I used the reported fatality rate as of 9/10/14. The model allows me to adjust this for people who have contracted ebola within the last 3 weeks (and are included in the reported cases number), but have not yet died. In the adjustment, I apply the same fatality rate to these recent cases, but divide by two, assuming that of all the recent cases, half of those who will die have already done so.
So your CFR is 50%, which I think is at the extreme low end, but at this point the CFR isn’t certain.
Thanks!