I actually like playing unbalanced games with a specific victory condition. I even wrote scenarios for other games like that. But it is a game and should be fun for both sides. Even a campaign.
A GM has to make a campaign interesting for as many of his players as possible. My players know that if they are looking for the "Easy" button for a reward in my campaign, it doesn't exist. Players in scout 'Mechs are given missions for scout 'Mechs, and not just act as fast, fragile targets on a battle map. If a scouting unit gets into a fight, he failed in his primary mission. To train people the value of scouting, and gathering intelligence on the enemy, scouts are sent out, usually alone, to determine the opponent's strength, numbers, location and movement, preferably without being seen, which may include some observation while out of the cockpit.
If the scenario is PvP (and mine always are, as I recruit other players to act as OPFOR in order to make it a true double blind, even on the strategic level) then the force that does not want its position and forces revealed to the opponent assigns screening elements to prevent scouts from gathering crucial data. Aggressive commanders will send out small teams to hunt down enemy scouting units in what I call "Stooging" (from the Three Stooges eye poking routine) in order to blind the enemy. Cavalry screens were used during the Napoleonic wars to prevent the enemy from locating the main body of forces, as well as keeping the enemy from gathering data.
I've had players as techs, commandos, even DS pilots, so I had to design scenarios suitable for them, as well. We had a player who had no interest in battlemechs at all, who commanded an all-vehicle unit.
Set, balanced scenarios have a place, but reserve them for training the new people who don't know how to play.