When you have measured data points in an assay you can plot these in a graph for example in Excel. The x value are values you know, like concentration values of the calibrators. You can then plot your measured values as y values. This can be absorbance, light units or whatever measurement parameter you have.

Curve fitting can be used to fit the line of best fit to these points. The line of best fit can be a straight line with function y= A + (B*x) . Alternatively non-linear functions can be used which follow the same method. From the straight line drawn, there is a distance of each actual point to the line. Since we know that our x-values are accurate, the distance of the data point to the line is only caused theoretically be inaccuracy of measurement of our y values. We can then calculate the distance of each y point to the fitted line. This is called a residual. The line of best fit is the line with minimal distance to the measured points. This is a line in which the total distance of the line to the different points is minimal. In practice we use the square of the residuals and add these up. This is called the sum of squares. Since we are using the minimal of the sum of squares to predict the line of best fit this method is called the least squares method.

The least squares method uses the distance from the data points to the line of best fit