Hi all,I am a newbie and still a student, so would appreciate help and advance apologies for mistakes.Problem: Calculate VaR analytically for a portfolio of options. The options are on the same underlying except that they have different strike prices and expiration. Our solution: Calculate the greeks for each option and add them up to get the greeks for the whole portfolio. Then treat the portfolio as one option and apply one of the various analytical delta-gamma approaches to calculate VaR.Questions: Does adding up the greeks make sense? Does treating the whole portfolio as one option have any pitfalls? Further question: How do we test the effectiveness of this approach? In general, how does one test the "correctness" of the VaR calculated by some analytic approach? One answer - use historical Monte Carlo simulation as a benchmark and compare against it. Any other approaches?Thanks!

Typically, yes, a weighted average of greeks will give you portfolio greeks, which are still just an approximation. With Delta-gamma VaR, that may be good enough for what you are doing, but I prefer to be more conservative.The only way, IMHO to properly measure VaR for any position with third-order or higher effects (like any option, but the common greeks only go to second degree and don't include many cross-derivatives) is to re-price the entire option protfolio at the threshold scenarios, and stress test them for extreme values.One major risk delta-gamma ignores is vega risk, which alone can devastate you...Jorion, in "Value at Risk" describes this as "Full Simulation"