A few tips on the AI ruleset; In essence, the ruleset means that a human player is going to go one of two ways, either they lose early because they cannot develop fast enough to ward off hostile AI nations and warred to death or they win because they research and develop the opposition to death, the game becomes boring because either humans lose early or win in the long term, knowing they are going to win well before the actual end. The AI scores rarely get above 400, so as soon as the human reaches a score of 400 they know they are winning, but that is just ⅓ into the game itself.
Set out below are some of the reasons why this happens; Differentiation between different combat units There seems to be no differentiation between different combat units. EG a sea unit will attack a ground enemy when there is nothing to be gained by it, purely because it can and will win and usually because the ground unit is weak. However, in doing so, it will merely weaken itself to counter attack from a stronger enemy sea or air unit. It is very rare for a sea unit to prosper in attacking an enemy ground unit outside of a city. Ruleset change; Sea units attack sea units or cities only. Ruleset change : A ground or air unit only attacks another unit for an objective (defence/attack), at the moment the objective seems irrelevant to the decision unless its critical. Warfare It seems that there is too much bias in the ruleset to begin warfare. Warfare is expensive, especially when units are constantly being lost in such warfare, it should really and only be a resort once (1) the nation is large enough to accommodate it and (2) the nation has the capability of replacing the units lost in warfare to the same or similar extent without affecting other advancements. The current system is so warfare orientated that an AI nation will almost be happy to commit suicide trying to beat an AI opponent and leaving itself completely unready for an attack from elsewhere. Furthermore, it seems that warfare in one part of the nation is far more important than development in another part of the nation. What happens is that two nations with the same border are constantly at war with each other, constantly losing units in that war and seemingly never ending the war despite the fact that both will be dramatically weakened by war and will lack the ability to actually defend against another enemy afterwards and they are so concentrating on war that they lose the ability to develop in other areas. This has improved with the diplomacy but is still an issue as diplomacy doesn't seem to be used by AI very much. Ruleset change : War should be a last resort not a first decision. Perhaps increase the parameters to make was less appealing. Also peace should be the first objective. Also, it seems that warfare, once instigated just goes on and on and on, there seems to be no period of consolidation, so an AI army will be like locusts, going around killing everything it can but leaving very little behind. Once a city is conquered, the AI should consolidate. Ruleset change : At least 5 turns before the next city is attacked. Also there is substantial unhappiness due to many military units being outside the city fighting elsewhere. This leads to a city having to cut production in order to deal with unhappiness. Ruleset change : only 1 military unit that is supported by an AI city is allowed to leave the city itself. Lack of research I find that the AI research seems to shoot along to a certain level and then the research slows down dramatically. The easy way to win is to stay out of the battles and make peace until you are way beyond the AI research level and then make units that are much stronger (such as stealth bombers). Research should continue as a priority in AI continually but I suspect the AI ruleset doesn't put enough emphasis on research and lets it drag at whatever speed it takes. Also, every city I build has a university in early build. Not a single AI city I take over has a University. That means that I can produce more research from 6 cities than an AI with 20 cities. Ruleset change : Put more emphasis onto research and less onto warfare. Lack of proper development of cities The AI really only gives a city the barest minimum requirements and usually not even that. EG my cities have 25 structures, the ones I take over usually have 1-2 and maximum 5. No city I take over has (1) factory, (2) bank, (3) Stock Exchange, (4) mass transit, (5) supermarket, (6) super highways, (7) sanitation. Due to this, the AI cities stay very small and don't develop. Once my cities are developed, the AI can never take them over or even match them. Ruleset change : Development of cities needs a much higher priority put onto it in the ruleset, unless a critical build is required. Lack of use of the research that has occurred already Many times I take over a city, it is defended by a Marine (usually many Marines and other units), which means that there has been substantial research, but the city itself never seems to represent that research and the research level is way ahead of where the city development is. What this means is that the city can build good defensive units and has lots of them, but doesn't have the ability to produce the right amount of product or money to properly defend itself against a continued attack and once the defensive units are gone, the city cannot replace them fast enough. Ruleset change : Unless there is a critical build required, the first build of a city should be the most suitable city development that relates to the level of research achieved. Instead of building an AI army, an AI city should build a bank. Roads and railways It is very rare for an AI nation to build extensive railways, the most they get to is purely interconnecting local cities. Roads & (once researched) Railways should be build in every square that serves a city. Ruleset change : Roads and railways to be built in every square that services a city Ruleset change : Objective - All a nations cities should be connected by roads and railways. Also some AI cities are isolated and not properly connected to the other cities in the same nation. Ruleset change: Priority build is road and railway link to another city in the same nation. Cities AI cities are almost built anywhere, they build too close to each other and on inappropriate squares. Ruleset change : An AI city cannot be built which encroaches upon another city in the same nation. Ruleset change : Different levels of square are given city suitability values for AI city build purposes. The good build squares (Plains and Grasslands) should have a much higher value than bad build squares (mountains and deserts). Ruleset change : If a locality is suitable to AI city build, then the decision as to where to build is decided by the highest suitability value square in the locality. Ruleset change : Cities cannot be built on mountains. Ruleset change : unless unable to do so, all cities should have irrigation and farming in the city itself as soon as possible. Terrain development AI terrain development is sometimes just wrong. For example, hills always seem to get mines, even if they have a special resource (such as wine) whilst mountains don't seem to get anything (not even roads). Also AI terrain development is sometimes useless by being outside any city's domain. Ruleset change : development of AI terrain should have basic rules set as follows (all developed tiles to have road and railways); Outside any city domain : none except roads/railways to connect cities. Forest : none other Plains : Irrigation & farming Grasslands : Irrigation & farming Deserts (with oil) - mining Deserts (without oil) - irrigation & farming Mountains : mines Hills - irrigation & farming Swamps (with resources) - no other Swamps (without resources) - irrigation & farming Jungle (with resources) - no other Jungle (without resources) - irrigation & farming. Tundra - irrigation and farming Ice sheet - mining Ruleset change : Development starts in closest undeveloped square to the city itself once the population of that city has occupied up all other developed squares of that city Just some ideas to get on with !!!!!!
_______________________________________________ Freeciv-dev mailing list Freeciv-dev@gna.org https://mail.gna.org/listinfo/freeciv-dev
