Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[GalenEvil]

Good luck on the new job!

Suggestion for roads, once you get to the need of implementing them:  Make roads not be just forced creations. Keep track of the paths taken between areas by "in-game" sapients and if it is greater than some threshhold per time period have them make a road there. That way they wont make a road between some out of the way tiny village and the metropolis, but instead will lead from metropolis to metropolis (generally) with offshooting roads along the route as needed. Once a road is established there could be a movement speed bonus, since it would no longer be through uncertain terrain but along an established and stable route.

Suggestion for cities/villages: You could simulate movement through the terrain with some number of entities over a period of time (without actually keeping them around) and then build cities at the most frequently passed through areas and villages at a lower threshhold. This way they will generally be along highish-traffic paths but will not look to be just thrown onto the world.

MUST SEE PROGRESS!!!!

* GalenEvil waits patiently ^_^

[Mephansteras]

Interesting suggestions.

I'm probably going to do something quick and easy for cities, initially, since I don't want to get too tied up in World Gen. Eventually I have some thoughts on what I'll do if I circle back around and deepen it, but for now I think a fairly simple algorithm for cities will suffice.

The map is divided into 5x5 Regions. Each Region will/can have a city in it. Cities will prefer to be by water and on either grasslands or hills if possible. Villages will then grow nearby as the city expands during worldgen and possibly during play. That should be enough for a first pass at things so I can move on to the actual game mechanics. Well...and a ton of UI stuff. I don't have much of that in yet at all, and there is a lot to figure out there.

Depending on how this first day at work goes, I may elaborate further on this line of thought.

[Mephansteras]

Been thinking about site generation a bit and how that will work. Also ended up adding some notes in about pillaging and conquering sites as well as army upkeep and supply chains.

Hopefully I'll get some decent coding time in tonight. If not, I doubt I'll get anything more done until Sunday at the earliest.

[Mephansteras]

Well, no real progress in a bit but the notes section has been fleshed out a bit more. I also have a bunch of programming notes set up so that when I do get to code (which should be tonight), I'll be able to dig in and get some good work in.

[Akhier the Dragon hearted]

   Notes are progress, just not programming progress. Half my projects I start would have progressed a good bit farther if I had written down notes about what I was planning.

[GalenEvil]

Yarr, notes are important! I try to write out as much as possible before writing anything more than boilerplate code for UI stuff.

[Mephansteras]

Yeah, yeah. It's useful stuff, just doesn't feel as productive since notes don't effect anything in the game until I implement it.

Coding has gone decently well tonight. Pretty much all background stuff so far, but I'm getting closer to having cities actually spawn. My goal is to get those to show up on the map by the end of the week.

[GalenEvil]

I like to think of notes as a roadmap for what you are going to implement. I use notes a bit to restrict myself on what to code, so that feature creep happens less often since I don't code anything that I haven't written out and at least thought about for a few hours while doing theory work

[Skyrunner]

The nice thing of merely implementing what is detailed in a paper is that the roadmap is there for you xD

[Mephansteras]

Between work and looking for wedding venues life has been very busy lately. But I'm still working on this when I get time. Hoping to get a few hours of coding in tomorrow night.

Currently pondering how the AI is going to work. I do have a question there for you guys: What kinds of things would you want out of an AI in a civ/conquest game?

[GreatWyrmGold]

Currently pondering how the AI is going to work. I do have a question there for you guys: What kinds of things would you want out of an AI in a civ/conquest game?

A few levels. Start with the easy-mode AI, it should be easier. Higher AI levels could make more intelligent decisions and perhaps probably cheat more. Ideally, a fair amount of the difference between AI levels would be improving the AI's strategic thinking.

[Mephansteras]

Yeah, the easy-mode AI will be pretty simple. Initially it'll be basic build-defend for neutral cities and build-expand for the 'active' AIs.

Ideally, I want to get it to the point where the AI doesn't have to cheat to be a threat. I'd rather that the difficulty levels had more to do with your starting situation than having cheating AIs. Having everyone love you and hate the other AIs should make the game a lot easier than having everyone start off hating you and being friendly with one another. I'm pretty comfortable with a neutral starting position being slightly in your favor as long as you know how to play the game well.

Diplomacy is something that is sort of a long-term goal, but I'd really like to get in and have it be good at some point. Something better than the classic Civ diplomacy of 'Well, the AI only accepts deals in their favor and is generally dumb about everything up to the point where you beat the tar out of them and are able to force concessions'. I'd like actual relationships between factions and for the AIs to behave reasonably. That'll mean more when the game has a victory condition better than 'Own x% of the world', which is how it'll be to start. When you have to kill everyone else to win it's hard to be reasonable about things.

[Akhier the Dragon hearted]

   You will probably end up having the AI cheat in some way. Thats fine, it happens but don't make it cheat badly. Extra resources? Fine, they need it to beat me but it should be either a one time startup amount or a modifier on their current income. Map hacks to know where everything is? Reasonable and may even go in just to make the game run faster depending but don't let their knowledge affect their diplo like how in Civ4 you can have someone hate you for a trade you made before you even met them. Better stats on their units? Why not but make the differences visible, being beaten when they have better swords is obvious but being beaten when the AI gets a hidden +10% to hit just makes people mad.

[Rolepgeek]

Looks extremely cool, the type of game I'd love. Are you taking suggestions or just updating us? It looks really cool, I just always love more complexity, thus meaning I take a look, and see how something could become more realistic/advanced, usually at the coder's expense. Which is why I would never be able to make one of these, because I'd never stop making it more complex.

[Max White]

Natural levels of AI can be pretty easy to achieve if you have the game mechanics that allow for it.
For example, lets assume that each AI can't cheat by telling a player from another AI apart. Anything that the AI does it does for its own ideals, not to interact with the player in a specific way. If you want to make the AI easy on the player and hard on the AI, you need to make it preform the same behavior that is would punish.
So if you have  system to give tributes to other players, then 'easy' AI might be very willing to accept any tribute and totally unwilling to give tribute to others, making the player able to buy their way out of trouble without all AIs forming an unholy alliance.
Another possibility would be if an AI was incredibly vengeful, always choosing to target players that have already attacked it, and also willing to zerg rush other players, leading to massive conflict between the AIs that the player can choose to step out of by playing more defensively.

This sort of AI paradigm is much better suited to unique roles such as 'Merchant AI', 'Warrior AI', 'Builder AI', 'Scientist AI' and others rather than the more linear 'Easy', 'Medium' and 'Hard'. It also allows the player to experiment and have fun with different configurations, and even set more unique challenges for themselves.