Following on my "historical options data" thread (
https://www.elitetrader.com/et/threads/historical-options-data.358125 ), I went medieval on the data I have and gave up trying to "correct" it, rather just drop anything that's bad behaved.
The marketdata I paid for has about 10,000 underliers. After decimation, the universe of acceptable ones was reduced to ... 3%. 300 out of 10,000. Maybe I'm too drastic but then please, clean up the data yourselves. The combinations of unforeseen crap that gets thrown at you is only as creative as it is limitless. For instance I should have made a picture and put it here (especially since it happens on multiple underliers). You know how out-of-the money options are decreasing in price, since (obviously) they value less and less. Well if the number 42 is in The Hitchhiker's Guide to the Galaxy by Douglas Adams, the "Answer to the Ultimate Question of Life, the Universe, and Everything", in my marketdata this number gets changed to
5. Everyone knows that in real life out of the money options, where there is still liquidity, will quote the minimum acceptable price tick for the exchange: $0.01. Well not in my data. They quote ... 5. Option prices converge to ... 5. On a lot of underliers! W
! T! F!