So if everyone is "useful" (whatever this nebulous term means...I mean honestly even using aid another in 3.5 every round in combat is of some "use") on the battlefield, their usefulness outside the battlefield doesn't matter. Well both are useful just at taking on two different types of enemies.
IMO, what you're talking about is not game balance it is specifically combat balance and impossible to nail down unless usefulness is defined.
How bout this...
Lets say You have Three Colors (
Red,
Yellow,
Blue) and I have Three Colors (
Green,
Orange,
White). Your colors represent PC classes, mine represent typical D&D encounters (combat & non-combat).
Now, lets set up a simple RPG. Yellow defeats Green, ties Orange, and fails against White. Red Defeats Orange, Ties White, Fails against Green, and Blue defeats White, Ties Green, and is defeated by Orange.
In theory, all colors are balanced. They defeat one, tie one, are defeated by one. However, this assumes that the game, over the course of the length of play, has equal parts Green, Orange, and White. In a game where (lets say) Green is 2x as common as White, than Yellow has a distinct advantage over Blue and Red, (esp Red). If the game features plenty of Green and little or no White, than Yellow is the most desirable, Blue is not a bad choice but will pale in comparison to Yellow, and Red is worthless in all but the occasional Orange situations and the once-in-a-lifetime White ones.
Of course, you switch Green and White around, and the whole thing inverts and Red is king, Yellow is useless, and Blue is stuck and the perpetual Brides-maid.
Would not a better solution to create a system where Each color has a 50/50 chance of defeating two colors, and a 90% chance of defeating the third?
For example, Yellow nearly always beats Green, but has a 50/50 chance of defeating Orange or White. Blue is the same way with White, and a 50% chance against Green and Orange. Red finishes it off with a near-perfect rate against Orange, but a 50/50 chance against White and Green.
Now, the game isn't so skewed. A Red stuck in a world of constant Green isn't stuck sucking, he has just as much of a chance as Blue. Yellow is still King, but Red and Blue still can handle themselves. Yellows "less than 100%" chance is also a key; he can fail and allow Red/Blue to occasionally shine.
More importantly, each class keeps it "niche" (defeats a certain color) without being shut out if their antithesis color comes up.
What's this got to do with D&D?
D&D (pre 4e) is pretty much R/Y/B if you replace them with Spellcasting/Combat/Skills. Orange/Green/White is Spell Resistance/Traps/Damage Reduction. It shutsdown one, ignores the other, and is trumped by the third. However, if your DM focuses on one element (say, lots of traps) than the skill PCs are more important, the spellcasters can mimic (or even supercede) the skill users, and the combat machines are hanging out waiting for ambushes. You can jigger the colors as you want to shut-down any or all of the PC colors.
A much better solution is what 4e tries: PCs excel at something (fighters are good combat, rogues at skills, wizards at magic) but nobody is locked out of the game for being the wrong "color". For someone who was familiar with high SR monsters, Trapfinding, or DR (for example) this is a shock and it can feel like the classes are the "same" but I assure you, being competent at most things (and better at one) is an all around better experience than being great at one, ok at the another, and utter FAIL at the third.