LS is correct. With perhaps one caveat (mentioned later for under-water break in).
First of all, the run in time is only intended to seat the brushes to the comm. Not wear them beyond the point where they have fully contoured to the shape of the revolving copper mass between them.
No offense to the long-term break-in guys, but running in a motor for more than minutes is only contributing to premature wear of the brushes and also the comm, regardless of the motor involved. Once the brushes have fully seated, there is no other part of the motor that needs to be "broken in". There is only ONE moving part with only TWO objects that ride against it. And if the comm is out of round, no amount of "break-in" will help since the brushes are softer than the comm and will not "make it round".
The reason for "seating" the brushes is to reduce (you cannot eliminate it) the arcing between brushes and comm that result in "pitting" of the comm surface. This occurs most prevalently when the brush faces are not contoured to the same curve as the comm.
This is best done by running the motor at slow RPM until the brushes adapt fully to the contour of the comm. I have yet to see any motor, including full-race strap motors, that need more than 30 minutes, even with high-silver brushes (yech!) to accomplish this. The brushes in a typical RTR motor, being of a soft compound, are fully seated in 10 minutes or less at low RPM.
I break-in motor brushes at the slowest speed possible. If the motor rotates at 3V then that's what I use. If it needs 5-6V, then fine. I oil the bearings/bushings before running the motor in, and again after - very sparingly.
I have used the under-water approach. The benefit to it is speed. Under water, the cavitation action literally erodes the brushes at a much faster rate. If minutes are good for dry break-in, then seconds are good for under water. But this is where I will use reverse voltage - just for the first few seconds (maybe 10-20), then the remainder of the time (maybe 30-60) with normal forward voltage. Why? Because this method can eat away at the trailing edge of the brush faster than anywhere else - that is where the cavitation is greatest. By reversing direction, the effect is minimized, though you still want the brushes seated in the normal rotation direction when finished.
With today's disposable motors, I only use dry break-in, and only for a few minutes. Anything more severe, and you are really doing nothing but wearing the motor out. Run it in for a few minutes, then drive it on track for a few slow laps, then have at it!
I also NEVER use "tuner cleaner" sprays to "blow out" a motor - the temperature change can distort the shape of comms or stacks, or even crack the insulation on windings if used to the extreme. Comm drops are okay, and recommended, especially on motors that can't be easily disassembled for manual cleaning.
JMO, as always.