Arathorn said:
I agree on this point. I think no one particularly likes that situation. I've found moving up in difficulty levels removes this problem. That and a ruthless crushing of opponents to get victory quickly.
Sometimes I feel it resolves the problem, but not always. There have been many games where I have struggled through the ancient era, managed to catch up to the AI by the end of the middle ages, then took the lead using stacks of cavalry. The early industrial age is fun, while I'm cementing my lead, but the late industrial era and the modern era get boring: I find that once I'm in the lead I seldom have anything to fear from the AI. And yet, if I go up a difficulty level (to try to get more challenge late in the game) then I'll never make it past the ancient era.
The problem is that the AI seldom capitalizes as much on the snowball effect as a human player can. I find that I catch up from behind
much more often than I fall behind once I'm ahead. Changing the difficulty level really only changes how far behind I am in the early game and how long it takes me to catch up (if I can). It doesn't really change the fact that once I've taken any significant lead the outcome isn't really in doubt anymore.
They could always increase the number of difficulty levels, to make it easier to find a level that was just right: you were behind for most of the game and could catch up in the modern era. Such a game would be challenging and interesting the whole way through. But that wouldn't address the basic problem that the game is only challenging when you're behind or about even with the AI. It would be nice to be able to take a lead and still feel some element of challenge... still feel the threat that an AI civ might catch up and overtake you if you weren't careful.
Splitting as an anti-snowball mechanism? Hmmm... Has a lot of problems, not the least of which is that lots of people like big empires and have no fun if they can't be big.
It does have lots of problems, but then, corruption and waste (the current main anti-snowball mechanism) is one of the most complained about aspects of the game, so it seems to me its worth discussing other options.
If it's possible to build a large empire safely, all the good human players will do that. The AI will either have to do it that way or will be far too dumb to be a useful opponent. Plus, you then lose all anti-snowballing effects, don't you? And aren't you back to the problem you noted in the opening sentence -- a large stable empire?
On a high enough difficulty level, a human player might not have the option to build a large empire safely. If you're falling behind the AI, you'd have to make a decision: play it safe with respect to "splitting" and risk falling further behind, or take a risk of "splitting" in order to catch up to the AI. And if you were even with the AI or a little ahead, you'd have to decide whether to try to get further ahead (but risk splitting) or play it safe, even though that made it harder to pull further ahead.
And that, of course, is the whole point of "anti-snowball" - decrease the ease by which you pull further and further ahead once you're ahead.
And of course, when you say "the AI will have to do it that way or be far to dumb an opponent," you could consider that there are multiple AI opponents that might be programmed with varying levels of tolerance for splitting-risk. If the AI next to you expands faster than is prudent, they might be "dumb" in a global sense, but if the wipe you out before succumbing to a split, then you're not around anymore to laugh at their misfortune.
Splitting could make the field of opponents arrayed against you much more dynamic and interesting. One AI civ might expand slowly and cautiously, with little or no risk of a split. Another might expand more recklessly, becoming the dominant power in an early age but then falling victim to a split (or several) and fading to relative meagerness. In other words, not only would such a mechanism ruffle up the balance of power between player and AI, but it would also affect the balance between AI and AI, and there's all sorts of ways in which that could make the game more interesting.
Plus, book-keeping to keep yourself "safe" (and with no details, I have to assume that the algorithm will become known and exploited) sounds like the opposite of fun to me.
This is a valid point and an important issue to consider. My personal feeling is that there should not be any specific level that becomes "safe" - just graduated amounts of risk. When you attack a pikeman, you take on a certain amount of risk: how much depends on which unit you use. If you use a warrior, its a longshot; if you use cavalry, the odds are in your favor. But even if you use Modern Armor, there's always some tiny chance that you'll lose or at least be wounded. No amount of book-keeping with respect to unit strengths can make you completely "safe," and no one considers it an exploit if you can predict that cavalry is more likely to win than a warrior. In other words, predicting the relative risk of "splitting" should be simple, and there shouldn't be any absolutes that encourage one specific play style.
@mojotronica -- carrot and stick. How will the stick be implemented? Is it a dice roll every turn? One frickin' roll of the dice can remove almost a large chunk of my empire (say I have a capital and one "safe" Forbidden Palace plus a Provincial Palace)? Or would it be just one city where I can put down the revolt so easily it's a nuisance like the current whack-a-mole pollution? YUCK! Or is it controllable? Well, then, you just remove the chance for revolt before you build the PP. Sounds incredibly painful either way -- or any other way I can try to imagine.
I can't answer for mojotronica, but here's my take on the matter:
Such events should not be so rare and powerful that they devastate you when they occur, nor should they be so common and minor that they're a frequent nuisance (like pollution). They should be rare (but still possible) when you play conservatively, and somewhat more common if you expand very aggressively. They should not be so powerful that no one ever wants to play aggressively, but should be powerful enough that there's also reason to consider expanding more slowly. I'm optimistic that a suitable middle-ground could be found. Perhaps something like a few border cities splitting off together (
not half your empire, but also more than a single city), and it would occur 0-1 times per game if you played it safe and 2-3 times per game if you took more risks and expanded less cautiously.