To the person who has no issues because they've constantly won, congrats you don't see a problem because you aren't looking. Where's the dog saying it's fine in a burning room meme when you need it?
Speaking to the topic of well-fare, we see it all the time in different forms. Whether it's the starter quests that now give newer players a leg up or the fact that newer players will never gear 10 or 11 Lando because he's no longer the king of AoE. So giving more rewards to a narrow 4 million GP over wider 4 million GP seems a little petty and over complicated at the end of the day.
But to the original poster's thought, it's a great idea but has a few major flaws. Some pointed out earlier. Mods and Meta's play a huge role in determining the chance one player has against another. Without taking into account both of these things in a matchmaking system you'll still end up back at square one.
A player who plays the mod game every day will run circles around players who don't. It's a painfully RNG convoluted system, but it's super powerful when played diligently.
A player who plays for Fleets is always going to stick out regardless of fleets being present. Maxing out Imperial Super Commando just isn't as great for a ground combat only GA as it is for a GA that features Fleets, and even now... well nothing matters if you have the Falcon on deck.
Lastly, Meta matters. A g12 Lando is in no way equal to a g12 Revan... not in GP and not in terms of viability in squads. And at one point he was in the metas, so a player who played during them is going to suffer that bloat for quite some time even if you cut the GP measurement down to a smaller section of that roster.
I don't know there is a great equation for as complex as this game has become, it'd be nice to see though. Especially when people want this sort of game mode where one side gets the high of being the TW king and the other half get the experience of setting defenses for GW because they have no shot at winning and if they don't set something the other kid throws a tantrum about getting a free win not being good enough. So I'm hopeful maybe we can get to an equation that handles this nightmare better, or at least makes GA less of something I want to turn off.