I'd like to do a montecarlo simulation of a $\Delta$ hedged strategy (long OTM call) to see how the PnL distributes on cases like:
- $\sigma_{bought} < \sigma_{realized}$
- $\sigma_{bought} > \sigma_{realized}$
- $\sigma_{bought} = \sigma_{realized}$.
For this, I calculate the purchased option price with $\sigma_{bought}$ at $t_0$ and then do a random walk for the underlying asset $S$ so I can modify the hedge amount on each step.
My problem is that in order to calculate the hedge amount variations on $t$, I do not only need a new $S_t$ but also a volatility as $\Delta$ depends on vol (at least with BSM formula: $N(d_1)$ where $d_1$ depends on vol).
Question: What volatility should I use on each step to calculate the new $\Delta$ of the option?