Leader Personality and Flavors Experiment

[gunnergoz]

@glider1:
Thanks for the information and the file. I will try it out in my next game.

Regarding the issue of giving settlers some defense points, I posed the question in another forum thread and got this reply from Spatzimaus:

There are four things you need to keep in mind.

1> Giving a settler unit ANY combat ability, even if it's purely defensive, will prevent that unit from stacking with combat units (as it'd now be considered a combat unit in its own right). This can easily make the unit LESS survivable in the long run.

2> The AI's behavior requires settlers to be escorted. If you're trying to make a unit that doesn't need escorting, then the AI can't adapt well to this change, because it doesn't have a behavior set for unescorted settlers.

3> If you're trying to have multiple concurrent settler units (and not have them form an upgrade chain) then a bunch of things can get screwed up, like the free settler from a certain policy. You can get around this by giving your new unit a resource cost or other requirement, which automatically prevents the AI from seeing it as a "primary" unit.

4> Settler costs are hard-coded. They're either set to a specific value, like a normal unit, or set to a dynamic value calculated by the AI based on the size of the city to be founded and how many free buildings will be given. This latter system is used for the core game's Settlers, and will change drastically depending on your starting era. So you can't make a second settler that costs exactly 100 units more than the original settler unit, because the game doesn't work that way.

I speak about these topics from experience. In my Ascension mod, I'd added the Colony Pod as a settler unit unlocking in the modern age, to be used mainly for games starting in the future eras. It had a combat ability (strength 50), but purely defensive (couldn't attack), had much better mobility than a normal Settler (movement of 3, and all terrain cost 1 MP) and could paradrop. To keep the AI from seeing it as a replacement for the Settler, I gave it a resource cost (1 oil and 1 aluminum). In the end, it actually worked very well, since in those future eras most units had high mobility and so an escort was still needed even for colonists that could defend themselves.

So there is the reason why we cannot alter settler defense values at this time. Pity.

 

[glider1]

Sorry for any ignorance in this question but the results seem very interesting. Is the end result to make this a modmod on top of GEM or to eventually to become part of the core GEM project?

At one stage under the VEM mod for vanilla, we were seeing big improvements in the AI's overall performance with a few of these changes. I was thinking of lobbying for a "strategic AI option" in VEM that would be kinda like an "aggressive AI option" we had in Civ4.

In GnK, we start all over again, because the AI now knows how to build and use a navy which on one hand means that it is overall more competent, but is now more difficult to "tune". In vanilla, we simply played on pangaea+ and turned off naval flavor and the AI became really challenging because of VEM's improvements.

As for the settler issue, thanks GunnerGoz and that's fine because the barbarian mini-game hurts the AI more than it hurts humans and personally I find it a bit tedious. It is wonderful to play civ like a game of chess. In other words you can move your settlers unescorted to the location you like, but it won't help you because the AI does the same and will go down on you hard if you make bad decisions.

Unfortunately, because of the 1UPT rule, you actually have to have the barbarian mini-game in place (unless you don't exploit it) because otherwise players will just exploit the AI by building settlers and placing them at their best location and render the AI unable to settle on that spot because of 1UPT. That's fixable but would need relaxation of 1UPT to allow non-military units of different civs not at war to sit on the same tile.

Ok, latest news on flattening out the leaders personality and flavours to midrange. Here is the hypothetical problem and why it is actually not a problem:

Theoretical problem: If you set all leaders the same, the game evolves according to "differences" that are intrinsic to start position and variations in traits. Therefore what happens is that one civ WILL become stronger than another and so a pattern emerges. It is a classic case of survival of the fittest. The strong cooperate to eat the weak. Then the strong themselves become weak and the remaining strong then eat the weak until you are left with A verses B in the endgame. How do you break this cycle?

Solution: You don't have to because it sorts itself out because there is a random element. In other words, what I am seeing is that all civs are competitive for territory now. One civ compresses up and tries to grow tall instead. Now you have the situation that the strong try to "eat" it cooperatively. However you cannot actually know which civ is going to eat the weak because there is a random element. Therefore if by chance one civ consumes the weak civ but it was not the strongest theoretically, then the geopolitics completely upset themselves and you get an imbalance in strength, very much like a wheel on a car that has lost the lead shot in the rim to stop it from vibrating at high speeds.

So even with all leaders = 6, you still get variation. The question is how much variation in behavior. That is still unanswered.

Cheers