Well with your parameters, the prediction will have massive range of errors. Besides, how will you even model this much chaotic distribution; with no reliable data. Good luck with it though.
I think only way is to see CIC run its one year course, or even 6 months, the distribution will become...