intraday forecasting
if 90% confidence intervals could be plotted into the future with options like formulas, or using neural nets, or hidden markov models, or probability techniques and pattern recognition probability analysis so that volume and price movements, based also on RT, recent past, and stochastics,DPO and ema velocity and acceleration, like kalman filters, both kinds, stochastic, could be used as multiple indicators of future mean price, this would be very useful for intraday and yearly graphs to give some kind of clue. it is not 100% reliable, as new news on a company or natural disasters or man made disasters can have a big impact. Read the Black Swan by Taleb to get more of a clue about the probabilistic problem and fat tails. Monte Carlo analysis can be carried out to give the Conf. Intervals. Read Also the quants to be sure we are not trading at the millisecond level like the supercomputers at the exchanges and elsewhere.