I like the new matching mechanism better than the old one as it provides more challenging setups. However, I feel it has its own shortcomings that could be fairly easily fixed.
The old system cast the net way too wide by assuming that every character counts in GA. In reality all those g8 toons will never do anything worthwhile in GA. Therefore I fully support the decision to ignore those going forward.
The new system however seems to assume that there is only
value in the absolute top toons. Meaning if I have eg 8 defense slots to fill only my top 80 toons are useful. And that it does not matter if someone has 50 more G11/12 toons after that or none. This ignores that those that have developed fleet have farmed toons as pilots that do little in standard GA. Also to have that many more GA viable toons provides the flexibility to field precise counters or vary your defensive setup. Therefore there is also considerable value in those toons over and above the current threshold. The further away we get from the top toons the less significant the additional value gets.
My suggestion would be to do as you do now counting GP from the top. But I would suggest to increase the cut-off point a little bit to include more toons. The formula of toons considered could be something like:toons counted=number of defense slots*5*2*(1+adjustment factor).
if you choose the adjustment factor to be 20% then you would count the top 96 toons instead of only 80 to take into account some of the still very useful toons outside of the top 80 while still ignoring the useless fluff at the bottom.
You might want to consider having a higher adjustment factor for the lower divisions as there might be more use for underdeveloped toons than at the top. The absolute number however would be smaller with a stable adjustment factor. Or you simply put a flat number like +15 as adjustment in there for all divisions.
(well my real suggestion would be to dispose of GP as a measure of how good/useful a toons is but I doubt that will fly