I would expect a much higher spike in %positivity if that were true.
Eh, the current value is pretty bad. They generally use above 5% as bad, below as better, and somewhere under 2.5% as good. In the 15%-20% range, the stats are practically useless for knowing the true total cases.
I'll bet a week's worth of Genie+ and maximum IAS that the case curve will not be backfilled.
How much backfill to collect?
I saved the CDC data set when it was through 08/22/2021 and again when it was through 08/30/2021 for both cases and deaths. Both of them have backfilled data for dates prior to 8/22/2021, but it's not equivalent for the two metrics.
Based on the differences, the death lag is really obvious. Basically, looking at the daily death charts today on 9/1/21, all the data through 8/1/21 isn't changing much anymore, just one or two here and there. All the data from 8/1/21 through 8/30/21 is very different between the two death data sets in a meaningful way. So, we can look at the daily death charts, but only as of 30 days in the past not through today. People have been saying that for ages through, it's why the last 2 weeks on the death chart always looks like a huge decline.
For the case chart, it was backfilled too, but the delta between the data sets was net 96 cases between 1/23/20 and 8/22/21. Some days up, some down, with the max additional 66 cases one day. There must have been a lot of days reduced too, since when I did this between 8/25 and 8/22, the 8/25 data set had 582 extra cases vs the 8/22 set. None of these numbers are different enough to make a material change to the curve shape. They're spread out over enough time, they can barely even shift the entire curve up and down. Compared to the total cases over the entire span of time, it's barely anything. But, compared to looking at just the latest daily new number, some days it is a significant amount.
So, there is a backfill happening, even on cases.
It's doesn't look large enough to change the curve shape.
It may have a slight impact on the curve height, but not huge. At least not between data sets a week apart.
It does hide some data in the past. Enough that the most current daily number is probably misleading.
I'll need to wait for the CDC to publish FL data through 9/22/21 before I can compare the data through 8/22/21 on what's known a month apart instead of a just a week to see how it holds up then.
Based on that so far, you may have technically lost the bet, but not in the spirit of of having a meaningful historic impact. I don't think anyone could collect it from you, yet.