg analytics and dealer 
> valuations (although I use it primarily for pricing vanilla swaps,
swaptions, and bond math).
> It is the most comprehensive function suite for interest rate 
> derivatives and benefits from being (I think) the most widely used 
> library in the industry. Documentation is reasonably good.
> 
> But it is a binary-only library. The world would be a better place if 
> there was critical mass behind something like QuantLib.

There really should be!!  The Quantlibbers just sent out something like a
call for 'wider participation' on the -devel list.  

My RQuantLib project was always meant as 'proof of concept' only -- if folks
want to contribute I am more than happy to include their code, or to replace
mine.  Dominick's rather nice additions are a good example.

Also, Joe Wang did some fabulous work with the _automated_ SWIG glue that
provides essentially _all_ QL functions directly to R. Still not that easy
to use and somewhat heavy on memory but way more than a start.

So if a few people would make a serious push, we could have some really
sharp R/QL integration in a matter of weeks.  Some C++ needed, of course,
but doable on both major OS flavours.

Dirk

--
Hell, there are no rules here - we're trying to accomplish something. 
                                                  -- Thomas A. Edison

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. 
-- If you want to post, subscribe first.