Hi all,
I am a bit worried. I ran a whole lot of simulations a few weeks back, and reported on them. I was asked to run a few more, and tried to replicate my old simulations to ensure I was setting everything up the same way, and I CANNOT replicate my old results.
The utilities and importances are all correct, but when I run the simulations I get different results from what I got before - some only minor differences, others fairly major. I saved my output from the earlier simulations and have checked against them - everything is identical except the results. Same weights, same sample, same segments, but different results. The results are (mostly) the same for my price sensitivity analysis (only 0.1 out in some cases), but are out quite a bit for ALL of my competitive sets.
It is an ACBC with an alternative-specific design, and I am using randomized first choice, with several continuous price variables. Why have my results changed? We'll be handing over the simulator soon and they are not going to be able to match what I have reported.