The main problem I see with this is that its way to centered on Rome alone. I don't think that two campaigns would be very different from each other given how its the same progress of the game and very little oppertunity to change the course of history in a new way. And therefore that once you've finished one campaign, well, you've probably seen pretty much every thing the game has to show.
It doesn't necessarily even have to be entirely roman-centric. There are actually quite a few empires and kingdoms in that time period that could be playable (after all, it is around an 800 year timeframe). Many of the core mechanics can be modified slightly for other nations and tribal groups. Some (such as the "noble" system) would need almost no tweaking outside of localisation for different cultures. For example, I would include Germania and Brittania and populate them with historical tribal nations. Tribal nations would have no "cities" to speak of (eg, all their pops would be rural pops) and there are often civil wars and inter-tribal rivalries. There could be a list of historical tribal names and other, similar-sounding ahistoric names to allow tribes to randomly generate in a region occasionally.
Frankly I find the whole "compose provinces of tiles" approach to be much more historically accurate (eg province borders can and have historically been redrawn, and empires can fight skirmishes over valuable border tiles).
Honestly, the tile approach has the potential to be much more efficient so long as Paradox doesn't go and do something stupid (like evaluate events against EVERY tile). Due to this increased efficiency, the following regions could probably be included without raising CPU requirements above 2.0 GHz or so:
Italia
Gaullia
Hispania
Illyria
Greece
Middle East + Ukraine/southern Russia
Chinese nations (possibly excluding Tibet)
India (at least northern India)
Germania
I'm not sure about Indochina, as I am not too familiar with the history of the region. Surely it wasn't a major player in the early part of this time period though, so it could probably be cut, or saved for an expansion.
In honesty, I personally would prefer a game that was Roman-centric, but apparently other people don't agree -- which is fine. These are just ideas anyway, I hope Paradox picks and chooses which ones fit and creates a unique game.
Basically, what I DON'T want:
EU4: Antiquity -- Many of the EU4 mechanics would feel weird here, and Europa Universalis has absolute s*** mechanics when it comes to internal politics. A simple provincial factions system (like in CK2, but with provinces instead of lords) would go a long way to spicing EU4 up, but I digress. Even with a provincial factions system, much of the game would be spent blobbing, and after a point this becomes woefully ahistoric in an antiquity game. For example, there is no feasible way Rome could have conquered and held Persia, or it would have. Similarly, the Han would also be hard pressed to achieve such a feat, and Persia would also fond troubles subjugating either power. As such most of the game should be based on external political power balancing and internal politics. In a Europa Universalis context this would feel pretty detached and apersonal, like we experienced in Rome 1.
CK: Antiquity -- Don't get me wrong, Crusader Kings is definitely a good game, but for most of the recommended time period it would also not feel right in context. During the Roman Empire, it was not only "nobles" who gained power. If I remember correctly, a number of relatively low-born generals came to the imperial throne throughout this time period.
What I do want:
CK2 character-based internal politics, the ability to play a relatively low-ranking military commander or politician and work your way up, and basically a unique RP experience combined with the external politics (guarantees, warnings, "vassalages," alliances and coalitions) of EU4 on the national level.
Whatever RP system PDX decides on should have the option to be meritocratic. For example, I want the player to, on their characters' death, determine whether to play as the closest relative of said character, pick from a list of current characters, or create a random new character in a position of low to medium authority to start the process anew.
Also, I would like a minimum of Persia and northern India with preferably China playable (though China should be the first to go in event of major cpu-related performance problems). I also want a tile/city based approach for the game rather than the stale, out of place in context province system.