Something got me thinking about unit diversity earlier today. Now, years ago I wrote about it, but it's long gone, so let me write a bit more about it now.
Basically, in my mind, there are two approaches to distinguishing units from each other. Tactical and strategic.
A tactical change is when units have the same fundamental abilities but in different magnitudes. For example, a tank and infantry can both move, attack, be hit, be seen by the enemy, cost money to build, etc. A tank probably moves further, attacks harder, takes more damage, costs more, and so forth. These are just grades of abilities.
A strategic change is when a unit has a fundamentally different ability. Not a different grade of ability, but something that changes how the battle works. For example, the infantry might be the only unit that can take buildings, or the tank might be limited only to roads.
Strategic changes can be relatively minor or extremely critical. If the mission requires you to take buildings, the infantry ability to take buildings is a keystone of your strategy. On the other hand, it's very easy to have a game where such an ability is just a curiosity.
Both changes require balancing of very different sorts, but the flavor of the game also needs balancing. Games with diverse units need to be careful to limit the density of the units, to keep things from getting overly difficult and complex.
There are a lot of ways to do this. One way is to limit what units are available/useful in any given scenario. This can be done by fiat if the GM has strategic control - just don't assign the units. However, it can also be made a function of the situation itself. For example, if there is a cloaking space ship that is only useful in a nebula, then the ship is not a useful strategic choice in any other situations. It's possible that the nebula ship is otherwise a relatively normal unit whose only significant, unique tactical consideration is whether to keep it safe for future missions that might be in nebula.
Another method to reduce complexity is to have some units that vary tactically but have the same strategic capacities, and other units that have very similar tactical considerations but very different strategic capabilities. This is not suitable for all games, though, as it produces a very specific kind of feel.
Aside from those considerations, balancing them can be quite difficult and is beyond the scope of this essay. However, there are some basic tactical and strategic variations that are common, and I might as well list a few of them.
Some tactical variations: damage, range, area of effect, percent hit, side effects of strike, armor, health, dodge capability, speed, terrain passable, fuel/mana, scan range, visibility, secondary attacks or special abilities that have tactical value such as grenades or lock-on.
Some strategic variations: long-range scanning, repair/medical, capture of units/objects, resurrection, science or intelligence, leadership, flight if nothing else can fly, construction of roads, units, or buildings.
There are some variations that kind of fall between: snipers, missile barrages, sprint, entrenchment, regeneration, tracking, assassination, hiding, etc.
What abilities can you think of?