1 case on day 0 2 on 1 4 on 2 8 on 3 16 on 4 32 on 5 64 on 6
7 day average is only 18 cases but the next day you have 128 cases and the average is still only 119. If there's a major down turn in case numbers, the average will be slightly higher. This is an extreme example but my point still stands.
Its still a helpful metric when you have infrequent reporting since the incubation time is not consistent, not everyone is getting tested immediately when they feel symptoms, the time a positive test result and official reporting may be delayed, and there are spikes of infections that likely occur on weekends or holidays when people are having social events. The best data would come from daily testing of everyone with immediate reporting but that simply was not possible.
So you can predict the future? OF course the data lags, it also lagged at weekends because the staff keying in the stats only worked Mon - Fri. and they also had keying in mistakes which meant only coming clean _once_ via the BBC news.
Now it just so happens the US and China back in 2015 had wanted to set up a global body to avoid such pandemics but Europe didnt want it, which is why every country did their own thing. You can read the studies from 2015 detailing Covid from a joint effort between Wuhan virology institute & the Uni of Virginia!
The whole thing reeks of stupid criminality at the top.
Actually, no. That's not what "rolling average" means at all. It means the average number from the previous seven days. There is no "lag."
It means that on May 9, it is the average from May 2, 3, 4, 5, 6, 7, and 8.
Then on may 10, it is the average from May 3, 4, 5, 6, 7, 8, and 9.
And so on.
Getting something that basic wrong makes me think there are other fundamental flaws in the rest of what you wrote.