Create the line which best approximates a collection of points. The line fitter algorithm aims to minimise the fitting error of the resulting line. The error is calculated by summing the squares of the distances from each point to the line.

The less line-like the distribution of the given points, the less predictable the fitted line result will become. In extreme cases, very small changes to the input points may result in very different lines being fitted.

An example of such an unstable system would be points evenly distributed along a circle, where every possible direction through the circle is equally bad, and the final chosen direction is subject to rounding errors and cosmic rays.