I'm a bit confused about how the simulator takes the individual results for a specific simulation (using randomized first choice), and then derives the results summary from it. I ran a simulation and requested the individual results. I exported the individual results to Excel, and assumed these results could be used to derive preferences by assuming that whichever product the respondent had the highest likelihood (percentage) for was the one they would pick. This evidently isn't how it works as I tried it and got vastly different results to the simulation summary.
Is there some rule for converting the individual results (which do not provide raw or centred utilities but instead provide % likelihood of preferring the product) into overall share of preference? If not, I'm struggling to work out what the use of the individual results option is.
Thanks in advance,