Hi Not recommended
Category: Probability and Statistics
Why? This is both true and not racist at all since NonGaussian.
No need to respond to such stupidity. He did not get our idea. There are ONLY 72 pandemics >1000 people in recorded history. We are using conditional distribution.
No mushroombrain I explain that naive falsificationism is incompatible with nondeterministic (statistical) approaches but asymmetry is CENTRAL in ranking evidence and classes of distribution. See reference (my stat book) before bullshitting.
If the model is not reliable, don’t take the risk.
Lesson from a long experience with model blowups: @neil_ferguson , if you need a model w/”thousands of lines”, this is not a model useable for real world risk & decisions–rather something with the FRAGILITY of a house built with matches to impress some tenure committee.
Something UK modelers (idiots) did not get with hospitalization that we traders and modelers in nonlinearities have know since (1997): The market is a large movie theater with a small door. Same for hospital capacity. Gabish? Same with insurance models. Loss from hurricanes aren’t what you think are losses from hurricanes, even if your estimate of the intensity is correct. Why? Because the costs of the material goes up nonlinearly owing to (conditional) excess demand.
This is NOT because of absence of causal inference. This is because correlation is not a good measure of association because of its convexity. PROOF: the problem goes away when using mutual information. No causation schmosation.
This is not the base rate fucking fallacy. Please stop using names given by ignorant psychologists to statistical phenomena.
Wrong. Bimodality does not require continuity. Evolutionary biologists should stay out of probability.
