I'm interested in analysis of day-on-day momentum of certain large indices. In particular, I'm interested in the predictive power of the sign of the price change of the first hour of trading with respect to the sign of the difference between the open and close price. For example, I'd be interested in the probability of a positive day-on-day price change conditional on losses/gains in the first part of the trading period.
Does anyone know of attempts of analyzing this idea? If not, would you expect a sufficiently great predictive power to make investigating this worthwhile?