You cannot run the model, wait for the output, run to your Governor and say “The latest model says social distancing works.” If your Governor had any sense he would say, “Didn’t you write the model code? And didn’t the code say somewhere that social distancing worked?”
Sunday, July 18, 2021
That's the way it ia …
… Models Only Say What They’re Told to Say — The Paper! (Hat tip, Dave Lull.)
Subscribe to:
Post Comments (Atom)
Models are meant to estimate what may be to come. They are never to be relied upon to make accurate predictions. It's rare that they do. But, here we have had a situation where we have had to place bets. It's like the conundrum that must be figured out, "Your going down a road and get to a fork. One fork leads to certain death, the other to your destination. There are identical twin brothers living at the fork. One always tells the truth. The other always tells a lie. You get to ask one brother one question, and there's no turning back." That's the situation we've been in. All models are meant to answer the question, "Which road shall we go down, given the facts we have before us?"
ReplyDeleteLet's say you and I are playing one-on-one basketball to 21 points. In the first game, you beat me by 2 points, 21-19. In the second, you beat me by 4 points, 21-17. The easy model by one onlooker placing bets, is to say you will probably beat me by 6 points, 21-15, in the next game. But, another advises, it could be 8, which is 2-cubed, 21-13, next game: 21-5, then shut outs from then on. Place your mandatory bets, everyone. (And as we know today, it is the unvacccinated who are dying. There's no such thing as not placing a bet. There's no going back from death.)
What did not get factored in, though, in either of the models, was that the curve was about to go down, not up, into my favor, as you get tired from not eating Wheaties. Now, we have a new factor like social distancing, in this case, eating Wheaties. Can we now predict that this is how our basketball games will go every day? Place your bets, everyone, and someone advise the Governor! No one can leave the court. It's urgent!
The longer we get into a situation, the better our models can be. Until, in the end, the only legit ones left are explanations of what happened.
The quote you made from the article, is not completely accurate. You run your model, you go to the governor and say, "I have a model that shows such and so."
Here is a key element missing in Briggs' critique of models. Briggs allows that they are "scenarios" versus definite predictions. Then speaks as if anything being false is a bad thing. Everyone on the court knows that every model is going to be "false" as he says. Not so. Every model is only an estimate anyway, based on what might happen.
If a parent says to a child, "Do not run out into the road, you could get hit by a car." When the child does not get hit by a car, we do not say that the parent gave a false statement. When we say, "Keep a 6-foot social distance," we are not saying either that this will definitely keep one from getting the virus, only lessening your chances. We all know that 12 feet is even better. We are also not saying that if you do not keep your distance, you will definitely get the virus, although the Democrat lawmakers from Texas got their shots, broke the mask and distance warnings, and contracted the virus. The early social distance models were based on droplets, which remains a way to contract the virus. Now we know enough to imagine detecting an aerosol, like a rancid trash barrel on the sidewalk, that we want to keep a distance from. That aerosol stays in the air longer than droplets, and travels a greater distance to noses and mouth. It was a new factor for each future model.
Many of us have taken the wrong fork, and died, some got torturously ill. Others like us made it through somehow. We each placed our bets. We each had our made-up models. Some of us even said that it was all a hoax, even while dying from the virus. You bet your life no matter. And there has been no simple question out of this conundrum, like, "What would your brother say is the road I should take that will lead to my destination?"