If you're going to presume to lecture me about a "silly response", then will you grant me the same privilege to say the same about your response? If your qualification to make that pronouncements is that
"In my professional life I manage a team of analytic professionals in a project-based industry", might you be willing to consider that there might be other people who manage analytic professionals in project-based industries, where decisions need to be made on the basis of inaccurate or incomplete information, and thus are equally qualified to determine what is a "silly response"?
What I deal with professionally has everything to do with public health exposures. I've been doing this for over forty years. I have worked for public health agencies, and later as a consultant to government and industry. I have personally been in a position where I have been the person who has received information on actual or presumed public exposure information, and have been responsible for making recommendations to senior governmental officials on what should be done with that information. I had to make decisions on the best information available. Often, working with toxicologists, we had to construct some kind of exposure assessment model, often incorporating some broad assumptions. When you have 4-hour deadlne to come up with an answer, you work with the best information you can get your hands on inside of two hours. And then spend two hours working up your model.
When a crop duster crash lands into a canal that supplies drinking water, I've been the person on the line who has had to offer information about what actions should be taken and what notices should be made to the public. Or imagine having test results from a drinking water sample that shows bacterial contamination. But that sample was taken 72 hours ago. Does that represent the water quality that is currently being delivered to customers? Do we tell people to not use drink the water? What are the uncertainties - what is the error?
I've gone through the dangerous nonsense of Superfund Public Health Evaluations, which is a process that I've always thought was deliberately designed to make people think that they were going to prematurely die of cancer because they happened to drive by the site. I've had to do risk assessments based on the assumption of that someone was going to harvest wild watercress growing by a creek, and have watercress for dinner every day of their life for 70 years. I actually bailed out of doing that work because even though it was lucrative, I could no longer participate in such inanity.
So please understand that when I see something such as the IHME model in use for public health decision making, it's actually much better than most of the stuff I information I typically had at hand when I was involved in making those kinds of evaluations. Even with the error bars
*************
And I might add that you've managed to avoid the quite legitimate question. Pretend that you are the person advising people who have to made decisions. What are you going to tell them? How would you advise them to make decisions? Understand that saying we don't have enough information to give you advice is not a possibility. Because if you don't give advice, they are going to obtain advice from someone who is less qualified than you. And how would it would be better to have decision makers basing decisions on input from someone who is less-qualified?
**************
So I repeat - if you throw out the IHME, what do you propose a replacement? Because for you to say that we should throw information out, without offering a better alternative, is a silly response to the real-life situation for someone who is actually engaged in real-time public health decision-making.
I have been a statistician in the market research industry for close to 30 years now and some of that time has involved forecasting models. If I had given a client a model whose predictions had so much variability (like in the IHME models) or were so incredibly wrong, we would lose that client permanently. And we are dealing the manufacturing of widgets, not peoples lives, the entire US economy, and livelihoods.
I don't think you appreciate a key difference. In most marketing, the "No Go" decision is an option. How would the situation change if "No Go" is not an option. A "Go" decision must be made among the options that are on the table. Collecting more data or refining the model is not an option.
.
