I was wondering if an anyone suggest a possible probability theory for a piece of work I'm doing. Essentially I have a sequence of values (all known values from a population), for which I'm looking to calculate the probable value of the next in the series within a range. However, this range needs to be variable, depending on a 0-100 input or measure of confidence (e.g., the probable value range of 90 is bigger than the range of 70, or vice versa). I suppose it's very similar to what a confidence interval would do for estimating population characteristics from a sample, but I'm not sure if this is exacty what I'm looking for (i.e., estimating from a sequence of data). Thanks in advance.