We know that extrapolation is not very meaniful, but why do actuaries and statistician still use it?

Question

Can extrapolation ever be useful?

Answer 1

Extrapolation is useful every day. It is used to predict future trends or events by estimating new values from a set of known data points. These values go beyond the boundaries of he original data. The method is used, for example, in population forecasting and weather forecasting, and is used daily in all the sciences and in engineering. In archaeology it is used to help predict site locations and to develop palaeo-environmental reconstructions. Yes, I would say it is useful, and if used correctly and critically it is also meaningful.

Answer 2

Why do you think it is not meaningful? It gives you something to relate a guess to, so it narrows down the likely answers so you can make an educated guess.

Answer 3

Actuaries and statisticians still use it because it give meaning to their lives.  Most of these extrapolations are used to give some logical rational to what is basically educated guess-work.

This is especially true in the case of insurance companies setting policy rates.  What you are not told is that the company always adds in their own "fudge factor" to ensure that they will remain solvent and profitable.

Occasionally, health problems like the spread of HIV and unforeseen disasters like the tragedy of September 11, 2001 show that their extrapolations are usually in vain.

Extrapolations can be useful as generalizations, but then again, generalizations seem to get humans into more trouble than handling everything on a case to case basis.

P.S. Anyone, including university level researchers, who use extrapolations probably see them as necessary and very helpful.

Answer 4

??? Sure it is. It's just that you hav to be carful (sorry, my E sticks!)

It's useful; you just can't do it mindlssly. You nd to ask yourself whether it maks sense in any particular contxt.

So, you can extrapolate a child's growth, but not if you carry it on past th growing years.