I've done a bit more re-reading, this time of the domain rules from the old "BECMI" D&D (the Frank Mentzer "Basic" edition) Companion Set. Since my reading of Birthright recently, I've been thinking about the Companion rules in contrast to them. It's quite a contrast: BECMI's dominion rules are much more bare-bones and abstracted, and much simpler to use. One clever twist is that the core reward mechanism from the adventuring game - each gold piece looted earns one experience point - is extended to the domain game, which makes rulership more meaningful for name-level characters.
One thing I had in mind when reading this system was a complaint I'd read on a forum a while back that the system is a "money pump" that churns out profit (and therefore XP) too fast and too easy. It's true that an uneventful domain turn produces income well in excess of expenditures and a comfortable population expansion (and therefore more income the next turn), but other factors can have very steep costs. Among these are hosting tournaments, which have quite detailed rules, and random events like natural disasters or economic unrest... which have no rules. Literally none at all. There's a chart giving the chances of various specific events (mostly bad ones), but the game effects of each are left entirely to DM adjudication. This is a non-trivial absence, since the rest of the rules system is simple and abstracted enough that it's hard to come up with ways that make for a distinctive difference between the consequences of, for instance, a fire and of a hurricane.
On the other hand - this makes sense in context. Birthright, in contrast, has detailed rules are meant for a Birthright campaign, and so they're a lot like empire-building boardgames and videogames in their "gameyness" - the rules themselves are meant to present challenges in gameplay. BECMI's dominion rules are meant for a game in which rulership is only one part of the campaign, and not the focus of the actual game. So, instead of the game mechanics containing embedded, "hard-coded" challenges of gameplay, they're a springboard for DM-driven challenges that let the player bask in his successful rule... until [random catastrophe] happens and the DM uses this as a hook to throw a problem at the player to solve. In other words, these rules aren't "The game", they are the foundation of "The game". There are some old-school-isms in design too - for instance a wise player will keep back some the profits from the domain when things are going well in an emergency fund to fall back on during a crisis, which is a test of player smarts.
The mass combat system, War Machine, looks very nice. I haven't seen how the system works out in actual play, and it's hard to tell just by looking at it whether the results make sense. It does look like the system is a little more complex than maybe it needs to be, because the effectiveness of a force is worked out in two separate but linked stages. Still, I really like how it requires no miniatures or any other props, requires very few dice rolls, and the resolution chart gives a variety of consequences on multiple parameters – each line on the chart includes winner casualties, loser casualties, won or lost ground, etc. In fact, in these days where charts and tables are badly unfashionable among most (non-OSR) gamers, War Machine is a great example of a mechanic that just couldn't be done as well without a chart. For instance, as you look at higher margins of success by the winning side, generally the winner loses less and the loser loses more. There are blips, though – the only result which has the loser suffering 100% casualties is not the highest possible success and also causes fairly high casualties even on the winning side, which invites DM narration of a brutal bloodbath. You could – maybe – with a set of rules and dice alone manage to link winner and loser consequences together with success level, but I bet it would be much more complicated and time-consuming than just looking up the damn chart!
BECMI's domain rules have a lot of good things going for them – they're quite simple, have some good flavour, and are built for BECMI so easy to use in Labyrinth Lord. And yet... in some ways they're so simply abstracted that I feel like I might be as well just skip them and have the player roll for income per peasant and skip straight to the rolls for random events, which is where the interesting gameable stuff happens. Except, of course, that the random events give no DM support whatsoever beyond saying "[thing] happens", so I just don't feel like the rules system is doing enough for me to justify its existence.
Following the Domain theme of my reading so far, I have AEG's "Empire" for 3e/d20 and "An Echo, Resounding" for Labyrinth Lord still to read.