HiI guess most of those who read this thread will be familiar with derman's static replication of a variance swap. However, this replication only works smoothly for variance swaps that payout in the numeraire (accounting) currencies (e.g. jpy for a usdjpy variance swap) - for e.g. a self-quantoed variance swap (usdjpy paying in usd), this hedge is imperfect, leaving signfiicant 2nd order (Vomma = volga =volgamma) risk.Hence my question: does anyone have an idea about how to adjust the standard static hedge for a variance swap in the self-quantoed case (maybe something like a convexity adjustment)?also: does anyone have an (intuitive) understanding of what exactly accounts for/causes this difference in the risk profile of a non-quantoed vs the self-quantoed variance swap? Many thanks in advance for any help!!!