Difficulty system not based on wealth [1.0]

[Snafu_RW]

I think raid strength should fundamentally be time-dependent.  A simple line or curve whose slope (or equivalent if a curve) increases with difficulty level (basebuilder would be pretty flat, extreme quite steep).  Difficulty *should* increase with time, requiring a colony to advance to survive. 

Cassy doesn't give that already? (See the storyteller description!)

@ Crow: /Colony total wealth/ already affects raid size/danger (see above OP complaints); however I agree that raids could be more varied: eg a 'snatch' raid to kidnap a specific colonist*; a 'grab' raid to gather <stuff> (food or high-wealth items etc); a 'kill' raid to wipe out your colony, etc..

No idea how this will develop into the game; I'm just throwing out thoughts..

Don't forget that the 'win condition' of this game is to escape the planet (ie build/find a starship). If you choose to ignore that goal & treat it as a base-builder/colony game then fair enough, but don't expect conditions to improve..


*Esp. if said colonist is a relation; perhaps ramp up that raid type probability?

[Greep]

The storyteller description is incorrect.  There is no time element.  There used to be a "long term rampup" which sort of but not really was time based, although that got switched into adaptation which is a little more reactive.  But even the old system was multiplied onto wealth.

[5thHorseman]

I hated the old timed system. It meant you were fine until you lost once and then you were screwed because the next attack would be bigger while you were smaller. I never found a situation where it was worth it to keep playing. I always did, it just was never worth it.

[NiftyAxolotl]

The game can't ever become easier or safer as a colony improves, because then it would become boring. Dangers have to scale in line with the player's ability to handle them, keeping the risk level the same.

See how ARPGs like Diablo deal with this. The player never makes any lasting progress on difficulty, but progresses on story and aesthetics. Chain Lightning versus level 30 monsters isn't any easier than Shock Bolts at level 1, but it looks and sounds so much cooler, and the control feels smoother. Often, the late-game monsters are the same as the level 1 monsters except for a palette swap. But having cycled through other monsters still feels like progress.

RimWorld lets the player decorate bedrooms, buy guns, orchestrate complex production lines, heal maimed colonists, train pets, install air conditioning - all sorts of goals to strive for, achieve, and feel good about. The problem with wealth-based raid scaling is that every accomplishment is poisoned by a feeling of helplessness. The colony is endangered because all threats get bigger. How can the player prepare against "all threats get bigger"?

But the player can prepare against specific, telegraphed dangers. I like the themed, motivated raid idea. Specific, well-communicated player accomplishments should unlock larger sizes of different raid types, which try to push the player back on that accomplishment.
- Thieves scale by how much carryable loot you have. And they try to steal it.
- Tribals scale by how much electricity you are using. And they try to destroy it.
- Mechanoids scale by how many Advanced Components you have or are using. And they try to destroy them.
- Infestations scale by how many excavated mountain tiles you are occupying. And they try to push you off of them.
- Slavers scale by the value of your colonists. And they try to kidnap them.
- Maddened herbivores eat your grain, carnivores eat your animals, boomrats burn your furniture.
Threats feel more actionable if they are specific. If I know that my Alpaca ranch is going to attract swarms of maddened wolves, I can prepare for wolves. Which feels much better than a vague knowledge that everything will become more dangerous.

I also like that a motivated raid has limited (but painful) consequences for opting out. Currently, fighting a raid to the death feels mandatory.

[East]

There is an easy solution.
In order to build a particular building, it requires specific materials.
And this material has to get a fixed scale threat. This threat can be done at the desired timing.
This battle can not be done by reducing property.
Fixed threats can be difficult to deal with with with low property.
It doesn't have to be a battle. It could be a huge mental shock or a long winter.

It seems to be a good result if you combine the increased threat according to property and the fixed threat of the desired timing.

If your property is high and you have enough fighting power, it's easy to strike a fixed threat, but on the contrary, a property-based threat will be difficult.
If your property is small, fixed threats will be difficult and property-based threats will be easy.

[erdrik]

...
These are good suggestions.  However, there still needs to be some mechanism to determine the basic strength level of any raid or pack that is generated.

I think raid strength should fundamentally be time-dependent.  A simple line or curve whose slope (or equivalent if a curve) increases with difficulty level (basebuilder would be pretty flat, extreme quite steep).  Difficulty *should* increase with time, requiring a colony to advance to survive. 

That said, the basic shape of the time curve should be modified by things like:
...

I don't really support time dependent raid strength, partly for meta reasons.
Eventually if you play long enough you will reach a point where your computer simply cannot handle the horde the game will send at you. If the intent is endless scaling that eventually kills you, there needs to be a hard cap where instead of adding ever more numbers, you get raids with ever stronger individual units.

Even if the intent is not endless scaling, I'd still prefer prioritizing the scaling the strength of individual units over numerical hordes.

[Greep]

Time based doesn't necessarily mean time only.  While vanilla has like 1000 lines of code in various areas related to the raid size, making it time based doesn't necessarily mean you need to make it a one liner.  But you can do a 1 liner if you like: exponents < 1 take care of what you say and slowly cap it in a gradual way.  Although having looked at a lot of graphs, I think you want it to be the other way around and be > 1, as wealth grows almost quadratically in rimworld.

The other part I agree with, however, that simply means there should be new units with higher points per unit.  The actually formula for raid points would be increasing all the time.

[seerdecker]

Thanks all for the comments! Your feedback has been very helpful to improve my own proposal. Even though I can't please everybody, I tried to incorporate as many of your suggestions as I could.

First, let's discuss some fundamental scaling issues. Suppose that the raid strength depends primarily on wealth or defence strength or time. Two things can happen. If your effective defence strength increases faster than the raid strength, then the game becomes too easy. Conversely, if the raid strength increases faster than your ability to defend against it, you eventually lose. Neither those situations are desirable.

In particular, if the raid strength depends on wealth, and the game becomes exponentially harder with wealth, then we obtain the current state of Rimworld. Most of the wealth becomes focussed on defence. Raid sizes become huge and tedious to clean up. The computer starts to lag. Non-pawn-based defences become the dominant mean of defence. It leads to dubious incentives such as damaging your weapons and limiting the luxuries of your base.

Now, if the raid strength depends primarily on how well on how you dealt with the last one, then eventually, no matter how well you play, you'll get a bloody nose because the raid strength will keep increasing until you lose. I don't find this very satisfying.

I propose to resolve these scaling issues with a system based primarily on progression.

The stronghold system

The stronghold system offers an alternative for the players who do not seek to escape with the ship at the earliest opportunity. It is fully compatible with this goal, however.

There are three variables used by the system: the luxuries (aka wealth), the defences, and the development level. You start the game at the lowest development level. Your goal is to progressively increase your development level and reach the maximum level you can sustain, both militarily and economically.

At the beginning, the other factions look at you with contempt. Look at the little settlement, they say. No electricity. Play with a horseshoes pin. Savages! Even the actual savages look down at you.

Accumulating wealth is very dangerous initially. As per Thane's suggestion, the pirates will leave you alone so long as you have nothing worth taking. If you start churning out golden statues while your defences consist of a grand total of three traps, you'll be sorry. However, hoarding wealth becomes less of a worry at high development levels. By then, it will be widely known that the stuff is guarded.

In terms of mechanics, the raid strength is very sensitive to wealth at low development level. The wealth contribution levels off at high development level. Eventually you can hoard luxuries as much as you want. In fact, you'll have to if you play on high difficulty.

The pawn expectations depend only on the development level, with some exceptions made for caravans. As per Koek's suggestion, expectations start very low, but as the development level increase, wealth must be obtained to improve the base to please the pawns. Critically, high wealth by itself does not lead to high expectations.

The defences are counted in a different pool than the luxuries. This includes walls, doors, traps, turrets, weapons, worn apparels, pawns, bionics, and animals (based on actual combat potential -- chickens don't count for much).

Like Awe proposed, the weapon quality has no bearing on its defence rating; it only affects its market value. Thus, obtaining a legendary weapon is a major reason for celebrating. Conversely, awful weapons are truly awful. This is true for armor and apparel as well. Apparels like cloth dusters that don't provide meaningful protection do not count as defence.

The defence pool is soft-capped by the development level. You can build sky-high defences if you want, and sit back and relax while your traps and turrets take care of the raiders automatically. You'll have a mutiny on your hand, though.

See, your pawns interest lies mainly on having a good time and making progress. They'll put up with building some defences, but they won't tolerate ugly turrets everywhere they look, traps that can blow them up at a moment's notice, and walls so high they can't see the scenery anymore. Wealth won't placate them indefinitely. Yes, that golden statue looks very nice, but they'd rather watch X-rated movies on that plasma TV that they don't currently have in their neolithic development level.

In terms of mechanics, the raid strength is marginally influenced by your defence rating. Lower defences lead to slightly lowered aggression, since your attackers believe you're an easy prey. It helps you to recover between raids. The main raid scaling factor comes from your development level. The attackers guess your probable defences based on how advanced you are. When your development level increase, the raids become stronger, but your defence pool also becomes larger, i.e. you can build more turrets without incurring a revolt because your pawns acknowledge the necessity.

The defence pool is a critical resource to manage. At high difficulty, it is too small to realistically repel a raid using only fixed defences. You have to use every tool at your disposal, especially your pawns. The defence pool size increases slightly slower than the raid strength, so the game becomes harder as you go on. But, since time is not a major raid scaling factor, you have a chance even on sea ice. As many have suggested, the difficulty can also change dynamically when you're dealt a bloody nose. Adaptation mechanics are already in the game. Importantly, you never have to deal with gigantic raids that are a chore to handle.

The player upgrades its development level by researching the corresponding technology.

The core of the proposed system can be implemented without too many changes to the existing mechanics (possibly in a mod). Mostly, the accounting needs to be changed and the research tree must be tweaked to integrate the development thresholds. The main challenge is the balancing.

Optional stuff

Ideally, it would be very difficult to obtain high-level stuff in a low-development colony. Trade ships and outlanders snub you. Raiders don't carry high-level gear (and if they do, beware). Basically, you're stuck with the stuff of your era. Quests could be an interesting exception.

It could be fun to make the defence cost of an item non-linear with the amount of that item that you have, e.g. to give freebies. Say, the first 2 power armors you have don't contribute to your defence pool. So, this encourages you to use all of the variety of the tools at your disposal, even those with lower absolute utility.

The number of pawns you can recruit could be affected by the development level. Not many people want to live in a slum.

The development level could scale off indefinitely, so you can brag that you reached development level 12 on merciless. This allows the difficulty to increase smoothly past the intended difficulty of merciless, for the players who want to. Conversely, for players who just want to build the base, the defence pool is effectively unlimited.

Another possibility to soft-cap the defence pool is to make the raider strength non-linear with it, rather than incurring a debuff. Don't scare the neighborhood -- they don't like it.

The development level could also have economic consequences. To make your stronghold rise to pre-eminence in the world, you have to bribe politicians, impress the world with your art, and feed, dress and heal the hordes of visitors who seek your protection and charity. Then, the challenge becomes to increase your throughput to meet the demand. This makes your pawn skills important. Rimworld meet Factorio!

tl-dr
- Separate wealth from defence.
- Defence pool is limited, incur major debuff if exceeded.
- Expectations, defence pool and raid scaling depend mostly on development level.
- Development level increases at the player's pace, through research.

[EvadableMoxie]

The more I play and understand the wealth based raid system (And adaption) the more I dislike it.  There is really no reward for good management and efficiency.  If it takes me 60 days or 10 years to get to 75k wealth it's all the same to the game.  If you're a good manager and your colony expands and progresses quickly you're generally punished for it.  On the other hand, you can sit at 18k wealth forever and wait for your colonists to research the entire tech tree with no punishment.  Likewise if you defend raids perfectly they get harder and harder, until eventually you can get a raid that's powerful enough to destroy the entire colony.  On the other hand, if you did less well and were taking small losses, you get adapted into easier raids.   Success is punishment.

It's to the point where I don't really want to play on Merciless anymore.  Not because I can't survive on Merciless, but because playing on Merciless involves you going against what I feel should be the actual point of the game: Making the best colony you can.

[RemingtonRyder]

The main problem with using wealth as a base for generating threat points is that not all of it translates into something which gives you an edge over the next raid. It isn't selective enough. One of the things my mod Combat Readiness Check does is to ignore the item wealth calculation in favour of assessing the wealth value of weapons and protective armour that the colony has available to them (equipped or not).

By the way, there's a cap of 20000 threat points in 0.19, so there is sort of a limit to how large raids can be.

[Greep]

I think combat readiness is a great mod if you're playing in a sandboxy way, sort of like a very adaptive difficulty.  However, I don't think it's great on a colony planning level, as there would be no real point to building any sort of defenses: the enemy would just scale exactly to those defenses.  And it kind of relies on the modder to have perfect knowledge of how much your preparation is worth in raid value, which I think it's better to be blind to.  One thing I think 1.0 kind of failed on is combat animals, and it's because it's pretty hard to choose a right number for how much raid points an animal is worth.  So you really shouldn't.

[RemingtonRyder]

There are settings so you csn adjust the threat points that things generate. The defaults mean that only light defences are needed, but if you crank it up then yoy need to be more prepated.

[Wanderer_joins]

One thing I think 1.0 kind of failed on is combat animals, and it's because it's pretty hard to choose a right number for how much raid points an animal is worth.  So you really shouldn't.

I don't know if it fails. Animals are fun and it's not easy to manage a large number of animals and use them against raiders. It's rather like a weird strategy which shouldn't be penalized.

[Nynzal]

What bugs me is that wealth should be a major indicator for raids. Ofc people want that legendary stuff you got going or the huge supplies you have. I actually have never bothered with controlling my wealth, but there are fair points that increasing colony wealth doesnt do anything and just makes life harder.
I dont think there is a perfect solution and "exploits" will always happen.

[bbqftw]

The game can't ever become easier or safer as a colony improves, because then it would become boring. Dangers have to scale in line with the player's ability to handle them, keeping the risk level the same.

See how ARPGs like Diablo deal with this. The player never makes any lasting progress on difficulty, but progresses on story and aesthetics. Chain Lightning versus level 30 monsters isn't any easier than Shock Bolts at level 1, but it looks and sounds so much cooler, and the control feels smoother. Often, the late-game monsters are the same as the level 1 monsters except for a palette swap. But having cycled through other monsters still feels like progress.

RimWorld lets the player decorate bedrooms, buy guns, orchestrate complex production lines, heal maimed colonists, train pets, install air conditioning - all sorts of goals to strive for, achieve, and feel good about. The problem with wealth-based raid scaling is that every accomplishment is poisoned by a feeling of helplessness. The colony is endangered because all threats get bigger. How can the player prepare against "all threats get bigger"?

But the player can prepare against specific, telegraphed dangers. I like the themed, motivated raid idea. Specific, well-communicated player accomplishments should unlock larger sizes of different raid types, which try to push the player back on that accomplishment.
- Thieves scale by how much carryable loot you have. And they try to steal it.
- Tribals scale by how much electricity you are using. And they try to destroy it.
- Mechanoids scale by how many Advanced Components you have or are using. And they try to destroy them.
- Infestations scale by how many excavated mountain tiles you are occupying. And they try to push you off of them.
- Slavers scale by the value of your colonists. And they try to kidnap them.
- Maddened herbivores eat your grain, carnivores eat your animals, boomrats burn your furniture.
Threats feel more actionable if they are specific. If I know that my Alpaca ranch is going to attract swarms of maddened wolves, I can prepare for wolves. Which feels much better than a vague knowledge that everything will become more dangerous.

I also like that a motivated raid has limited (but painful) consequences for opting out. Currently, fighting a raid to the death feels mandatory.

the current system is also boring, as ultimately there is no sense of progression, and that most strategic choices are meaningless besides a very specific set of ones that lower your wealth. In fact it is almost worse in the sense that early game outcomes are dominated by combat choices, late game is dominated by considerations like "how do I destroy all these weapons that sell for 20% price but count for full market value" and "corpse disposal party", since your colonists will break at any simultaneous sight of rotting and regular corpses.

I like to discuss optimizing economy, but the conclusion is that none of it really matters. Play well, or play badly in this regard, arguably the latter is better.

Same concept with mood management. Everything you know about making space efficient bedrooms that pawns like goes out the window when you realize that it means nothing in the face of "make sure you don't go up an expectations tier"

PS there is absolutely progression in Diablo, the difference in efficiency of geared and ungeared character is immense.