The Bayes' Theorem approach really isn't that helpful?

[tonychanyt]

Prof Michael Heiser said:

The Bayes' Theorem approach really isn't that helpful because you have to presume certain things and plug certain assumptions.

The Bayes Theorem is useful if you use it properly. You are not supposed to just plug in a subjectively arbitrary number based on your presumption.

You can more or less manipulate them to find: lo and behold, the person who wrote this post that Jesus doesn't exist, doesn't exist either, according to Bayes' Theorem.

That's because they failed to assign the weights correctly.

Once I have assigned weights to propositions, how can anyone tell I didn't do it arbitrarily or whimsically?

Fortunately, there is a mathematically sound answer to this question. The goodness of my subjective beliefs can be measured objectively through the process of wagering based on my weights. My personal beliefs should remain coherent, even if they are subjective. Formally, a set of beliefs and preferences is referred to as coherent if it cannot lead to a Dutch book; that is, my weighting scheme does not guarantee that I lose money in the long run due to my betting habits. If your beliefs permit this outcome, then you are incoherent. You are effectively committing to a losing money scheme due to your habit of being too subjective in your assessment. I apply my coherent weighting scheme to bet against individuals whose bets are not coherent.

You want to train yourself to be a reasonable and coherent bettor. In general, you can use Bayes' rule to make any life decision in the most optimal way. In practice, the more accurately you estimate the three input probabilities in the Bayes Formula, the better your decision will be. This is the essence of actuarial science.