Road System

[Aku]

I think there's a lot of possibility, but i wouldn't call roads "free from bandits", because bandits know they'll be people there. I think the safety comes from having more patrols on the roads more often.

[B-weasel]

Computer logically thinking, I'm not sure this is an "easy" task, and in some regards, would have to use "hidden" knowledge, as in, something an indvidual really shouldnt know, but as a computer AI could know.

Suppose the map were created with a grid system. The computer could plot the points of all characters every specified period of time. Well traveled points would have significantly higher values. Computer AI would direct bandits to those points, with some intelligently designed rules, of course.

In regards to the player knowing any of this, I am only suggesting that the player know his own paths of travel. A traveler's memory, so to speak.

[Jake_j_a]

On the note of no starting roads, this dose not make any sense at all. The world was not born with your character, so at game start there would already be well traveled roads in existence.
Give terrain a movement speed modifier, then have it mathematically calculate the fastest rout from point a to point b. This would give you a more realistic road system at game start, avoiding hard terrain when convenient, but going through at the shortest point when avoiding it is too far. This also would give the desired bonus for traveling a formed road as the more formed a road the less terrain movement penalty there would be.

Then there could be a 'danger' rating that would make the computer think this rout is slower, for the purpose of calculating rout taken, the speed penalty affecting a radius and increasing based on how many have been attacked and how many killed, but only if there were survivors, that would be the tricky part, figuring out how to calculate if there were survivors that 'knew' what happened. There would have to be a time delay as well, giving the information time to circulate.
If a setlement were sufficiently nearby, with sufficient force, it may send soldiers out to this 'danger' zone, removing the danger speed penalty from the calculation.

For the player, perhaps there would be a way to put dialog in nearby towns where random people (primarily merchants and guards people) might tell the player the road is dangerous. Additionally perhaps you could create 'missions' for the player (or AI) to clear that area. Unlike conventional rpg's the ai would not simply wait around for the player to solve it's problems, so the mission would expire when that area were cleared.

[Sprayer2708]

Then there could be a 'danger' rating that would make the computer think this rout is slower, for the purpose of calculating rout taken, the speed penalty affecting a radius and increasing based on how many have been attacked and how many killed, but only if there were survivors, that would be the tricky part, figuring out how to calculate if there were survivors that 'knew' what happened. There would have to be a time delay as well, giving the information time to circulate.
If a setlement were sufficiently nearby, with sufficient force, it may send soldiers out to this 'danger' zone, removing the danger speed penalty from the calculation.

Wouldn't the danger rating also go up if a caravan went missing and nobody knew what happened? Or if somebody using the road stumbled upon the remains of the looted caravan?

[Jake_j_a]

The caravan just disappearing it's hard to say, too many variables I'd think (was it a well known merchant who would be expected, just some traveler no one really new, a guy just setting out to make his fortune etc) In some cases it would make sense to up the 'danger' rating, in others it really wouldnt. Perhaps an easy solution would be to up the 'danger' a little less if there are no witnesses instead of not at all.
As for other travelers coming upon the looted/butchered caravan, that would be a witness, they may not have seen it happen, but they saw what happened.

Honestly I have no idea how hard it would be to code in witnesses to begin with, so it's really just a thought I was throwing out there.

[RebelMonarch]

Having the roads dynamically evolve as your game progresses is a fantastic idea! I'd love to see something to this effect implemented.

[MugenNoKensen]

The road system, i think, would be more enjoyable if it was completely aimless (not unless you had some sort of compass, map and/or directions from the nearby tavern)
Another suggestion would be that if you were to initially brave the walk to a mapped place, the likelihood of getting attacked or raided along the path from a to b, was to be lowered
EDIT:
Also neutral caravans/merchants would more likely be travelling through

[Argus]

Warning: Crushed dreams ahead.

Dynamically evolving roads would be terribly complicated to program, not to mention it would be really taxing on your processor to keep constant records of every location in the world AND of every location the AI goes. Consider the fact that some people already experience lag and the ground is just a texture mesh right now with no data stored about it.

Then we have to remember that the only part of the world that is actually loaded at any given time is the chunk(s) your squad is on, which means even if we did it successfully and it didn't tax the processor too much, you would only develop roads in the square area around wherever you spend your time, which could mean a really weird map if you spend a lot of time sitting in a player outpost, for instance.

Finally, what's the marginal benefit of a dynamic road compared to just laying down some roads between major cities? There's not much gain for all the effort: roads that don't connect the major cities by definition don't go to anything particularly interesting, and there's no means of AI's growing a new city organically, so expanding the road system beyond connecting the existing cities won't do much, meaning little difference between the much simpler static roads vs. dynamically evolving roads.

I like the idea of a road system, but there are other things (like fleshing out research tiers, implementing the faction system, adding hunger/thirst/needs) that I think should have a lot higher priority.