[Tristan Lall]

Shifting gearboxes are cheap this year, and the stupid cylinder restrictions are long gone. That means the set of circumstances where a shifter would not be beneficial is exceedingly small.

That said, the main disadvantage for a competently built shifting gearbox is the fact that it inherently spends some of its time out of gear before resuming under power. (CVTs don't have that issue, but they're quite rare in FRC.)

Think of it in terms of a diminishing return: for each ratio, you add complexity and unpowered time every time you accelerate and shift through the gears, but gain output efficiency (by running your motors more optimally). At some point, the losses of time equal the gains in acceleration. Consider the situations in which your robot will be expected to perform, and estimate whether you're running into the practical limits of the design's utility.

Also, don't underestimate issues like being in the wrong gear and running out of air, or synchronizing two shifters.

(For my own amusement, I enjoy having two speeds and six drive motors very, very much. Incidentally, it's a good combination for a lot robots in a lot of games.)