I think the init limitis a reasonable suggestion, though I don't expect it'll have
that much impact.
As for tag sizes, I've historically been against lowering them, for most of the reasons Butcher listed. However, the increased political flexibility that would result from splitting up the player base into more and smaller pieces may be worth the downsides. Proceed with extreme caution, but I think a gradual reduction in tag size might be worth investigating. (And no, I'm not interested in rehashing that discussion for the umpteenth time.)
Another option might be to move away from fixed round lengths. Right now, if you want to win at tick 1177, the best way to do that is to be winning at tick 1000, and the best way to do that is to be winning at tick 800, and the best way to do that is to be winning at tick 600. This (along with a host of other factors that are not relevant here) encourages alliances to forge round-long blocks against their opponents for quick easy victories, and then honouring those deals even when the political landscape changes. Make no mistake, despite what AD posters would have you believe (when it suits them, anyway), honour is a
bad thing in PA politics. The inevitable aftermath of inflexible politics is early stagnation, and that's a loss/loss situation for everyone involved. 2 weeks of sitting on your hands is boring even if you're winning. In such a scenario, ending the round early would be preferable.
And on the other side of the coin, a round that's still undecided at tick 1150 might well be fun for another week, or even two. Though right now we don't feel the pain of a fun round ending too early as viscerally as we do that a bad one ending too late, I think it's just as big a problem. Variable round lengths would fix both in one fell swoop.
There are 2 big caveats that have to be addressed.
One, the decision when any individual round ends must be taken by algorithm, not by a human. The last thing we need is for wins to be questioned (rather than the methods of achieving wins) and for PA Team's impartiality to be cast in doubt. It must be clear
in advance what the conditions for ending the round are, and players must be informed of the progress of the round ending as it goes on ("alliance X is Y score/ticks/votes away from winning!"). A approach might be that the round ends when #1 alliance is X% ahead of its nearest competitor (possibly not counting alliances it's napped to?), or when a majority of alliances (weighed by size?) chooses a victor (regardless of the victor's score), or when an alliance has spent a certain number of ticks at #1 (consecutively? or not?), or whatever crazy scheme(s) you come up with. As long as the win conditions (there could be multiple,
4X-style) are determined pre-round, they're viable.
Two, PA's monetization scheme needs to be adjusted. The system of using a slowly accruing bonus at a certain point in the round depends entirely on the round being of predetermined length. If round length is no longer fixed, then the bonus needs to be adjusted, probably to a continuous one. This takes away a significant strategic choice, but I feel that's an acceptable loss if that means politics becomes more dynamic. Alternate solutions to the bonus/revenue problem would obviously also be welcome.