The wandering drunk analogy is pretty dumb. Nobody seriously imagines crossing a busy highway is particularly safe while sober, much less while drunk. Even drunks know that. What he is really observing here is lack of appreciation for the effect of variance when analyzing statistical probabilities. Here is the essence of risk management: understanding the probability of an event while weighing the impact of that event. Multiply the two to get risk factor.
You mean calculating the Expected Monetary Value (EMV) of each risk event. That's still problematic because it uses single-point estimates instead of true variance. Savage is correct in saying the best tool for the average person is Monte Carlo simulation, which teaches us to see risk stochastically vs. deterministically.
I don't think it's dumb at all. The importance is not to the person being drunk, or even about crossing a busy highway. The point is about how if there is a random movement that has an average that is safe, depending on the cost of deviating from that average, the result could be catastrophic. And why considering only the average can be dangerous. He literally says, "suppose his average position is at the center line of the highway".
I really liked that interview, I think I might buy his book.
Does this seem like how we the whole world is not handling coronavirus correctly?
Hmmm...I wonder what the Vegas odds makers think of this?
The wandering drunk analogy is pretty dumb. Nobody seriously imagines crossing a busy highway is particularly safe while sober, much less while drunk. Even drunks know that.
What he is really observing here is lack of appreciation for the effect of variance when analyzing statistical probabilities.
Here is the essence of risk management: understanding the probability of an event while weighing the impact of that event.
Multiply the two to get risk factor.
You mean calculating the Expected Monetary Value (EMV) of each risk event. That's still problematic because it uses single-point estimates instead of true variance. Savage is correct in saying the best tool for the average person is Monte Carlo simulation, which teaches us to see risk stochastically vs. deterministically.
I don't think it's dumb at all. The importance is not to the person being drunk, or even about crossing a busy highway. The point is about how if there is a random movement that has an average that is safe, depending on the cost of deviating from that average, the result could be catastrophic. And why considering only the average can be dangerous. He literally says, "suppose his average position is at the center line of the highway".