Back on September 1st, Ontario reported 133 new COVID-19 cases.

@imgrund reported that the Rt value was 1.04, but less than a week later it was revised up to 1.2

2/
Rt is the current transmission rate. When Rt is > 1.0, new cases will grow exponentially. When Rt < 1.0 new cases will shrink exponentially.

See this table:

3/
Starting with 100 cases/day at the start of a month, an Rt of 1.0 means you'll have 100 cases/day at the end.

Rt = 0.9 -> ~50 cases/day
Rt = 0.8 -> ~25 cases/day

Every time you reduce the Rt by 0.1, you get half the cases/day at the end of the month.

4/
But the opposite is also true.

Starting with 100 cases/day at the start of the month...

Rt = 1.1 -> ~200 cases/day
Rt = 1.2 -> ~400 cases/day

Every time you increase the Rt by 0.1, you get double the cases/day at the end of the month.

5/
Through April, May, and June, Ontario did a good job of keeping Rt below 1.0. There were some blips above 1.0, but we did reasonably well.

https://twitter.com/imgrund/status/1299709840261087234?s=20

6/
Starting in mid-Aug, Rt started to creep up above 1.0 on a consistent basis. By Sept. 1, Ontario had about 25 consecutive days with an Rt > 1.0.

It makes sense then that we would see the doubling of new cases/day from about 75 at the start of Aug to 133 at the start of Sept.

7/
By that point though, the Rt was 1.2. Let's look at the table to see what that means. If Rt stayed at 1.2 for the entire month, we would see about 532 new cases/day at the end of the month.

8/
But Rt didn't stay at 1.2. It grew to 1.3. If we had a constant Rt of 1.3 for the whole month of September, we would expect to see 1064 new cases/day on the last day of the month.

9/
Well look at that. Rt has been hovering between 1.2 and 1.3, and new cases/day are at 700, neatly between 532 and 1064.

Math doesn't lie.

10/
Let's take things forward.

First of all, COVID has an incubation period of about 6 days. That means that numbers are going to increase exponentially for another 6 days even if we implement a complete, military-enforced lockdown today.

11/
Therefore, we're on a bumpy trajectory upwards to 850 new cases/day already. @fordnation and @celliottability have already made that bed and Ontario has to lie in it.

12/
In 2 weeks, hospitalizations will increase dramatically. Why? Because the most recent infections are breaking out of the 20-something cohort.

Good job uni students. You're sending mom and dad to hospital.

13/
Deaths will follow 1-2 weeks later. This is the pattern we saw in April and May. It will repeat itself.

Again, it's too late to do anything. @fordnation, @celliottability, and @Sflecce made the decisions that will lead to these consequences in the first week of Sept.

14/
If we act TODAY, and go back to the most intense lockdown, we will start to see new cases/day start to drop about 2 weeks later, just like we did back at the end of March.

https://twitter.com/imgrund/status/1299014520426422279?s=20

15/
The best time to take action to prevent widespread death was 2-4 weeks ago. The second best time is now.

If Rt stays at 1.3, and if the cases/deaths proportion remains constant, then I think that every week's delay doubles the upcoming deaths.

16/
@fordnation @celliottability @Sflecce: Every week you delay in locking Ontario back down will double the number of deaths in the province. Your delays for the last few weeks have already quadrupled that number.

Will you seriously let it double again?

17/17
P.S. When the death toll for Ontario in October is finalized, remember that it could have been 1/4 the size had the government acted earlier.

18/17
I wrote this thread venting into the void, but it seems people are paying attention to it.

You should follow @imgrund and @DFisman. Do it. Now. No excuses.

They're the heroes. I just get angry when people ignore the math.
You can follow @egerlach.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: