Last night I was thinking along this line: the problem with Ming shouldn't be its government type or religion. The actual problem was nobody could have ruled such a large empire with no local autonomy.
Consider a thought experiment: if you somehow changed the government type of Ming to a merchant republic or an archduke, it should only fail harder at governing all these land. Celestial empire should make sense in 1368, be a little outdated in 1444, and only totally suck after western nations get more modern government types. After all, it was the improvements in administration technology that finally makes a huge empire without autonomy possible.
In contrast, the game's modelling of LA is a bit off. Basically you get -0.1% LA per province per month, which implies a large empire can get rid of LA at the same speed as a two province minor.
A simple solution would be assigning two values to each government type. Say A and B.
A is the old LA reduction per province per month, but the values get heavily nerfed. Say -0.017% per month (200 years from 40% autonomy to 0%) in early game and -0.042% per month (80 years from 40% autonomy to 0%).
B is "total LA reduction over all provinces". If you have N provinces with positive LA, they gets B/N LA reduction per month. An early game government may give, say, -0.5% per month "total LA reduction". So if you have only 5 provinces with LA, you get as much LA reduction as before.
By separating A from B, we can differentiate the government types a bit more. Say merchant republics get more B, so when these tiny nations conquer one or two provinces, they get autonomy quickly reduced. They would also prefer conquering a few rich provinces instead of many poor provinces, which makes sense. A despotic monarchy can have no A but some B, since the monarch ruling over everything is probably not a scalable way to govern a monarchy. An empire, on the other hand, can have some A but no B, which is better suited to large nations.
Now let's look at Ming. It was formed in 1368 by Zhu Yuanzhang, a 6/4/6 ruler. Surely we should assume a ADM 6 ruler knows how to choose a government type. The only limitation to him should be the poor technology level he had. Chinese had almost no knowledge about theocracy or republics, so he needs to choose between Celestial Empire and Despotic Monarchy. In the A/B system proposed here, it should be obvious for him -- Despotic Monarchy had no per-province LA reduction, while Celestial Empire can have some built in (say, -0.017% per month). The total LA reduction didn't matter for a huge country like China.
In this way, we can make China not to its full potential at 1444, but can actually explain everything without hard limits...
Consider a thought experiment: if you somehow changed the government type of Ming to a merchant republic or an archduke, it should only fail harder at governing all these land. Celestial empire should make sense in 1368, be a little outdated in 1444, and only totally suck after western nations get more modern government types. After all, it was the improvements in administration technology that finally makes a huge empire without autonomy possible.
In contrast, the game's modelling of LA is a bit off. Basically you get -0.1% LA per province per month, which implies a large empire can get rid of LA at the same speed as a two province minor.
A simple solution would be assigning two values to each government type. Say A and B.
A is the old LA reduction per province per month, but the values get heavily nerfed. Say -0.017% per month (200 years from 40% autonomy to 0%) in early game and -0.042% per month (80 years from 40% autonomy to 0%).
B is "total LA reduction over all provinces". If you have N provinces with positive LA, they gets B/N LA reduction per month. An early game government may give, say, -0.5% per month "total LA reduction". So if you have only 5 provinces with LA, you get as much LA reduction as before.
By separating A from B, we can differentiate the government types a bit more. Say merchant republics get more B, so when these tiny nations conquer one or two provinces, they get autonomy quickly reduced. They would also prefer conquering a few rich provinces instead of many poor provinces, which makes sense. A despotic monarchy can have no A but some B, since the monarch ruling over everything is probably not a scalable way to govern a monarchy. An empire, on the other hand, can have some A but no B, which is better suited to large nations.
Now let's look at Ming. It was formed in 1368 by Zhu Yuanzhang, a 6/4/6 ruler. Surely we should assume a ADM 6 ruler knows how to choose a government type. The only limitation to him should be the poor technology level he had. Chinese had almost no knowledge about theocracy or republics, so he needs to choose between Celestial Empire and Despotic Monarchy. In the A/B system proposed here, it should be obvious for him -- Despotic Monarchy had no per-province LA reduction, while Celestial Empire can have some built in (say, -0.017% per month). The total LA reduction didn't matter for a huge country like China.
In this way, we can make China not to its full potential at 1444, but can actually explain everything without hard limits...