Kenneth Chang

For Italy, it's sublinear. For the U.S., it's straight, with a slight upward curve the last few days, which could be the result of increased testing in New York, but even that is a slight deviation.

Which again makes sense because the US has been jacking up the number of test kits it's been ordering and tests being given. You completely fail to take into account rate of testing. That's poor science, and on par with what antivaxxers do with autism rates!

Linear + linear is still linear, with a steeper slope. It's not exponential. You can scribble that down calculation, right?

The operation isn't addition but multiplication.

Look. If you have 5% of the population effected, randomly sample, and you run 100 tests, what's the average number of positives you'd get? 5. If 10% of the population is effected, and you run 200 tests, what do you get? 20. If you have 15% of the population effected, and

you run 300 tests, what do you get? 45. If you have 20% of the population effected, and you run 400 tests, on average what do you get? 80!

My goof. But still not exponential. On a log plot, the curve is quickly bending and you would easily and quickly infer the underlying linear character of the spreading if that were indeed the case.