Who cares what the best fit is. There is simply not enough data. You're talking twelves data points with considerable variation. Using statistic doesn't turn muddy and ambiguous data into a clear signal magically.

My point is that there is no extrapolation of the data we’ve seen to date that yields the blue line on their graph.

Sure, but I think they are tracing the worst case scenario based on the underlying assumption that the virus increases exponentially.

What going to happen is indeed inherently uncertain so we pretty much have to prepare for the worst plausible series of events. Once it becomes clear that a better outcome is fairly likely, then the plans can shift.