What failure looks like

 In case you were wondering (and you know you were), this is what failure looks like:

Figure 1

Allow me to explain.

The UK reached its low in daily new covid-19 cases back on May 5th. The U.S. reached its low back on June 21st.

This is what the two countries' experiences have been in the time since each one's low:

Figure 2

(As the labeling on the chart indicates, the data are from Worldometers. I don't know if there's a way of downloading their data, so instead I set it to show me the 7-day average of new cases, found each country's low point, and then manually found and typed up the values. To save time, I picked up the values ever week rather than every day.)

This graph has two problems in terms of comparing the countries: the action isn't happening at the same time, and the U.S. is substantially bigger than the UK, so we expect to see higher numbers here, even for an equal severity.

We can align the two countries' timelines by counting weeks since the low, rather than calendar dates:

Figure 3

It doesn't look great for the U.S., but being bigger, it's reasonable that our low is a significantly higher number than the UK's low, and maybe our eyes are doing a poor job of correcting for that.

The obvious way of accounting for differences in population is to look at something like cases per 100,000, and that together with looking at weeks since the low gives us:

Figure 4

So, again, not looking great for us, but the UK numbers get so big over toward the more recent part of the data that the comparison at the left side of the chart gets squashed.

So let's look at just the four weeks for which we have both countries' data:

Figure 5

Oh.

That is not reassuring at all. We reached our low noticeably worse than the UK low, but not a lot worse. We were about 28% higher than them.

We're now four weeks past our low. We are more than double the rate that the UK had four weeks after their low.

We're getting worse faster than they were.

To control for each country's different starting point in case rates, we can normalize each country's rate so that its low equals 100, and compare the data on that basis:

Figure 6

One week after our low we were doing trivially better than the UK at their comparable time.

Two weeks after our low, we were a little bit worse off than the UK.

Three weeks out, and our course was clearly worse than theirs.

It's now four weeks out. In that time period, the UK had not quite doubled. We've more than tripled.

Where might this be headed?

Over its first 28 days of increase, the UK rate grew at 2.2% per day.

Over our 28 days since our low, we've grown at a rate of 4.0% per day.

From day 28 to day 56, the UK rate grew at 6.1% per day. Over the remaining 20 days of data, their growth slowed down to "only" 4.7%.

What is our next 7 weeks have the same growth experience as the UK?

That looks like this:

But of course, we're already rising faster than they were. What if our next 4 weeks are faster growth than the UK's comparable time, by that same 1.8 percentage points? Then we miraculously revert to their growth rate over the next 3 weeks?

These are both pictures of failure, one worse than the other.

On your birthday, you get to do things that you consider fun, and I generally enjoy playing with charts, but these are not exactly fun.