What's fascinating? The fact that reality doesn't match the predictions produced by your analysis? That is a shame, I agree but atm I'm a bit busy and not in a position to adjust reality to fit your model predictions, please accept my apologies on this matter. I know it'll be a bitter blow but you just might need to face the inconvenient prospect of having to deepen your model in order for it to produce less unreliable projections.

Cheers

Graham