Newton-Raphson and Brent for Implied Volatility

1. The Black-Scholes vega is $\mathcal{V} = S e^{-qT} \sqrt{T}\, \phi(d_1)$. Why does Newton-Raphson fail for deep out-of-the-money options where $d_1 \gg 1$?

2. Brent's method requires an initial bracket $[\sigma_{\mathrm{lo}}, \sigma_{\mathrm{hi}}]$ satisfying $f(\sigma_{\mathrm{lo}}) \cdot f(\sigma_{\mathrm{hi}}) < 0$. For implied vol inversion of a European call, a valid universal bracket is:

3. Near the root $\sigma^*$, Newton-Raphson converges quadratically. This means the number of correct decimal digits:

4. The Black-Scholes call price is strictly monotone increasing in $\sigma$ for all strikes, maturities, and vols — therefore the implied vol root-finding problem always has a unique solution for any market price.

5. In the hybrid Newton/Brent implied vol solver, the bracket $[\sigma_{\mathrm{lo}}, \sigma_{\mathrm{hi}}]$ is updated at each iteration. The update rule is:

6. The Brenner-Subrahmanyam approximation gives the initial guess $\sigma_0 \approx \sqrt{2\pi/T} \cdot C_{\mathrm{mkt}} / (S e^{-qT})$. This is derived from which ATM approximation?