>>The nice thing is that numbers don't lie. The distance between the reality and the result they get is the same as the distance between the reality and the initial set of assumptions.
>
>I've read on a book by an economist
>
>He declares that something is like y = f(x) = x * 1/a.
>
>Then he talks many pages about what he can do with that.
>
>After that a tiny paragraph defines:
>
>Initial value [a] on this equation need to be intuitive guessed.
>
>Wouldn't it be easier to guess y directly? I mean I can draw a line with little help.

But when you think of it, 99% of economics (and other human interaction... areas of knowledge which are called 'social sciences' here, or 'humanities' in some other places) are like that. They go to lengths to describe the behavior of their model, and even get prizes for that, while they don't quite explain how did they get the initial set of assumptions into the model. Not to mention that (at least the old ink-on-paper ones, who didn't have computers) they don't really try different sets of assumptions. But many were shown (including Marx) to have fudged the numbers.

And yet their theories are put into practice.

nb. I have already put a bet against myself that my mentioning Marx will take this into a different kind of dispute from what I intended. Now this note introduces a counter-assumption which throws the whole theory off, rendering the model invalid. But I'll still stick to the original theory based on the model without it.