I get the feeling I'm in a minority on this issue, that I see the death of any AI faction (in particular one I didn't decide to kill - which is rare indeed) as a failure. The game becomes just a little but less interesting for me if there are now fewer players on the stage, especially if it's one that was historically significant.

I was playing in a semi-historical 1.0 Romani game a long time ago, not paying attention to the AI factions outside my sphere of influence when an announcement popped up that Pahlava was no more. It really took the enthusiasm out of that game thinking that I'd never have the chance to face off against them in a century's time as I'd been hoping.

Since then I tend to tinker heavily with the AI factions (by console commands, Force Diplomacy and raids), with the general aim of slowing down changes in the ownership of provinces and protecting the smaller factions from being steamrollered by the larger.

Its one of the reasons I like seeding random areas with provinces belonging to factions who aren't based there. Galatia and Scorcouw as a way to preserve the Arverni and/or Aedui if things go badly in Gaul, for example. I tend to put serious obstacles in the path of AI rome (like continually boosting garrisons in Mediolanum, Segesta, Bononia and Patavium) early in the game to stop them conquering Gaul by 250BC.

Another real annoyance is the way many factions seems obsessed with expanding northwards. Hayasdan are the worst for this, seeming to want to become the masters of the steppe. Baktria and Makedonia do it too if you let them. I've already mentioned the Romans.

Does anyone else engage in that sort of thing? I tend to notice a lot of campaign maps with the AI run riot, and it is an extra layer of effort to the game. Though I find, particularly the raiding aspect adds something when you're not blitzing or otherwise expanding a lot.