(2009.Oct.01 07:15 AM)fngMalvos Wrote: But why not try to fit the whole equation with all the data: Happiness, End and Energy? The format Dingus posted should be enough to determine if a given equation is valid if we get a wide enough range of people.
That is the end result. I'm not sure how much Zen would frown on it though. We obviously can't get the exact one, but my current model for example is accurate to within about 2% on most of the data I have tried. I am definitely missing some elements though as some outliers are almost 5% off and I haven't tried to integrate the trainers/ donor vs non-donor numbers at this point.
I haven't posted my version just because I don't want to spoil Biff's effort. He may come up with a more accurate model using different assumptions, and I don't want to sway him to my assumptions and risk missing a factor he takes into account.