harmonic42 harmonic42

[Balance] Player vs AI power progression in FE (and all 4x games in general)

▼

[Balance] Player vs AI power progression in FE (and all 4x games in general)

from Elemental Forums

Hello.

I'd like to bring up a balance issue that I've experienced with every 4x game in existence, other than Civ4 BTS. It is actually extremely apparent here in Fallen Enchantress, perhaps moreso than any other 4x game. I love Stardock to death, but GalCiv2 is another big culprit. If you the player are looking for a challenge, even a moderate one, you'll pick the difficulty where the AI is at its "smartest" and at the very least, the AI gets no economic penalties, or even gets "cheater" bonuses. What happens?

 

1) Players start weak, while AI starts strong.

2) AI seems to thrive based on small formulas, small scope, and has either economic advantages or perfect algorithms for early expansion.

3) On higher difficulty levels, early game is by far the most dangerous for the player. You can be overwhelmed by a warlike AI (cough, Yithril with its juggs) or in GalCiv2's case, Drengin with its dominator fleets.

4) Due to diplomatic systems, it is almost impossible to broker a peace deal with these early, strong, aggressive AI factions. If you're going to die, or give up, this is the time. The time between 4 and 5 is the most fun you have, because you transition from desperation to being in a position of power.

5) However, if you survive, you make it to mid-game. I define mid-game as the point when your faction has hit critical mass. It now has the infrastructure to pursue all pursuits, technologically, militarily, economically. It is the point when you've basically already won, because...

6) The AI gets worse and worse as the game goes on. The more formulas the AI has to figure out based on changing circumstances and a sentient, intelligent human being adaptive, the worse it performs.

7) Since the AI is indeed just an AI, it doesn't have emotions or instincts, it cannot determine that the status quo is not working. It cannot determine that the human player has already won, and is on an exponential growth cycle toward a crushing victory, unless it gets bored and quits.

8) Late game. The human player is overwhelmingly powerful, invincible in fact. There is still room to grow, but why bother? Nothing will pose a challenge anymore. You can either end the game now, or if you're like me, and like EPIC games with epic challenges, try to help one enemy AI faction to the point where they are huge, and have lots and lots of nice juicy targets for your awesome army to kill. I mean, you spent all this time polishing, perfecting, building, training, and you have a huge army. Or god-like champion/sovereign. And you need a test. But SINCE you have those things, there is no test. The test is getting those things, which is rather anti-climactic.

9) You get your nice victory message and celebrate if it's your first couple games. If it's not your first couple games, you sigh.

 

Thus, virtually speaking, what we typically define as "late game" is really more like "post-game." Early-to-mid game is where the game is won, bar-none. Has anyone in the history of 4X games made it PAST that mid-game critical mass point, to find that there was actually additional challenges afterwards?

I have experienced this in Civ4 BTS. Late game AI knows how to explode economically, and build obscenely an powerful military ,and use it on you where it hurts most.

Is this an inherent problem with AI? With 4X games? Is this a Stardock weakness? Does anybody else experience this game flow?

38,134 views 27 replies
Reply #26 from Elemental Forums Top

Quoting MarvinKosh, reply 25
Okay here's a theoretical situation, let's say you had two equally-skilled human players playing this game, do you think that the player who suffers the first defeat at the hands of the other could pull it back and win?
End of MarvinKosh's quote

It depends on the variance and the swinginess.

Variance meaning the randomness inherent in the game, both in terms of the actual RNG and in terms of different outcomes happening despite the same set of actions. If the variance of a game is high and the defeat is narrow, the initial defeat might not matter much.

Swinginess meaning how quickly the game can turn around. Sometimes, this refers to variance being so high that things just happen. More frequently, it might mean one player unlocked a certain game-changing tech or found some game changing equipment/champions/monsters that's enough to tip the scales in the other direction.

I do believe that if one player manages to pull far enough ahead to overcome the variance and the swinginess, then yes, the other player can't possibly come back.

Of course, I don't have enough experience in FE to know what that point is or how easy it is to reach that point. Also, since FE doesn't currently support multiplayer, I don't think we'll ever be able to test this.

Reply #27 from Elemental Forums Top

Of course, on the faction power graph it's usually easy to see where the eventual leader started to win.  It's where their faction power begins to rocket up really quickly.  The reason why this happens is because they've cleared enough land to settle more cities, they have enough gildar to rush buildings, they've conquered cities and the unrest from occupation has finally gone down the a level that makes them useful as unit-producers or research-producers to the expanding civ.

This is why, if the AI is losing a war, it becomes all the more important for them to protect the cities they have left (which hopefully they will now in the 1.1 update) but also to destroy the ones that they can't protect.

Razing your own cities is admittedly a strategy of desperation because it can reduce your ability to support units, but unless you have the units to defend it, or it is safely sealed away by a chokepoint fortress, it's just going to fall to the enemy anyway, which will eventually make them even stronger, unless they decide to raze it.  If you manage to push the enemy back, you could rebuild later.

Galactic Civilizations IV