Re: Derivates docs

[Neil Booth <neil daikokuya demon co uk>]

Anthony wrote:-

> 1) Wouldn't it be simpler if time-to-expiry is given in years,
> to be consistent with the annualization of the vol and the risk-free
> rate - saves the usual 365/365.25 tedium.  Also the docs don't appear
> to state the compounding assumption of the risk-free rate (I assume
> it's annualized as well).

We discussed this, and it might be done for a future release.  The risk
free rate is continuous.

> 2) These ten funcs are often collapsed into one by many financial
> spreadsheet libraries - they take a parameter "C" or "P" for the
> option type and a string argument to specify the requested output.
> (If the output argument is an array one can return various values
> for the price of a single calculation - not too important for BS
> since it's quick but it's a helpful style for more heavyweight stuff)
> Also when one is sceptical about say the gamma you can change an
> input cell to see the delta, not edit the function call also risking
> that the signature is different.

Exactly; I suggested all this too.  I also suggested something that
no library I've ever seen provides - S for straddle.  So simple,
yet so annoying, if you're a trader, that it isn't provided.  Also,
functions to solve for forward price and vol.  I think this improved
function won't take a risk-free rate at all; it should take a
forward rate and return an undiscounted premium, IMO.

I hope to add these.

> 3) Spreadsheet gurus (wrong word, I really mean "nargs" but I suspect
> it's a britishism) will often implement their own functions which
> improve the builtin stuff.  The COUPNCD stuff I was just reading is
> a prime example, and without disrespect to the B-S code referenced
> here I might be doing the same.  I imagine this is quite common, so
> that statistics nargs have their own regression functions
> with extra bells and whisles and diagnostics and so on.  One issue
> comes to mind: if I implement a function with the same name and
> signature as a function in derivatives.so, can I have both
> libraries loaded? I presume not?  If not, should I be able to,
> calling derivatives::blaa or mylib::blaa explicitly? I'm
> imagining the plugin manager could have a toggle to 'export' the
> names into the unqualified namespace...

Dunno.  I reckon we should prefix all the non-Excel built-ins.  e.g.
calXXXX for calendar functions, optXXX for option-related stuff etc.
I think its cleaner to use like that, too.

Neil.