RFC: Reshaping the research penalty

[JediClemente]

Have you ever, having more than 10 cities, conquered another city (not always to keep it; maybe for enforcing capitulation and giving it back) and seen your research cost increase dramatically?

While a 3 or 4-cities Netherlands is 5 techs ahead of you, and increasing?

This is what I refer to as the "static" part of the research penalty system: it doesn't matter if some or a lot of your cities give out a poor research. A war of conquest, even not to keep the cities a few turns later, having more than 13 cities, is a nightmare.

Spoiler :

Looking at the actual code, let's make an example, because you may not notice completely during the game: it's worse for the human player than for the AI.

For 10 cities, you'd get your actual base cost for the tech (although other parameters affect it: namely the era, having certain techs already, or being in the last turns of the game).

For 11 cities: 8% extra cost for the human player, 5% for the AI.

For 13 cities: 24% extra cost for the human player, 15% for the AI.

For 18 cities: 64% extra cost for the human player, 40% for the AI.

As I remarked earlier, the problem isn't penalizing the higher number of cities, but doing so regardless of the nature of the cities; they may be unproductive.

And remember the more cities, the higher maintenance costs and civic upkeep = less money = less science.

My idea: the penalty should depend on your research rate: the higher it gets, the more the research cost should be increased.

To begin with, the penalty will be calculated for having a research rate higher than 5 times (iMargin parameter) the base research on the capital.

The modifier (iMultiplier) will be 5 for the AI and 8 for the human player, as before. This is the extra % cost for techs.

Hm?

int iNumCities = GET_PLAYER((PlayerTypes)getID()).getNumCities();
int iMultiplier = 5;
if (GET_PLAYER((PlayerTypes)getID()).isHuman())
	iMultiplier = 8;

int iMargin = 5;
int iResearchRate;
int iCapitalBaseResearch;

if (getID() < NUM_MAJOR_PLAYERS)
	if(iNumCities > 10){
		iCapitalBaseResearch = GET_PLAYER((PlayerTypes)getID()).getCapitalCity()->getBaseCommerceRate(COMMERCE_RESEARCH);
		iResearchRate = GET_PLAYER((PlayerTypes)getID()).calculateResearchRate(eTech);
			if( iResearchRate > iMargin * iCapitalBaseResearch){
				iCost *= 100 + iMultiplier*( iResearchRate / (iMargin * iCapitalBaseResearch) );
				iCost /= 100;
				}
		}

This means having as research rate for the total of your cities more than 2 times "5 * base research in capital", but less than 3 times that, will add an extra 16% to the current tech cost.

This penalty depends on having a lot of developed cities. Also, having Bureaucracy as civic won't affect, as it's the base research considered (or so I think, anyway, may be mistaken ).

With the idea introduced, it's a question of finding the optimal value for the iMargin parameter.

And that means testing. So if any voluntaries can put the edited DLL on the Assets folder (after making a backup of the older one!) of RFC and run some games (esp. American games, to see the tech level of all civs by 1775), it'd be great.

I also attach the edited CvTeam.cpp (the function is CvTeam::getResearchCost(TechTypes eTech)).

Edit1: the new DLL crashes the game. So I take it out until figuring out why.
Edit2: solved. Also, now the "more than 10 cities" is again taken into account, but just previously to apply the penalty, and doesn't influence it.

 

[embryodead]

The problem is, one can pretty much screw himself by focusing his capital on research (which is what I do often). That's not fair or balanced either, I think. The original bonus is a bit high though, that is true.

 

[JediClemente]

The problem is, one can pretty much screw himself by focusing his capital on research (which is what I do often). That's not fair or balanced either, I think. The original bonus is a bit high though, that is true.

On the contrary, the problem now would be to have cities other than the capital with a lot of research.

I think Rhye intended the penalty to prevent huge empires from having a lot of science, and unbalancing the game.

Having a lot of cities can give you a lot more science, but not always will (in fact, increasing your empire in a city more by founding or conquest will only produce a scientific advantage, if so, several turns later). This was the point I was trying to make.

I need some way to measure the tech rate to increase the cost progressively. It doesn't have to be in comparison to the capital. But then what?

 

[Wessel V1]

Quite a while ago I had another idea to 'fix' this problem:

Nothing happens up till 10 cities. Then the trouble starts.

The 11th city, in research order, will not have it's full output anymore, but only 90%. So, if your 11th city has a base research output of 90 (after the modifiers), it will only produce 81 beakers per turn.
The 12th city, in research order, gets a similar penalty, but increased. It will produce 90% * 90% = 81% of it's original output. So if the city had 80 beakers (lower than 90 of course), it will produce 64.8 = 65 beakers per turn.
So, in more mathematic form, the nth city produces (base * 0.90 ^ (n - 10)) from it's original output. That means, that a 20th city will only make a very small impact on research, hardly making any difference. This way, it doesn't hurt science to expand your empire, but it won't benefit to it either. 90% Is of course an estimate, probably 85% or 80% would be a more reasonable value.

 

[LuKo]

I think that Wessel's solution is very good (I think 90% is good unless tests prove that it isn't). But I know nothing about .dll and I'm afraid it can be hard to do.

 

[JediClemente]

That is an interesting idea.

But too complicated, isn't it?

It's always easier to increase the tech cost than to reduce a city's research output. Also have in mind this has to be computed each turn, for each civ.

That's why Rhye had a very simple mechanism, same for my first idea.

For this we'd have to sort all cities in order of research, and apply the reduction for each one beyond the 10th.

I was also thinking about adding the whole research from the cities in the core area and comparing it to the total research. If it representes less than, say, 60%, apply penalties.

 

[LuKo]

Still, I don't think it's as good as Wessel's plan. The perfect solution would be easy and efficient system, but as for now I prefer efficient over easy (I think that changing just a bit the system is not worth amount of work needed to compile).

 

[JediClemente]

I don't quite understand. Compiling the SDK is a piece of cake once you've done it before.

The other day I recompiled to remove the "barbs can raze holy cities" nonsense. It involved changing a line of code.

It doesn't matter if the change is big or little. If it adds functionality, at least I'm interested.

 

[Davinci07]

Thank you so much this will help- the Russians very much