CV and mad panic behaviour

[mateinone]

Hi mio.

Long time.  I largely agree with a lot of what you've written.  Can I ask why you have opted for 7 or 10 days vs 14?

Its hard to disagree with what you've written.  I think we ran out of the first lockdown too fast and paid for it.  We eased 3 times in 3 weeks with 7 days as the target, and it went backwards.

[member=105]Thryleon[/member]

You asked me this a while ago, I will share something I posted on my facebook page
It is pretty long-winded (about the norm for me  :-[ ) and it is just an example of how the longer average timelines mean it takes longer to determine your outcomes, but also leaves you behind the curve in relation to making relative decisions (imo)

[img width=1100]https://scontent.fmel7-1.fna.fbcdn.net/v/t1.0-9/120217738_10159055656953760_1181839184017882066_o.jpg?_nc_cat=111&_nc_sid=730e14&_nc_ohc=Rwqd4Ui-I44AX9ohk_x&_nc_ht=scontent.fmel7-1.fna&oh=3525eb62750e9fd89e9b0971c61299c9&oe=5F95D410[/img]

Why lead time and average days in the stats matter..
So I have banged on an on again about the number of days that are being used for the averaging and the number of days needed to determine if the steps are working (to be sure you are not releasing restrictions to early or tightening them too late).
I have been particularly vocal around the 14 days averaging and around the 28 days that was deemed necessary (now reduced to 21 days).
I have tried to do up a little graph that shows this. What this particular graph is showing is decisions made, 7 day average loads, 14 day average loads as well as as where we were at 16 days after a decision and 21 days after a decision.
I was going to add in 28 days after a decision, but decided to leave that out, because it is quite clearly obvious by the data shown above that 28 days is absolutely not a useful piece of data to respond to a virus such as this.
So the first part
7 days vs 14 days.
The problem as shown with this data is that using 14 day averages means your data is always behind the curve, it shows it on the ascending and the descending phase, which means it is difficult to make decisions in a timely manner. On the way down, that might mean keeping restrictions for longer than is required, on the ascending phase it will likely mean that by the time your decisions are made, they need to be much stronger and for longer than what was previously required, we know when this virus cuts loose that every day matters.
With regards to the 7 day averages, I have stated previously that around 14/15-18 days are required to see if the previous steps have been successful, but this is just to conclusively state this, in reality you can already often see these trends forming over the previous days, but the 7 day average show it around this time.
With the 14 day averages as you can see below, 21 is not completely unreasonable, but in reality it should be visible a few days before this.
The other thing with these elongated averages as it makes it difficult to determine if previous steps have impacted the spread.
Take for example if you use a 14 day average after the stage 3 restrictions with mandatory masks were introduced. There is a leveling at around day 16 and a clear drop at day 20, the problem is that this day 20 is also day 10 after the level for restrictions were imposed, is there enough information to draw a clear conclusion that the masks had a serious impact on the virus? Well you can make a case, using that data, but there are questions.
With the 7 day average, we can see a clear turning at around day 13 in this graph and by a few days later (days 15/16) it is conclusive.
Of course these cannot be attributed to the level 4 lockdowns, as they were only introduced 3 days before the worm began turning.
What is also significant is that whilst it is reasonable to make a case that you can see the impact of the level 4 restrictions (you would have expected more of a leveling out), it is also fair to say that the biggest impact was seen when the level 3 restrictions were put in place, which indicates that it is not unreasonable to at least propose that level 3 restrictions with masks is all that is required to manage the situation in the current environment.
Now in any case, 21 days is "acceptable" as a method for monitoring this during the descending phase (clearly 28 days is not), but it does still seem like overkill. People can say, well what is an extra say 3 or 5 or 6 days, well if it is say 5 days now, 5 days the next level and 5 again for the last, it is an extra 2 weeks, so it is still significant, but nothing like the significance of the original dates published.
Anyway, I hope the graph is clear, it's sole intention really is to show the impact of the different timelines on decision making.