[ThaddeusMaximus]

Originally Posted by Zebra_Fact_Man

Based on nothing else other than helping rookies before, and having MANY rookie students go through my teams, chances are your problem is due to motor optimal rotation direction. All motors have a preferred way to rotate, and when rotating the opposite direction they rotate slower. The math isn't so simple, but it averages out to about 85% speed of the optimal direction. When you build a robot, normally motors on one side spin optimally, and the other side spins in the opposite direction, non-optimally, causing you to veer off in one direction consistently.

A few ways to fix this:
a) make all motors spin in the same direction. Not so easy if using the KOP drive train, direct drive a wheel to the gearbox axle, or like symmetrical chassis designs.
b) use some sort of autocorrect system, like a gyro or encoders.
c) quick and dirty fix - reduce the optimal speed side limit to 85% max speed in the code. You can play around with the number until you are happy and the robot is driving straight.

I have never seen this phenomenon play out with CIM motors in an FRC application.