Ion Saliu's "Fundamental Formula of Gambling" (FFG) is described by Ion here with a formula. In brief, for a specified percentage "Degree of Certainty" (DC) you can calculate the number of immediate previous draws that will contain the numbers in the next draw for a given jackpot lottery game.
So, for a 6/49 game, a sample of results is (always rounding up):
DC # of previous draws to analyze
99.9% 53
99% 36
90% 18
Once you have the number of immediately previous draws to analyze, you can then check all these draws and determine what numbers are used.
I have tested the FFG on Canada's national Lotto 6/49 as well as the (Canada) Western 6/49. I have found it is not overly useful as a primary selection tool, but it can be useful as a secondary tool or as a "sanity check" for numbers selected by other methods.
I don't find it useful as a primary tool because I've found that with a DC of 99.9% it *is* accurate, however, it usually only eliminates 0-1 numbers of 49. At lower DC percentages, I've found it eliminates more numbers, but even at 99% DC there is an unfortunate tendency that 2-4 eliminated numbers actually do get drawn on the next draw (but it's not predictable enough to use that to my advantage, either) - I didn't calculate it out exactly, but it seemed to be occurring more frequently than the 1% I would be expecting from the DC.
Also, even when testing this with a DC of 50%, it still has a tendency to come up with 17-19 numbers, which is, for my liking, still a lot to wheel for a relatively low percentage.
So, I don't know if this is just a quirk of the games I'm testing with, or if that's the case for all games, but that has been my experience.
Hope this helps.