The font scaler will always add two “phantom points” to the end of every outline. If the entire set of contours for a glyph requires “n” points (i.e., contour points numbered from 0 to n-1), then the scaler will add points n and n+1. These points will be placed on the character baseline. Point “n” will appear at the character origin, while “n+1” will be placed at the advance width point.
Both points (n and n+1) may be controlled by TrueType instructions, with corresponding effects on the sidebearings and advance width of the instructed glyph. The side bearings and advance width that are computed using these phantom points are called the device-specific widths (since they reflect the results of grid fitting the width along with the glyph to the characteristics of the device). The device-specific widths can be different from or identical to the linearly scaled widths (obtained by simple scaling operations), depending on the instructions applied to the phantom points.
This is just a guess. It looks as though the last two points are provided as a quick look up to work out the width of each character. My guess is that the x, y values are reversed for the advance width point.