That model isn't consistent with data on the ground though. Many more states are showing clear and obvious growth over the past 7-10 days (which is about the best fidelity that measurement can achieve anyway). Check out the link below for my personal list of "worry states". A few might have ticked down in the past 2-3 days, but the overall trend of all of these states is upward and there's no reliable evidence of a genuine halt in growth.

https://91-divoc.com/pages/covid-visualization/?chart=states...

Edit: in fact I went back and spot-checked Iowa. That link you posted says that Rt dropped below 1.0 20 days ago. In fact the real world new case rate is almost three times higher than it was 20 days ago! I'll just say it: that site you linked is garbage.

There are more confirmed cases because there's more testing.

That explanation made sense for the hyper-growth mode (2-3 day doubling) in, say, NYC's early explosion, because there was giant pool of infected being discovered with new tests. It doesn't explain states like Iowa which are doubling only ("only", sigh) every 6-10 days. In that realm you'd expect positives to track linearly with test growth, and they don't.

I'm sorry. Those graphs you linked are pretty, but they're just dead wrong.

Doubling of cumulative cases in 6-10 days, after adjusting for increased testing, is not that far off from Rt=0.94. Even Rt=1.1 still falls within the uncertainty interval.

No. You're going to need to cite me some analysis with real numbers if you genuinely want to convince me that a 3x growth over 20 days is consistent with an Rt less than one.

I'm serious: that site is bad. It's a bad model. You have to stop reading and linking it. The only way to make it "right" is to assume (as you are) wild divergence between measured case incidence and "actual" unmeasured infection statistics, which isn't borne out in any papers I'm aware of from any outbreaks anywhere.

Basically: I'm telling you that measured data makes this model very wrong. And your response is to tell me that this is because measured data is itself extremely (like 3x!) wrong. That's bad science.

I'm not trying to convince you. I'm just saying that the data you quoted does not contradict the model I linked to. New cases in Iowa are essentially flat for a while now, which is consistent with Rt close to 1. https://coronavirus.jhu.edu/testing/individual-states/iowa