Re: [Quantlib-dev] Confronting the upfront

[Jose Aparicio-N. <ja...@fr...>]

Quoting Ferdinando Ametrano <qf...@am...>:

> I wonder how to bootstrap including an upfront if depending on the
> TermStructure reference date the upfront might not enter the NPV?
>
> ;-)
>
> ok, ok, I'm joking... sorry... I just couldn't resist... :-D
>
> ciao -- Nando
>

Its not sci fiction, its in there:
and it boostraps from it

    void MidPointCdsEngine::calculate() const {
[.....]
        // Upfront Flow NPV. Either we are on-the-run (no flow)
        // or we are forward start
        Real upfPVO1 = 0.0;
        if(!arguments_.upfrontPayment.hasOccurred(settlementDate, true))//true?
            upfPVO1 =
                probability_->survivalProbability(settlementDate) *
                discountCurve_->discount(settlementDate);
        results_.upfrontNPV = upfPVO1 * arguments_.upfrontPayment.amount();
[.....]

"hasOcurred" hits back...
Settlement lags are relevant here indeed (ordering combinations of
T_settle=upfrtPay  Today and T_protection_start), I did not enter the discussion
but followed with interest because of this but yes the meaning of today is
relevant.

Regards
Pepe