Relevant, but off topic due to too much unnecessary detail:
I did a my disseration on BL, so I can answer your question somewhat. Technically you can calculate without excel the efficient frontier of each portfolio. You use calculus in respect to a desire level return and solve to minimize variance (standard deviation^2, so power rule)…. However if there are more tha two assets in the portfolio you need to use matrix multiplication, which I don’t think most ppl taking the CFA is able to do. The minimum variance portfolio is not the with the highest Sharpe, but lowest variance on the frontier. I think it’s just term confusion here, each corner portfolio will have the highest sharpe for that level of return, but there is only ONE minimum variance portfolio. We assume the highest Sharpe is the market portfolio due to supply and demand (CAL and BL uses this logic). If there are assets with a higher sharpe than market, ppl will buy it pushing price up (cost base of the asset is higher) and return down.
When CFA tells it to solve using the corner portfolio, it is a quick and dirty way that ignores correlations. This will over estimate the actually VaR.
There is an easy way to see if a portfolio is a corner portfolio or not due to limitation of knowledge and time on exam. If there is a portfolio with a lower return and a higher standard deviation than it’s neighbour ( P1: R= 6% Stdev = 11%, P2: R = 7% stdev= 16%, P3: 8% stdev= 12%). You know that P2 is not on the frontier usually do not kink (I have used actually data and gotten a zig zag line, but my frontiers allow leveraging and short selling of all asset classes), due to a whole lot of mathematicially and supply and demand factors.