|
PureBytes Links
Trading Reference Links
|
Hi Lee,
as frequently seen in discrete maths, it is easy to find a recursive formula
for such a problem; the task of finding a non-recursive formula is slightly
more difficult. You might recall the Fibonacci sequence which is very easy
too express recursively; it requires some non-trivial effort, however, to
find the direct formula.
Sorry, but I am afraid the formula I gave is correct. You are quite welcome
to post an instance which you believe to be incorrect, indicating the
intermediate results of your calculation, so that we can compare notes.
As for the derivation of that formula, I'll come up with an explanation of
the method used when I have more time later.
Best regards,
Michael Suesserott
> -----Ursprungliche Nachricht-----
> Von: LScharpen@xxxxxxx [mailto:LScharpen@xxxxxxx]
> Gesendet: Friday, October 26, 2001 14:15
> An: MikeSuesserott@xxxxxxxxxxx
> Cc: omega-list@xxxxxxxxxx
> Betreff: Re: Hard sums, easy formula
>
>
> In a message dated 10/26/01 4:31:12 AM Pacific Daylight Time,
> MikeSuesserott@xxxxxxxxxxx writes:
>
> << > > Hi Ian,
> > >
> > > try this:
> > >
> > > Floor(0.1*(Sqrt(5)*Sqrt(8*Acct-75)+5))
> > >
> > > Floor might be called Int in EL, I don't remember now.
> > >
> > > Best regards,
> > >
> > > Michael Suesserott >>
>
>
> Hi Mike,
>
> I'm not quite sure how I got hooked into this thread but I did.
> Perhaps it
> was the very patient way you were dealing with the guy who asked the
> question! <g> Anyway, I did get hooked. I wound up with the value for A
> (account size) being a recursive function of C:
>
> A (for C contracts) = A (for C-1 contracts) + (C-1)*5
>
> I then tried using your 'solution' [
> Floor(0.1*(Sqrt(5)*Sqrt(8*Acct-75)+5))
> ] and found that for a given value of A it was NOT calculating the
> proper # of contracts. It yielded values for the # of contracts
> which was
> consistently too large by 4 contracts. Changing the last 5 in
> your formula
> to a 1 produced the 'correct' results. Am I all wet here or is
> my analysis
> correct?
>
> My efforts to find a way to calculate the # of contracts to be used for a
> given value of A (account size) couldn't get any further than
> finding the
> recursive relationship I noted above. So I AM curious how you
> derived the
> formula you came up with. Is there something I should know about the
> properties of recursive formulae that has escaped me?
>
> Regards,
>
> Lee Scharpen
|