Difficulty system not based on wealth [1.0]

[zizard]

If one clicks back and forth between the wealth and debug pages, one finds it difficult to say "raid strength isn't primarily based on wealth" with a straight face.

[Bolgfred]

C) The solution

Fixing the wealth issue is much more difficult than listing its shortcomings. What do you think could be done to improve the game with respect to difficulty? Discuss!

Keep in mind that it has to be simple to have a shot to be implemented, i.e. implementation effort must be handled as wealth

I think that the wealth system makes people who know about it feel threatened. Those who doesn't know about instead, get a very balanced gameplay feeling as the wealth adaption is actually a pretty good idea.
In this way, same like the unknowing, nobody should be afraid of the wealth system. It affects the difficutly, but in a good way.

A rough Example:
Let's say you will craft a excellent LMG. it's a strong weapon that makes you stronger. If i get the math right thats LMG 300 x excellent 2 = 600 wealth. itemWealth gets divided by 100, so its 600/100=6. The next raid will have 6 points more than the last one.
A raider has a point value of 35-210 depending on its type. The difference of 6 means, that from 20 Raiders attacking you next time, the one with the two peg legs, will only have one peg leg!
This means, while you get a rambo rifle, the raider gets a new leg. Now I ask you: Is a LMG able too fight a leg? I'm pretty sure it is.

Here's the formula:
P = (( C × 42 ) + ( iW ÷ 100 ) + ( bW ÷ 200 )) × ( sR × lR ) × D × T × R
C= Colonist
iW = Items
bW = building

Further information here:
https://rimworldwiki.com/wiki/Raider

[Jpagano]

It's always interesting the difference between how players interpret increasing difficulty as "punishment" versus progression.

...

In RW interpretations seem to shift back and forth across this line very fluidly and it's hard to see exactly what drives that difference. It seems very personal and very situational, and (at least on the forum) can even be driven by meta concerns or theoretical analysis of code or XML.

...

I think it's not just increasing difficulty, but the way in which it increases (both cause and variety). I enjoyed my 1st unstable play through, haven't been able to pick it up again yet due to work, but appreciate that you're listening to feed back.

I think Oblivion vs. Morrowind is a decent case study re: effects of scaling. IIRC, Morrowind had no scaling. While this occasionally lead to unwinnable situations, I thought it enhanced the story telling, and gave the PC a true sense of progression: mid-level ancient ruins that obliterated the PC at a low level are memorable because of the danger, and once the PC has gained strength, that strength has tangible meaning (the ruins are no longer an impassable threat). This does lead to eventually "capping" difficulty, but I think that's fine (whether it's above or below what is feasible to defend against).

In contrast, everything was scaled in Oblivion: everything from end-game dungeons to mud crabs matched the PCs level. This caused several issues:

1) It homogenized encounters to the point that everything felt like a DPS calculation - if every enemy grows to match you, are any of them really that different?
2) It pushes the player away from non-combat growth - it's not so much the increased difficulty as it is the active discouragement from different builds (i.e., homogenizing builds/play styles).
3) It sabotages the internal game system of progression: it's fun to figure out how to optimize to handle a challenge, but when the conclusion is to simply avoid growing in level, it's a little disappointing. As a player, you're either "cheesing" the system, or you're choosing to ignore the optimized solution (and figuring that solution out is usually the fun part of the challenge).

The difficulty slider in that game is even worse - it just directly scales damage and HP. I remember turning it to max once; I ended up kiting a boar for about 5 minutes, and I realized it was just a "tedium" slider.


All that said, I think the best way to circumvent the problems with scaling involve different methods of increased difficulty. I think AI is the best way to do this: adding a handful of different raid types has already forced people to adjust their strategy, I think there's still a ton of room to explore there (especially with e.g. Changing targets, shooting at doors, etc). The launch sequence is another great step - it's a huge ramp up in difficulty, but it's an active choice with a tangible benefit. I think the caravan quests are a great prototype to a sort of mid-game progression step, with a ton of room to flesh out more significant events (rescue multiple/special colonists from slavers, assassinate a faction leader, capture a large established base, etc).


Re: other changes, such as the "man in black" thing I've read about - I think that's another example of over-scaling. If it happens with a degree of predictability, that sort of difficulty reduction just ends up making the rest of the game feel homogenized, or even "fake." Even if the random colony deaths can be brutal/frustrating, I think the game is much better if they exist; I still have fond memories of failing to build fire-breaks before rain was guaranteed, and frantically sending out most of the colony to try to contain huge brush-fires. If it rained, it felt like a miracle - now that it rains predictably afterwards, that type of story moment becomes meaningless.

[bbqftw]

With certain pawn count you are probably better off burning archotech legs than installing them.

For the vast majority of things in this game, combat utility gain does not really match raid point contribution (at least on respectable difficulty).

As a result, on merciless with clean play you end up facing 3-4x your numbers in midgame, which progresses to increasingly degenerate 10v1 odds late if you don't deliberately wealth control. With the new bionics tax its probably something closer to 13-15v1 odds if raid cap isn't met.

Bionics are good, but not good to the point that its worth adding another 3 raiders per cyborg

And you can make that calculation for most things, and they are found wanting.

[Greep]

Almost every game gets harder as the player progresses. DOOM, StarCraft, WoW, Portal, Street Fighter, Jagged Alliance 2, etc. It's an attempt to match challenge with player ability so the player is neither overwhelmed nor bored. But players don't tend to interpret, say, DOOM's increasingly difficulty fights as "punishment". Like, "Hey, I defeated a room full of 10 zombies, why am I now punished by having to fight a room of 15 zombies?" Nobody says this.

Here is the big misunderstanding I think:  Raid scaling by wealth is not equivalent to this scenario.  An equivalent scenario is if you scaled further enemies by, say, how much ammo + health + armor at a given time.  Most players would understandably find it odd if you took on a big bad baron of hell and, at the final end of the game, simply fought 12 basic possessed simply because you ran out of ammo and have to use a pistol.  What a let down that would be!  Conversely, you would be punished for doing well if you did well.

Rimworld's situation is slightly worse than that even, though: much wealth is not useful in a fight and it's unclear to newer players that wealth is the main determining factor.  Using the DOOM analogy, it would be as if enemy scaling also considered how many times you jumped in the game.  It would be a bizarre and pointless sense of progression, but a smooth one nonetheless that can be worked around in weird ways and hurts some players randomly (probably ex oblivion players  ).

The equivalent scenario here to what DOOM actually does would be timed based raids:  In DOOM the progression simply increases the further you play and in Rimworld it would increase the longer you play.

[spidermonk]

But players don't tend to interpret, say, DOOM's increasingly difficulty fights as "punishment".

Imagine DOOM with permadeath and when you pick a pack of shells threat level increases significantly. When you pick up a gun threat level increases even more. Now you no longer can grab all the ammo you see, and you no longer can pick all that shiny new guns. You have to think careful what to select and what not. And that's kind of frustrating, because you really want that shiny toys.

Now, even then you'll want to progress in that inhuman DOOM, because it has clear sense of progression - completing levels. In RimWorld your progress can be increasing level of safety and well-being of your colonists, and if not developing your base means more safety for your people, then you are very motivated to not develop it. But you really want to.

An explicitly articulated "notice" system sounds like a great solution though.

[JimmyAgnt007]

Since there is no BOSS fight or explicit end of the game, we shouldnt be progressing as if there were.  In DOOM, we can recover a lot more easily than in RW and there is an end point.  I dont like the 'play until you inevitably die' logic in a constantly escalating threat.  Id rather a level of challenge being maintained.

Also, some raids should be easy.  It happens.  But trying to avoid wealth to manage raids is not how it should be played.  Thats why they should be scaled with your capacity to repel them.  Give or take a little.

[bbqftw]

I also like how you insinuate how most concerns re: raid scaling are from theorycrafting or code analysis.

That is a very unfair characterization. I tested all my 1.0 games without deliberate wealth control, and without looking into the specifics of the raid formula.

However, after getting threats like 41 centipedes / 220 tribal sappers / 300+ insects  for 22 pawns, when early game my fights were more on the 3-5 to 1 odds, I realized there was something I didn't understand about raid scaling. It felt off. If you played it, it would feel off too. I hope.

Only then did I spend an evening dev testing calculating the contribution of pawns, adaptation, and animals. At which point I realize that wealth scaling to threat is actually higher than first order.

But I did not go into 1.0 thinking I needed to wealth control. Play experience taught me it is not only the most important thing, it is the only important thing.

This makes the game a lot less appealing to me, because wealth management is actually pretty simple (and yes it works, consistently smaller raids - 2-3:1 odds are the trickiest you will face on merciless if done right). In contrast, there's enormous space to optimize in economy management and its something I am pretty bad at compared to top players.

But all economy optimization does is make your game significantly harder.

[Pestilence11]

What about a "Useful Wealth" metric for raids? Since some colonists require lavish surroundings that raiders wouldn't find practical, why wouldn't only things they find useful be counted? Like for instance if you have tons of higher quality food, drugs, armour, weapons etc. just lying around then word would get out to other factions that your settlement could be a juicy target.

A system like this would allow players to keep themselves busy by further improving their colony rooms without being "punished" for it, without having to play on a lower difficulty.

Unless as JimmyAgnt007 above said, the idea is you're supposed to die eventually to finish the colonist's story and to move onto another. 

[Tynan]

If one clicks back and forth between the wealth and debug pages, one finds it difficult to say "raid strength isn't primarily based on wealth" with a straight face.

This happens because you're playing clean games without significant population losses. If a formula has several inputs, and you hold several of them constant (adaptation/population) while changing just one (wealth), obviously the output will respond to the thing that you're changing.

Try holding your wealth steady and losing 90% of your population in a battle and see how things change.

Also when I say this I'm not saying wealth isn't really important to threat strength, it is, for sure. But from many posts in this thread, it's obvious some people have the impression wealth is the only factor, or nearly the only factor. Many people here are suggesting things like losses-adaptation/recovery, population-based changes, time-based changes which are in the game already, which suggests that it's not widely understood that the threat strength isn't just a factor of wealth. That's all I'm trying to communicate. There's no value in taking an adversarial stance here, this should be a collaborative process.

just because certain raid point term is ∝ wealth * population, does not mean its not proportional to wealth!

I never said it's not proportional to wealth. I said, "In truth, raid strength isn't primarily based on wealth (wealth is a major component of it, but so is population including animals, health, recent damage taken, and some other bits)."

If it was, hypothetically, wealth*population, it would be proportional to population exactly as much as it is proportional to wealth. What I'm responding to is the general myth that it's based only on wealth or primarily on wealth, which is the premise of this entire thread and which is false. People just tend to notice the wealth effects more because wealth changes more, as I noted above.

Regarding the very strong threats you've mentioned in your extreme late-game Merciless run - I'd say that it seems you are in fact surviving these threats indefinitely, so far from being too strong, they're probably not strong enough. It's just that you've played so far (years and years of game time) at such a difficulty level that you're reaching a level of development the game can't match since it can't scale infinitely and we don't want to cap your progress or force an endgame either. This is another "easy to describe and hard to solve" problem.

Since there is no BOSS fight or explicit end of the game, we shouldnt be progressing as if there were.  In DOOM, we can recover a lot more easily than in RW and there is an end point.  I dont like the 'play until you inevitably die' logic in a constantly escalating threat.  Id rather a level of challenge being maintained.

Also, some raids should be easy.  It happens.  But trying to avoid wealth to manage raids is not how it should be played.  Thats why they should be scaled with your capacity to repel them.  Give or take a little.

"Play until you inevitably die" is not and has never been the logic behind the RimWorld design here, especially at non-extreme difficulty settings. The goal is to not bore people with trivial non-challenges, nor crush them with impossible odds, in a game with no fixed levels or story. This thread was about discussing ways to do that.

Scaling raids with your capacity to repel them would create the same complaints you see here to an even greater degree, plus bizarre exploits like "never build defenses, just build a mega-wealthy super-base and burn every weapon you find".

--

Jpagano: I think the Morrowind vs Oblivion case is really interesting. But they key here is that in a TES game the player can choose where to go and what to do. He can choose to enter a high-level dungeon, or not enter it. RW as it stands is more reactive - threats come to you, not the other way around. This has upsides and downsides.

E.g. Imagine in Morrowind if 70% of the fights were actually enemies tracking you down wherever you were and attacking, forcing a player response. How difficult are these enemies? It's a problem I don't think those games really have at all because the structure is fundamentally different.

I've thought a lot about this and come to understand that it would probably be better if the player was actually the one initiating the most dangerous events in the game, as it is with the end-game ship sequence in unstable build right now. DF also does this a lot considering how easy it is to wall yourself off from sieges, and how the player can choose when and whether to dig deep and reveal the greatest threats.

This would solve a lot of problems and allow us to put up a "menu" of challenges the player can choose to undertake or ignore. It's a direction I'm interested in, and one I've moved in with recent releases (e.g. quests), but it's a very different structure from what's there so I'm not going to retool the entire game right this moment near 1.0 release of course. But, along with "explicit visibility" as a mechanic, it's something in mind for sure.

--

This thread was going really well, I hope the spirit of the OP can be respected going forward. Specifically the idea of thinking about solutions and new options, instead of just pointing out downsides of the existing system.

[Polder]

Technologies don't give passive bonuses but allow the creation of new items and buildings. It makes no sense to increase raid strength based on progression through the technology tree.

[JimmyAgnt007]

"Play until you inevitably die" is not and has never been the logic behind the RimWorld design here, especially at non-extreme difficulty settings. The goal is to not bore people with trivial non-challenges, nor crush them with impossible odds, in a game with no fixed levels or story. This thread was about discussing ways to do that.

Scaling raids with your capacity to repel them would create the same complaints you see here to an even greater degree, plus bizarre exploits like "never build defenses, just build a mega-wealthy super-base and burn every weapon you find".

I didn't mean that's how it is, a constant increase in difficulty was mentioned and I didn't want the game to turn into something like that.  I love my long-term colonies.

Currently Ive been building an open colony with no killboxes, no meta managing of wealth and minimal turrets.  The result is that the raids feel a bit over powered.  Rocket launchers devastate where regular guns would have been exciting.  My mountain fort killboxes on the other hand would have absolutly needed that type of raid.  Thats why I suggested some kind of reaction to how raids were repelled.  If the turrets got the most kills, send in the rockets.  If its mortars then use drop pods.  Lots of walls, use sappers.  and so on.  let the player be forced to mix things up as they play.  let the challange be in adaptation rather than a pure increase of numbers.  When I did get smaller raids, in the early game, they were a lot of fun.  Having a running gun battle in the streets as people responded to the threat.

Not sure how good the 'burn all weapons' exploit would be in my system.  Hostiles would still show up and kill you if you arnt armed.  It would devolve into fistfights maybe but one manhunter pack and you are dead.  Different kinds of raids to deal with different kinds of exploits would be interesting.  The one exploit i thought of for my system is to only allow pacifists.  Since they have no capacity for dealing with raids, they would be immune?  Thats why I suggested some variance.  Also, there would be a min strength to each raid, even if its just a guy with a club.  Every system will have its flaws, I just think we need one with the fewest possible.  Currently, a dev mode delete of the most devastating weapons is what works for me.  Though my open colony is getting a perimeter wall and I will be adding more turrets.

Is there a difference in raid calculations for tribes vs colonists?  Primitives would have a harder time fending off raids.  Some people like to impose their own rules as well.

[Greep]

Well, my argument for a constant timed increase was simply that if you look at your colony fun points graphs, it is a constant increase anyways, superlinear even in most cases.  A timed increase that was linear would actually be more forgiving, while at the same time removing some metagaming weirdness.  Unless you're arguing both against the current raid system and timed increase, jimmy.

There could be an argument that maybe the different difficulties should actually have separate raid formulae altogether:  From the feedback I've seen in 1.0 thread, medium and under players don't really care for a constant increase at all, while rough and over do.  In which case a fully adaptive raid design like you suggest might even work best on medium and under.  Would be nice to see a poll on that.  I think people choosing medium don't care to metagame in the first place or at least only do so when it's overwhelmingly hard not to (e.g. no gold statues), so there would be no concern there.

Edit: meh, I've bickered enough, removed sillier parts of post.

[5thHorseman]

This would solve a lot of problems and allow us to put up a "menu" of challenges the player can choose to undertake or ignore. It's a direction I'm interested in, and one I've moved in with recent releases (e.g. quests), but it's a very different structure from what's there so I'm not going to retool the entire game right this moment near 1.0 release of course. But, along with "explicit visibility" as a mechanic, it's something in mind for sure.

This is a really interesting idea. Maybe quests to specifically anger factions? We were just talking about spicing up downed refugee quests. Maybe instead of refugees, they stole something big from another faction, and by going to get them you get the item(s) but also trigger - some time in the next 15 days - a HUGE raid.

[Tynan]

I actually think a constant timed increase mode would be interesting.

There's the problem of dead man walking states (e.g. you took damage and now you just have to wait 10 days until the next threat and then die, there's no way to recover).

There's also the problem of:
-How are caravan threats handled?
-What if you start another base or migrate your base?

Seeing these it's just a can of worms I chose not to open. Lots of more fundamental stuff to work on.