He describes his invention: "Generally when an insurer performs certain calculations, it includes data from all its policies. If it has a million policies, that means a lot of processing as various scenarios are considered. Sometimes, the work can take days. I believed I had a solution to this cumbersome and costly process: create subgroups from the database, sample policies from each, repeat the process several times, then combine the results. My technique provides results similar to those from studying all policies, and saves time and money."
He doesn't note this but his procedure allows him to calculate not only the probable result but also a level of confidence in that result i.e., that this result will occur with a predetermined desired level of probability. He goes on to recount the actual process of getting the patent, but fails to note that there is nothing particularly original about his statistical process; it is a perfectly ordinary statistical sampling, designed to achieve a certain level of accuracy at minimal cost, based on a tested assumption about the distribution of the sampled population.
I wonder how long it will be before someone challenges his patent as neither unique nor original.