Dear Quacks & Experts: Your Loss Function Is Not Our Loss Function

Note to the colleague who wrote me. I lost your email! You asked excellent questions about the post Classical Statistics Has Outlived Its Usefulness: Here’s The Fix, which I meant to answer, only now I can't find the email, even in trash. I am terrible with emails and names and always behind. If you read this, please email again, and apologies.

Going to start with something that will at first seem random. But stick with me.

Lithium is a known anti-depressant, and has been associated with a lowered risk of suicide and increased docility. Suicide is undesirable, while docility, especially in the time of occult white supremacy, is desirable. Therefore, lithium should be pumped into all municipal water supplies.

Make sense to you? No?

What if an "expert" said it? It is convincing then? Still no? What if that expert, a genuine expert with a credentialed certification in expertology, is employed by the mighty United States government?

All right, hold that in mind and let's move to the next random argument.

The odds of winning the jackpot in an average Las Vegas slot machine is about 1 in 49,836,032. Call it 1 in 50 million for ease. These numbers are true and reliable. Winning the jackpot is desirable.

Therefore, the government should mandate each American citizen (people who broke the law to come here are excused), once a week, bet a minimum of $1,000 on slots, as replacement for Social Security.

This one make sense? Same modifier as before: this is expert opinion. You going along?

Now let's move from these two seemingly random arguments to one with which you are more familiar.

The risk of dying with, and not necessarily from, coronadoom for those under 44 years old is no higher than 0.0001700, and indeed is much lower the younger one is. The risk of dying from something other than coronadoom for the same age cut off is at least more than 14 times higher, at 0.00240.

Therefore, you should not be allowed outside without a mask, potentially forever, and you can't go to your "non-essential" job, either, also potentially forever, because you might die with, and not necessarily from, coronadoom.

Before you scoff about that "forever", thinking it satire or a bit of fun hyperbole, read this article about an crazed expert embedded in the UK's grossly inaptly named SAGE group. This professor John Edmunds said face masks might have to be worn forever.

Not just face masks forever, but other restrictions, too, such as the Rule of Six, which allows police to arrest or disperse groups of more than six. Why not seven? The science is too complex for you to understand.

Here is the point of these seemingly disparate arguments.

It does not follow that because lithium might lower the risk of suicide and increase docility that everybody should be forced to take it. It does not follow lithium (or soma) should be mandatory even if experts, yea, even government experts, decide it is a good idea.

Experts, incidentally, did decide lithium in water supplies is a good idea. It hasn't caught on with governments or the public yet, but only because they haven't worked themselves into a sufficiently fearful state over suicides or the presence of the non-docile.

No act follows from knowing a "risk" or a probability, such as winning a jackpot or dying from a disease. Just because winning a jackpot is good does not mean betting is. And just because dying from a disease is bad, it does not mean removing liberties to avoid it is good.

This point was made in more depth earlier in this article about Scott Alexander's Bayesian Rationality.

Experts do not know what I have just told you.

Experts have decided that since they know more about facts, like how lithium causes docility, that they therefore get to decide on the morality and propriety of any and all acts based on these facts.

Even if we grant experts their claims of knowledge---a generous and absurd admission---it just does not follow that experts know best about what to do. About anything.

Because a doctor says you might have a malady, it doesn't mean the treatment the doctor recommends must be followed. The horrors of the treatment, to you and which the doctor will not suffer, may outweigh the chance the malady strikes you. Or you may prefer the suffering due to the disease over the pain and expense caused by the treatment.

What is important to a doctor is not necessarily important to you. The doctor doesn't want to be sued for malpractice, or must follow a rigid protocol, and is over-cautious. And, as said, he doesn't have to suffer the decisions he makes for you. You do.

In decision analysis we say the doctor's loss function is not the same as yours. The consequences of actions don't carry the same weight for everybody. But we now pretend they do. Because experts.

We used to know all these things. We don't anymore.

Now we have experts who say idiotic insane things like we have to wear masks forever. Pronouncements which we have to pretend to take seriously because the expert is in government employ.

This expert won't be tarred and feathered, or even laughed into oblivion. He'll almost surely be promoted.

Subscribe or donate to support this site and its wholly independent host using credit card or PayPal click here