What is a good process to sample data (option prices and the corresponding underlying prices) to calculate implied vols and the IV curve. I mean the problem I am facing is this: say i sample the prices (bid and offer prices for options and underlying) at a time t. Now, I dont know what happened just before t. If there was a large change in the underlying prices due a large order, then the curve i would end up with would be out of sync with the "actual" curve in the market. Also, if I sample a number of points, I might be able to reduce this effect, but might not be able to totally get rid of it if such a scenario keeps repeating.
Does anyone have pointers to share?