
Hi all 
Been a lurker and have tried to learn from prior posts, linked papers and books. I am trying to create an intraday trading strategy focused on trading S&P 500 Equities (including past constituents to avoid survivorship bias).
The strategy would go long near market open with the intention to capture a fraction of an estimate for the 1day Friday Volatility based on historical volatility. Specifically, this would use the Parkinson Estimator of Volatility in order to get a forecast of the intraday Friday move. In this case, daily OHLC data is corporate action & dividend backadjusted.
1. As I will be processing intraday stock data starting on the Friday open, does anyone have a recommendation for modeling a historical probability table till market close? E.g. if i have a HV daily Parkinson forecast of 4%, I would like to model the probability that the price will close past the open * a fixed percentage of HV daily forecast (Open P * (1 + (HV Daily * fixed fraction)):
Would it be recommended to try & model this as a cumulative distribution function? Any paper recommended for such a model? A. Probability that stock closes above 25% of daily HV forecast (Open P * (1 + (HV Daily * 25%)) B. Probability that stock closes above 50% of daily HV forecast (Open P * (1 + (HV Daily * 50%)) C. Probability that stock closes above 75% of daily HV forecast (Open P * (1 + (HV Daily * 75%)) D. Probability that stock closes above 100% of daily HV forecast (Open P * (1 + (HV Daily * 100%)) 2. Any guidance on a mental model for evaluating the lookback period for HV calculation in practice?
Thanks 


