So, let’s get right to the table!
Date Cases % Increase N Cases N Tested Prevalence Deaths
3/19 422 46.5 134 1 in 14
3/20 585 37 163 1 in 12
3/21 753 29 296 1 in 7
3/22 1,049 39 896 1 in 2
3/23 1,285 28 236 1 in 8
3/24 1,535 22.5 250 1 in 8
3/25 1,855 21.5 320 1 in 6
3/26 2,538 37 683 1 in 6
3/27 3,026 19 488 1 in 8
3/28 3,491 15.4 465 1 in 8
3/29 4,596 31.6 1,105 1 in 4
3/30 5,056 10 460 2,684 1 in 6
3/31 5,994 18.5 938 4,779 1 in 5
4/1 6,980 16.5 986 5,159 1 in 5
4/2 7,695 10.2 715 3,272 1 in 4.6
4/3 8,904 15.7 1,209 4,392 1 in 3.6
4/4 10,357 16 1,456 5,533 1 in 4
4/5 11,256 8.6 899 5,402 1 in 6 274
Why is this our best Illinois day? I’m glad you asked!
When you combine a robust 5,402 testing number with an 8.6 percent 24-hour new coronavirus case increase – our lowest increase yet – it’s beyond statistically significant. If our 899 new case number isn’t just a downward blip, though I suspect it may be, we’re in really good shape.
But what’s utterly undeniable is that more testing is NOT leading to a commensurate increase in new cases, and if that lack of a direct correlation holds up, Illinois may already be on the COVID-19 downside.
We’ve tested 58,983 citizen, the Illinois mortality rate went up a notch to 2.4 percent, the prevalence backed off to 1 in 6, and we held steady at number nine in the coronavirus case state rankings.
So, now that we’ve discussed this science for a while, let’s have some statistical fun! I know that sounds like an oxymoron but let me try to convince you otherwise.
My friend Paul sent me a column by Nate Silver of 538 Blog fame. If you recall, Nate is renowned for applying a proprietary statistical model to uncannily predict sports and political campaign outcomes. Much to my and every Cubs fan’s chagrin, 538 forecasted the Cubs would finish a disappointing 84 and 78 last year and that’s exactly what they did!
BTW, if you think our statistics lessons are somewhat daunting, read Nate’s rather long column linked above. It’s really quite fascinating, but to spare you that Herculean effort, Mr. Silver essentially says coronavirus case and testing numbers, when considered by themselves, are meaningless.
And they are!
Just as we’ve regularly discussed, Illinois coronavirus case numbers are rising because that’s exactly what we’d expect as we get a better handle on the disease. Put more simply, an increasing number of daily cases doesn’t necessarily mean social distancing isn’t working. Silver also duly notes that countries with lower testing rate will have fewer “official” cases, and vice versa.
But while we’d be wise to heed Nate’s warning, if you combine a healthy dose of pattern recognition with some serious statistical cross checking, you’ll always come out alright!
Once again! I’ve been citing political polling as the most basic form of pattern recognition. If pollsters can access a statistically significant sample of the appropriate voter universe, they’re rarely wrong, because the odds of encountering a counter-trend with a sample group are infinitesimal. Homo Sapiens instinctively moves with a herd mentality.
Here’s a perfect example!
On election night, a Kane County Board candidate I managed finished ahead by just 18 votes, but despite 64 outstanding absentee ballots, I called to congratulate him. Why? Because it was statistically impossible for him to lose.
Of those 64 ballots, only half, or 32, were likely to come back. That meant his opponent would have to secure 26, or 81 percent of those votes to snatch victory from the jaws of defeat. And that’s statistically impossible. As the County Clerk told reporters, absentee ballots always break along the same lines as the polling place vote.
Sure enough! After the election results were certified, my candidate prevailed by a scant seven votes! But the best part of pattern recognition is understanding how indirectly related statistics will invariably affect each other.
When was the last time you heard about vote by mail ballots changed an election night result?
The reason I was concerned with some states’ climbing mortality rates is because it could’ve indicated a gross coronavirus case undercount. Thankfully, with the exception of those four states, the rest of the country sits well within the mortality norm.
One reader tried to tell me, based on some ill-defined CDC “recovering cases” data, the government was hiding a 67 percent COVID-19 mortality rate. My response was, “Where are the bodies? Why aren’t our hospitals overrun with the dead and dying?” I live within eyesight of Delnor Hospital and that would be kinda difficult to hide!
So, how do we know our trend is still intact? Because every last number supports it!
Coronavirus cases continue to rise, but at a consistently slower rate. Though our testing capacity has doubled, the number of new daily cases has not. The prevalence generally remains between a 1 in 6 to 1 in 8 ratio, and, as expected, the Illinois mortality crept up a bit, but it’s well within the bounds of median reason.
Furthermore, my medical professionals tell me the local hospitals are well under capacity, a greater number of formerly infected people are recovering – the Mayor of Aurora among them, and all the epidemiologist are revising their original terrifying forecasts downward.
When all the data lines up like it has, there’s only one possibility! Our social distancing efforts are working. While Nate’s certainly correct regarding those seemingly lonely numbers, cross checking your statistics is almost as powerful a force as compound interest!
Put more simply, our good work is working, so let’s keep it up!