Messages

I'd suggest a change to how tasks are assigned, as follows:

Currently tasks are assigned based primarily on whether a colonist has a task already, with distance to the taks being assigned used as a tie-break. With colonists only able to remember one task at a time there's not much more that can be done, but what if colonists were able to plan ahead and store a `to-do' list of tasks? (This could be as simple as a FIFO queue, though of course a priority queue would allow for more sophisticated systems.)

With a list of tasks for each colonist available, tasks could be assigned based on a function of both distance and the number of tasks on the colonist's to-do list. (Note: In this case it makes sense for `distance' to mean distance to the location of the last task on the colonist's to-do list, rather than distance to current location.) The function could be determined heursitically according to what works and also be exposed to modding so that players can change behaviour to suit their own playstyles.

For example, let's say the function is (total time of tasks on colonist's to-do list) + distance/(colonist speed) and suppose there are 15 idle colonists around the map, and 10 metal tiles we want to build. The first tile gets assigned to the closest colonist. The second gets assigned either to the same guy if there is no-one who can get there before he completes the first, and to the closest free colonist otherwise, etc. If there were 3 nearby colonists and many far-away ones, then those three would complete 3-4 tiles each and the rest would sit tight.

 -- Manual Prioritization - Prioritize nearest task:
Currently colonists apply their own priorities to tasks even when Manual Prioritization is active. This can create difficulties in certain situations, as in the following example.

Suppose I want a colonist to deal with the nearest available cleaning or repair task.
If I assign equal priorities to the two tasks, the colonist will focus on repair even if there is cleaning to be done nearby. On the other hand, if I assign cleaning a higher priority then they will focus on cleaning even if repairs are needed nearby.

It would be nice if Manual Prioritization overrode all such considerations; i.e., if colonists selected the task with the highest selected priority, ties broken only by distance. This would give players a lot more fine control over colonist behaviour.

-- Intelligent Hauling:
When a colonist has just selected a new task which is far away and Hauling is enabled for them, they should check whether there is an object which
(a) is near to them, and
(b) has a desired destination near to the task just selected.
If these conditions are met, the colonist should forget about the task and do the hauling first.

This would greatly increase efficiency in certain situations; for example, long-distance mining colonists would pick up the metal just mined when going home.

-- Combined prioritisation and designation:
When right-clicking on a tile, show the following tasks:
- Mining Rock/Mineral even if not designated (if selected, this designates the rock/mineral to be mined and then prioritizes the task).
- Hauling even if forbidden (unforbids the object and prioritizes).
- Debris hauling even if not designated (designates for hauling and prioritizes).
- Cleaning/Fire-fighting even if outside home zone.

Currently in the case where we just want to send one colonist on a specific job, we have to designate, find the colonist again and select him and then prioritize. Also this pretty much has to be done while paused, otherwise another colonist may take the job.

It's difficult to think of a raiding team that could survive that sort of defense!

Having re-read Sky_walker's post and thought about it a bit more, I can see where he's coming from. Knowing which path raiders will try to take in advance makes defending much easier - chokepoint or no chokepoint. I also think the AI he laid out seems very reasonable and would make things more challenging, while still allowing the player to respond (by coming out and attacking the demolitions team before it's too late).

Creating a bigger hole is basically what I had in mind. It seems like all of these funnels and killboxes rely on getting the raiders into a small area where they can be blown up and/or shot at, right? At the very least, by attacking surrounding walls there would be a chance that the AI destroys some vital component of the player's trap. I had envisaged the AI using grenades to destroy structures; those work well against rock as well as constructed walls.

If the raiders got shot at while trying to demolish a structure, they could simply forget about demolition for a moment and return fire. Alternatively, as you suggested for the specialised team, the demoman could continue his work while the rest defend him and engage the player's forces. So as soon as they blew a hole in a wall which revealed a turret, they would shoot at the turret rather than trying to blow up more walls.

The narrow corridors thing is a really good point and I don't see a solution without the raiders having some sort of proper objective. As it stands, the raiders objectives seem to be "blow stuff up and kill anyone who gets in your way". And actually, turning an underground base into a large open cavern fits that objective remarkably well!

Once they have proper objectives, then the demolition behaviour could be limited to when initially trying to breach the player's defenses - after that they could revert to their usual tactics.

[different]

Well, it is and it's not. Part of the reason for this type of squad to exist is that it'd have a different attack strategy than regular raider AI.

Sure. I guess I was also thinking of having regular raiders with demolitions capability, which could deal with...

Formations don't help with door exploit (you still got that one narrow point which gathers all the raiders and sends them through into the turrets / blasting charges).

The point I'm trying to make later on in my previous post is that it *would* help. If there is one narrow point, the AI will realise that its formation cannot fit through the narrow point without breaking formation (i.e., bunching up together). This could be the trigger for it to try and avoid that narrow point - possibly following a routine like the one you suggest, or possibly just destroying some of the walls/terrain around the narrow point to make it bigger.

 -- Weapon: Rocket Launcher:
Fires powerful rockets.
+ Fairly long range (about the same as a Lee Enfield).
+ Deals high damage to buildings (about the same as a grenade).
+ Splash damage to buildings.
+ Deals moderate damage to humans.
- Very slow firing rate.
- Very inaccurate.
- Heavy (slows movement rate when carried).

Basically this would prevent the player relying too much on entrenched defenses.
The low accuracy and lesser damage to humans would prevent this from being overpowered in the player's hands.
Possibly it should be very rare and/or only occur on higher difficulty levels.

AI use: Should be straightforward; just use the same AI as for Grenadiers, but fire from longer range.

-- Raider: Sappers:
Sappers can disarm Blasting Charges, rendering them harmless.
They show up with raiding parties and are used to clear a path through minefields.
There is a small chance (~5%) that the Blasting Charge will *instantly* explode (without waiting the normal time for the fuse to burn),
instead of being disarmed.
Colonists can rearm the charges later (so they are not lost permenantly), but again suffer the same ~5% chance of an instant explosion.

AI use: During the "staging" phase, the AI computes a path from the landing site to the player's base (I assume it has to do that later anyway!). The sapper tries to disable each blasting charge whose area of effect intersects the planned path. If the charge is activated and the Sapper is in range, he runs away. Once the staging phase is over he acts like a normal raider.

-- Mechanic: Sanity Meter:
Introduce a Sanity meter alongside Loyalty/Happiness/Fear.
The Sanity meter depeletes whenever Fear is greater than Happiness and recovers whenever Happiness is greater than Fear. Movement of this meter is slower than the others, so colonists will not quickly go insane just because they are frightened. Rather, insanity would occur if the colonist is subjected to consistent stress for a period of time.

If the Sanity meter drops below 10%, they will suffer a Mental Break (works just as it is now). If the Loyalty meter drops below 10%, they will *not* suffer a Mental Break. Instead, they will just try to leave the colony.

I had some thoughts related to this - specifically, what sort of AI routine should a demolition squad, or indeed a regular squad with some explosives experts, follow?
I think this is closely related to the raiders AI in general situations.

One thing about the AI currently is that it always attacks in single file. This gives more firing time to the defenders and reduces the amount of fire the player faces at any one time. I believe it would make a big difference if the AI was able to move its units in a formation; for example, a line parallel to the player's defenses would allow them all to enter the player's range at approximately the same time.

Further, if the AI could handle formation movement then it would also be able to detect chokepoints! More specifically, anywhere the AI could not move its troops while remaining in formation is by definition a chokepoint of sorts. If the AI discovered a chokepoint, it could try to overcome it by attacking the surrounding terrain with its weapons.

Of course some flexibility would be required here so that the AI doesn't try to carve massive paths through mountains to get to the player. But this could be overcome by first having the AI attempt to change formation, or break up into smaller formations and have each formation advance in turn. Attacking the terrain could be made a last resort, so as to reduce the amount of cumulative damage to the terrain from successive raids.

I think this would be a great idea. Very early on in the game, when you have some colonists but only one weapon (or a few, weak weapons), this would be very powerful and give you a good chance to survive early raids. Later on, however, you would have to choose between colonists using the weapons you have (which are now decent) and manning the turret - thus avoiding breaking the late-game altogether!

Another advantage is that it would encourage having multiple layers of defense; that way, if your turrets in one layer were destroyed then you could fall back to the other layers.

Hi! I found out about Rimworld via browsing Kickstarter, and very much liked the idea. I feel it's a shame city-building games haven't seen that much focus in recent years, and that space colonisation games seem to be almost non-existant. I played Outpost many years ago and have often wished someone would make a better game along similar lines. And yes, I have played Dwarf Fortress and it is quite amazing.

My favourite game is Starcraft, for its considerable strategic depth and the sheer amount of excitement and tension present in each game. I also loved Baldur's Gate II and Planescape: Torment for storytelling and the depth of the world they created. And... well, I've played lots of games, so the list of ones I like is quite long.