3 ms·
Not really. Last week, the error bars were 0.1% to 16%, which was much more uncertainty IMO. This Nowcast model is "only" a 3x fold from lowest to highest. So
by dragontamer 5y ago
Not really.
Last week, the error bars were 0.1% to 16%, which was much more uncertainty IMO.
This Nowcast model is "only" a 3x fold from lowest to highest. Sooooo... this is actually a lot more accurate and certain than last week's prediction.
- ianai 5y agoRight, that’s pretty much how a 7 day in the future prediction is going to look when rates can double daily. The pattern fits from SA data roughly though. But that’s not ceterus paribus, etc.
- jfim 5y agoIt really depends on the question. My observation was from the standpoint of "is omicron the most prevalent variant," which this week's data suggests probably, but it could go either way. Last week's error bars would suggest that the answer to that question would be "very unlikely." As you point out, a 95% CI of 0.1% to 16% is also huge.
- dragontamer 5y agoA 1% Omicron would suggest we have 6 doubling-periods before dominance, roughly 18 days. A 10% Omicron would suggest we only have 3-doubling periods before dominance, roughly 9 days (seems to be doubling every 3 days on the average). -------- The 30% vs 70% dominance question is "is Omicron going to be dominant on Thursday, or was it Dominant yesterday?". Which... isn't really that big of a difference anymore. We're just solidly in the era of Omicron now.