If you had a time machine, could you have bought the option at the price your model outputted? If so, have you tested this? Or do you get historric implied vols and plug them into your model?
Some obvious facts...
- Thread starter mutluit
- Start date
If you had a time machine, could you have bought the option at the price your model outputted? If so, have you tested this? Or do you get historric implied vols and plug them into your model?
He didn't actually say that he derived BS without ito's lemma, which is obviously possible (though obviously, doubtful that he can).
He said he 'eliminated' it from the formula (or, 'automaton' as he called it in another thread). The fact that he tried to post some code showed that he meant the implementation, not the derivation of BS.
With that, I asked him to point out where in the formula is ito's lemma (or ito-lemma, as he called it), at which points he started with the whole "it's my IP", and "it's irrelevant" tangent...
..because, as is so common on ET, he's a crank.
Come on mutluit, is it really that hard to put up your claim? Show us exactly where in that two line BS formula is your so-called 'ito-lemma'?
He said he 'eliminated' it from the formula (or, 'automaton' as he called it in another thread). The fact that he tried to post some code showed that he meant the implementation, not the derivation of BS.
With that, I asked him to point out where in the formula is ito's lemma (or ito-lemma, as he called it), at which points he started with the whole "it's my IP", and "it's irrelevant" tangent...
..because, as is so common on ET, he's a crank.
Come on mutluit, is it really that hard to put up your claim? Show us exactly where in that two line BS formula is your so-called 'ito-lemma'?
Quote from sle:
Black-Scholes pricing formula for vanilla option can be, indeed, derived without the use of stochastic calculus and Ito's lemma - that derivation uses pure probabilistic and risk-neutral approach. However, if you can show that you can derive Black-Scholes SDE without the use of Ito's lemma, that would be a feat.
There are numerous fields, finance one of them, that are full of "cranks" - people that come from elsewhere and think that they can figure it all out in 20 minutes and make a killing. Yours truly arrived at a hedge fund after completing a PhD in biophysics and thought that in a month I will have it all figured out. A decade and a half have passed, I have changed a few jobs, made hundreds of millions for the various firms, survived 9/11 and muddled through the credit crisis and yet I do not have it all figured out. However, if you think you have, flaunt it while you got it
Ito's lemma is indeed missing in my code because I purposely eliminated it! Got it now?
I'm currently a part-time trader, not a researcher. As said it would take me some time to understand my own work after 3 years, that's normal in programming. One just needs to study the old code, but I don't have time for that wholly academic stuff at this moment.
And: WHY ON HELL SHOULD I POST MY CODE TO YOU?
I don't want your code.
I want you to take the standard BS formula, from wikipedia, and highlight the part that you think is 'ito-lemma' that you eliminated. That's it. No private information you need to give up.
You can't do that because ito's lemma is part of the derivation, not the formula itself. And since you keep talking about code, you are talking about implementation, not derivation.
Again, to be clear - I don't want your IP; I don't want your 'code'. I just want you to point out where in the standard wikipedia listing of the BS formula is this 'ito-lemma' you keep claiming that you've made 'unnecessary'.
I want you to take the standard BS formula, from wikipedia, and highlight the part that you think is 'ito-lemma' that you eliminated. That's it. No private information you need to give up.
You can't do that because ito's lemma is part of the derivation, not the formula itself. And since you keep talking about code, you are talking about implementation, not derivation.
Again, to be clear - I don't want your IP; I don't want your 'code'. I just want you to point out where in the standard wikipedia listing of the BS formula is this 'ito-lemma' you keep claiming that you've made 'unnecessary'.
Yes, I'm talking only of an implementation w/o Ito's lemma.
If you can't come up with a solution w/o Ito's lemma, then, hmm. it's obviously your own problem I would say...
Here are the steps (in German):
// step0: stddev_for_t berechnen
// step1: z ermitteln:
// step2: z0 berechnen, d.h. halben stddev_for_t <-- das ist der KEY von BS !
// step3: z1 und z2 berechnen
// step4: Call berechnen
// step5: Put berechnen using Put/Call-Parity
I don't know what sle's 3rd column is supposed to tell us,
but when I do the same calculation I get different values (year=253 trading days in my engine):
Code:
Period(253 days) : Vola=20.00% +1SD=22.14% -1SD=-18.13% C/+1SD=22.14% C/-1SD=0.00% P/+1SD=0.00% P/-1SD=18.13%
Period(253 days) : Vola=50.00% +1SD=64.87% -1SD=-39.35% C/+1SD=64.87% C/-1SD=0.00% P/+1SD=0.00% P/-1SD=39.35%
Period(253 days) : Vola=80.00% +1SD=122.55% -1SD=-55.07% C/+1SD=122.55% C/-1SD=0.00% P/+1SD=0.00% P/-1SD=55.07%Yeah sorry, duration is the only sensible usage. Anyway, rates are low, but any model that results in that input is flawed as you could trade the conversion at zero rates. Obviously not going to happen in reality, but it makes the model output GIGO. So I fail to see the utility of backtesting a model that reflects zero rates. Like modeling XOM at zero dividend (inside the payout period).
In the 100C/P listed... the model trades the conversion at par which implies the 40-day fwd is at par as well. IIRC it's worth $0.07 or so.
Maybe I am missing something as I have only glanced at this thread. What is the sheet supposed to show us?