example
Day 0
5.8
Day 14
6.1
+0.3
Two weeks is a convention
Nothing in the scoring scale says "14". It is a chosen rhythm: short enough to keep track, long enough for a real change to have time to show.
Same protocol, or nothing compares
Same distance, same light if possible, face straight on. Measurements relative to face width vary by 8 to 11% with the distance of the photo. Change the distance, and you are measuring your arm.
Two ratings, two references
Each rating is frozen on its date, and compared with the population of that day. If the population changed too much between your two scans, the progress screen refuses to subtract the two ratings rather than show fake progress.
Look at each criterion
The progress screen shows more than the rating. It is criterion by criterion that you see what moved, and what did not.