"You must spread some Reputation around before giving it to noah.gleason again."
@Mark: 3620 has not done what Noah describes, but we have used current as a measure of excess drag torque in two ways:
1) we dial in alignment and belt tension in our drive trains by measuring motor current required to drive them with wheels-up; i.e., chassis on blocks. Any current beyond N*I_free, where N is the number of motors per gearbox, is excess drag. We never completely eliminate that excess, but we aim to get it below 1.5 Ampere. We also aim to get it balanced left vs. right, so the robot will be easier to drive on a straight path. (We are a skid steer team, up through 2017.)
2) we measure motor current required by scoring mechanisms to determine how much friction we are fighting, and use that information either to explore ways to reduce the friction, or to assess how much our speed-reduction ratios must be adjusted to prevent motor overheating and/or brown-outs.
Both of the above are performed during prototyping and again during final integration of the robot, using in-line ammeters between the motors and their controllers. We use the same procedures to diagnose problems that arise during practice or at competition.