Modeling didn't do a bad job. The tornado threat was overlooked by a lot of people who were distracted with things like ERCs, Recon passes, ADT numbers, and nitpicking every run of every model.
The north wobble might be starting to correct itself, it's back on course towards Bradenton/ Longboat key. Pretty much the consensus over the last couple days.
Hmon and HWRF still have wind trajectories that look like they are sucking water out of the bay rather than pushing it in, and from this storm size/shape it would have to track way north through Port Richey for TB to get the maximum surge.
That's every global model showing landfall within about 10 miles. Amazing agreement, this didn't used to happen. Even 10 years back models would have a much larger spread 36 hrs out.
Hurricane models actually did a good job forecasting an ERC and multiple intensity peaks. Dorian actually did a similar thing. Maybe not an ERC, but there was a pause in the intensification for almost a day.