In pre-print on Feb 7, this paper initially estimated R0 to be somewhere between 4.7 and 6.6. People dismissed it as being not possible. It was never picked up by the MSM.

Now we're seeing from the both the real world and other studies that it's pretty plausible.