Understandably, we all use 7-day, or 14 or 15-day or even 30 day moving averages to review trends. This is very useful when the underlying data has wild swings.
This chart from the Iowa state government, counting total deaths due to Covid-19 illustrates the issue super well.
In the top graph, the red line is the 14-day moving average – note how it has time shifted the actual peak, to the right. See also the purple line in the 2nd graph – where the time shift of cases is more pronounced.
Is this a problem – not a big problem but coupled with the delays in reporting deaths, it further time shifts past deaths into the future.
Right now, when many metrics are improving greatly, the media is focusing on daily death reports – which I note, are necessarily of past deaths, and often of very distant deaths getting rolled up into “today’s report”. This biases our perception of the state of the pandemic.
My hypothesis is that the U.S. has probably peaked several weeks ago – but reports of past deaths continue to accumulate and are rolled into present day reports, time shifting past deaths into the present.