Smart analysis on the uncertainty we tend to overlook. We need to get better at understanding probability and just how much we cannot know, in politics, and in life. https://t.co/ur5E6RoYOf

— Damien Cave (@damiencave) November 1, 2020

The problem isn’t understanding probability, though. I understand probability just fine, certainly better than most people, but the actual problem is that we are assigning probabilities to areas where it’s just inapplicable, where it makes no sense at all.

Applying probability to low-frequency, non-repeatable, model- and forecast-influenced events is just insane. It cannot work. And it does not. Saying that Biden has a “90 percent chance” of winning is essentially meaningless with the uncertainty present even in a normal election much less one occurring during a pandemic with an unhinged kinder-President.

It’s bullshit all the way down. Since the above is all true, you can’t even do normal statistical things like calculate proper confidence intervals because you don’t even have enough sample data (elections) to even begin to gin one up. And realistically, you only have one election, which hasn’t happened yet.

Yes, yes, I understand that the models are running simulations in pseudo-worlds much like ours where various perturbations are applied, etc. But elections aren’t the weather and again, they occur only once in reality.

It’s all worthless, but we pretend like it’s science while quant-y snollygosters sneer at us even though they are little better than an auspex peering into a crystal ball. Don’t buy into it despite what the “experts” tell you that you must believe or you are “anti-science.” This isn’t science; it’s superstition and dominance displays dressed as science.