Open interest answers your first and second questions. If you look at the change in open interest across all strikes and expirations, I suspect that you'll get some idea of how exposure to / openness to risk on the product is changing. I suspect that monitoring volume over change in price (was price going up or down on volume) might help provide some clarity regarding 3 and 4. Regarding 5 and 6, not sure it matters. Those are always open questions regardless of how one trades or what one uses to enter a trade.
Volume and open interest and different. One is a count and the other are open contracts. In a simple example open interest is zero. I buy 10 calls from you. Your software shows 10, the OI is 10 and the volume on the OCC is 20, just to confuse us. Tomorrow, I sell 10 to another person to close, they are opening. The volume that day is 10, the OI is still 10.
exactly, and if @Robert Morse would trade the last 10 to sell to the initial seller... then volume = 10 and OI after is 0... because the initial buyer and seller have reversed their trades with each other and have no position... and nobody else has either. The OI is only updated on the next day though, so when trades are done, the OI changes the next day.
This is also useful to know: the open interest lags by one day. This might be too slow, too late, depending on the trading strategy.
Not by a full day. At least for US options OI is available before the next day's open, usually by ~5:00 am EST. You can pick it up on the OCC report download page at no charge. @RobertMorse: one day log changes in OI ratios and ratios of ratios may be predictive of one day forward log changes in the price of the underlying.
How can you use OI to determine stock movement? EG. OI in a stock without active options, 100, than 5000 on the 25 call with the stock 23. How do you know if the person that initiated the trade, bought them or sold them? Would you then expect the stock to go up or down? If you have a better example, I'd like to see it.
One of the biggest mistakes I have made repeatedly in the past was to try to reckon things out from first principals. I ended up dismissing proposed features or indicators that were, counter intuitively, useful or predictive of my chosen dependent variable(s). In this case, I repeat, one day lagged log changes of ratios of ratios of changes in OI for different segments of constant maturity curves contain mutual information (Shannon Cross Entropy) with one-day forward directional changes in the underlying or its dominant driving or main idiosyncratic factor -- essentially Granger Mutual Information. Significant at the 0.001 level. This is easy enough to test if you have access to the data. Obviously constant maturity OI is a blend or fit of two or more expiries' OI's. Also with cross entropy calcs involving one discrete and one continuous series you may want to correct for binning artifacts.