H1N1 ... again

I'm always amused by the "mathematical" fitting of curves to the worldwide cases of H1N1 virus. I started with the standard Logistic fit. How well the curve fits the data is measured by the Root-Mean-Square (RMS) error. That currently gives this: Note the predicted Eventual Value. I migrated to a sexier curve, with more parameters and a Weighted error (where the recent errors are weighted more heavily than the ancient ones)... and I got this: Note the predicted Eventual Value. Of course, one could assume (almost) any curve with sufficiently many parameters and get a reasonable fit. Indeed, even something simple like: y = A xb gives a respectable fit: However, it's that Eventual Value that should be reasonable. von Newmann said: "With four parameters I can fit an elephant and with five I can make him wiggle his trunk."