For new(ish) players who weren't around during the initial launch of LSTB, you may wish to read some of the threads that occurred then before reading this.
Galactic Power (GP) was introduced in September 2017, right before the debut of LSTB. At that point, its only purpose was to allow us to "use" our DS rosters, which were next to useless in LSTB. Many people chose to raise unused characters to G6/7 + Lvl 53 to "pad out" rosters, allowing them to gain more TP by deploying more every day.
Back then, people with top heavy rosters were penalised heavily, myself included. I was getting 113/114 CMWs in LSTB, yet ranking about #15 in my guild due to my abysmally low GP of 1.7M. GP was intended to reward older players with some old-meta characters half-geared and sitting in their roster, while penalising those who only focused on a few characters.
Then came TW in December. TW was, in this aspect, the inverse of TB, as those with strong rosters and lower GP would win much more handily in TW (while losing out on better rewards due to guild GP bracket). Here, those with top-heavy rosters shone; they scored much better than everyone else of comparable GP by the sheer number of max teams they could throw. Padding out GP became a liability, as one had useless half-geared characters lying around in the depths of their roster. At that point in time (early 2018), the jury was out on padding rosters; one helped their guild in TB, while hindering their guild somewhat in TW.
This all got thrown out of the window with the introduction of Grand Arena, as the matchmaking system became one that only compared GP instead of GP and G12-to-GP ratio like it should. Those with padded rosters immediately found themselves competing with the likes of myself, who (as a 2.7M) have yet to face anyone with either Traya or Revan despite owning both (and their respective teams) and using them at G12 +3.
This obviously becomes problematic, as those clamouring for a GP calculation overhaul fail to realise that this kittens over all those who padded GP early on, making them less effective in TB and adding insult to injury. Meanwhile, the current GA matchmaking isn't very viable either; I have only lost a single match, and by about 10 banners against someone with a comparable roster of NS, BH, JTR and the like.
My proposed solution is as follows:
1. Use something akin to swgoh.gg's Collection Score (CS), and name it "Roster Completeness". Make this score percentage from 0% to 100% (100% being people with a completely max roster). RC will then determine a player's "Overall Strength", by multiplying it with GP.
2. Continue to use GP in TB deployments.
3. Use a guild's mean Overall Strength as the first criteria for TW matchmaking, then GP as a secondary criteria. Follow the reverse for GA; try to match those with comparable Overall Strength first. This is because TW GP brackets still exist.
4. Remove the nonsense "Community TW initiative".
This will solve most of the gripes people have with GA and TW.