The similarity of two unfolding profiles is defined by the area between the curves.

For this, the area between the curves is calculated by looking at the difference in the ratio signals of the two curves at each temperature. Then, with a little bit of statistics, the area is corrected for the noise in the data, and the resulting corrected area A_corr is used to calculate the similarity value as follows:

As a consequence of this type of calculation, a small shift in the brightness ratio (causing a shift of the profile along the y-axis) will result in a small change in similarity, whereas a shift of the T_i (causing a shift of the profile along the x-axis) will result in a comparably larger change in similarity.