Research agreements
- Thread starter joltis
- Start date
Seek
Deity
- Joined
- Aug 7, 2010
- Messages
- 3,773
It's really easy. For example:
Just copy/paste this into your mod and change the values to whatever you want.
Code:
<GameData>
<Eras>
<Update>
<Set ResearchAgreementCost="200"/>
<Where Type="ERA_ANCIENT"/>
</Update>
<Update>
<Set ResearchAgreementCost="300"/>
<Where Type="ERA_CLASSICAL"/>
</Update>
<Update>
<Set ResearchAgreementCost="450"/>
<Where Type="ERA_MEDIEVAL"/>
</Update>
<Update>
<Set ResearchAgreementCost="600"/>
<Where Type="ERA_RENAISSANCE"/>
</Update>
<Update>
<Set ResearchAgreementCost="800"/>
<Where Type="ERA_INDUSTRIAL"/>
</Update>
<Update>
<Set ResearchAgreementCost="1100"/>
<Where Type="ERA_MODERN"/>
</Update>
<Update>
<Set ResearchAgreementCost="1400"/>
<Where Type="ERA_FUTURE"/>
</Update>
</Eras>
</GameData>
Spatzimaus
Mad Scientist
Yes, it's been done in several mods. In my own the RAs cost almost exactly twice what they do in the core game.
There's one MAJOR balance point to consider though; the AI. Basically, the AI doesn't think more than one move ahead. It doesn't PLAN.
Right now, part of the reason the AI makes so many research agreements is that it hits the point where it has, say, 350 gold in its treasury with an income of maybe +20. It sees that it can buy a Research Agreement for 300. But what else could it use the money for? Bribing city-states (250/500/1000), sure, but the AI appears to be coded to not buy Influence unless it can have enough of a buffer to hold the bought bonus for many turns. (Not much point in buying 60 influence to get Ally only to watch it decay to Friend the next turn.) So effectively, there's a minimum of 500 there from the AI's point of view. It'd be seven or eight more turns before the AI has that much money, and that's seven or eight chances for it to say "okay, I'll buy an RA."
Likewise, it can rush units or buildings, but those generally cost significantly more than that 350 gold. So again, you're sitting on that pile of gold for several turns. The AI's probability-based, so it might randomly decide to save its money for a turn or two, but eventually it'll pull the trigger and buy it (assuming there's another AI out there with enough cash).
So the AI is sitting there, has this gold burning a hole in its pocket, and has no ability to plan ahead to realize that MAYBE it should save up for something else. So it buys the RA, almost whenever possible. If the civ has just ended a Golden Age it's even worse, as the income might be almost zero and it'd take forever to reach the threshold where there are other things you can do with the money.
But if you ramp up the RA costs too much, the problem inverts. The AI will never enter into an agreement, because once it has enough money in its treasury to rush a building or bribe a city-state, it'll do that instead. In some ways this is a good thing, but it becomes a major balance issue, especially on smaller maps (where the odds of two AIs hitting the threshold at the same time are so slim).
So really, what you need to do if you're adjusting RA costs is make it such that the cost of the Research Agreement is comparable to the other things the AI could be doing with the money during that era. No amount of Flavor tweaking can make up for this.
There's one MAJOR balance point to consider though; the AI. Basically, the AI doesn't think more than one move ahead. It doesn't PLAN.
Right now, part of the reason the AI makes so many research agreements is that it hits the point where it has, say, 350 gold in its treasury with an income of maybe +20. It sees that it can buy a Research Agreement for 300. But what else could it use the money for? Bribing city-states (250/500/1000), sure, but the AI appears to be coded to not buy Influence unless it can have enough of a buffer to hold the bought bonus for many turns. (Not much point in buying 60 influence to get Ally only to watch it decay to Friend the next turn.) So effectively, there's a minimum of 500 there from the AI's point of view. It'd be seven or eight more turns before the AI has that much money, and that's seven or eight chances for it to say "okay, I'll buy an RA."
Likewise, it can rush units or buildings, but those generally cost significantly more than that 350 gold. So again, you're sitting on that pile of gold for several turns. The AI's probability-based, so it might randomly decide to save its money for a turn or two, but eventually it'll pull the trigger and buy it (assuming there's another AI out there with enough cash).
So the AI is sitting there, has this gold burning a hole in its pocket, and has no ability to plan ahead to realize that MAYBE it should save up for something else. So it buys the RA, almost whenever possible. If the civ has just ended a Golden Age it's even worse, as the income might be almost zero and it'd take forever to reach the threshold where there are other things you can do with the money.
But if you ramp up the RA costs too much, the problem inverts. The AI will never enter into an agreement, because once it has enough money in its treasury to rush a building or bribe a city-state, it'll do that instead. In some ways this is a good thing, but it becomes a major balance issue, especially on smaller maps (where the odds of two AIs hitting the threshold at the same time are so slim).
So really, what you need to do if you're adjusting RA costs is make it such that the cost of the Research Agreement is comparable to the other things the AI could be doing with the money during that era. No amount of Flavor tweaking can make up for this.
Hi Spatzimaus,
Thanks for info. One thing to keep in mind is to how human reach to changes in RA cost. Too high is too high for human also. I have been trying to fiddle those money use values which are in defines AI_GOLD_PRIORITY find them. Minor civ stuff is elsewhere.
Thanks for info. One thing to keep in mind is to how human reach to changes in RA cost. Too high is too high for human also. I have been trying to fiddle those money use values which are in defines AI_GOLD_PRIORITY find them. Minor civ stuff is elsewhere.
Spatzimaus
Mad Scientist
Actually, no. That's what I was trying to explain.
A human will know that at his current income, he can save for X more turns to afford a research agreement, so he won't spend the money even if it could be useful right now for some building or unit he needs. Or he knows that he can sell off an extra luxury he has to gain the cash needed, if he sees an AI that he can make a deal with. Or he might see that he's close to starting a Golden Age, where his income will go through the roof. These are all commonplace methods of gaining money quickly, but the AI won't account for ANY of them. The decision on whether to buy will be made purely on the basis of how much cash he has that turn, and if there's anything else he can use that cash for right away.
The human player will save his money if he's getting close to being able to buy what he wants most. An AI player won't link cause and effect that way. Now, since all RAs are between two players, anything that prevents the AI from saving up enough will also prevent the human from doing the same. So there is some effect on the player, but it's not the same kind of absolute cap I've been describing.
