[about Rome's 80BC start being weak]
as an opponent for the human player or vs AI as well? they seem to survive fine when I run 100ad test starts.
[...]
I think the 1st part, Rome conquering Persia, is rare. the way the system is set up now its impossible to collapse completely before your scripted "fall" date unless you lose your core completely. that could easily happen to Rome by 420ad (their fall date is 410). when I ran a bunch of Visigoth starts (420ad) Rome was hanging in there in almost every one.
vs AI, as I was testing loading 0-100 AD test starts from 80 BC. It is tied to the Balkan barbarian problem I mentioned, most probably. From 80 BC, they seemed more prone to crisis, perhaps because from 320 BC their empire grows "naturally".
Of course, it follows that against a human opponent the start is also fragile. They don't really have anything to counterattack you with if you rush into Anatolia or Greece.
Regarding Sassanid-Roman warring, while usually the Sassanids hold the line, or even conquer some Roman territory, I've seen a few instances of Romans getting lucky and taking the Sassanid capital, which often leads to terminal collapse of Sassanids. If anyone should collapse the Sassanids, it should be the East Romans rather than the West Romans. I think walls in the Sassanid core cities would be all that is necessary to prevent Romans from sneaking in a core capture.
---
Also, loaded a Japanese game from 220 AD, opened WB and found that Epidamnos is an independent city in the middle of the sea of Roman red!
Edit:
Also of note: West Roman cities in Mesopotamia and Armenia do not flip to East Rome at their spawn.
Solution: West Roman cities East of the East Roman spawn should become independent at East Roman spawn.