To RNG or not to RNG

[5thHorseman]

With very high skill play in X-com or FTL, you can win nearly 100% of runs you attempt.  Would you call those games 'non stop victory generators'?

And this is based off what evidence, your personal experience? When I play FTL I find I can rarely get to the endgame, there are too many variables that I can't control that affect the outcome. I don't tend to blame my failings on bad game design however.

My personal experience with 1000+ hours of FTL: Old "normal" (which was the hardest mode) I got to the point where I won over 90% of the time. New (as of a few years ago) "hard" (which is now the hardest mode) I win over 50% of the time. And most of those, I can see what I did wrong and why I could have avoided it. And I'm not talking "I could have known ahead of time that jumping to that star would have been a battle that killed me" I'm talking "I should have realized that focusing on shields was not the correct play against this ship."

And each game takes 2 hours or less, so losing isn't that big a deal. If FTL games took 40+ hours and then RNG screwed me (on those very rare occasions where RNG is the actual culprit) then I'd never have played it for 1000 hours.

[NiftyAxolotl]

Most players assume that a lost fight or crippled colony is game over. You can blunt some of the frustration by priming the player to keep playing:
- If colonists are kidnapped, pop up letters to represent the ransom quests. Now the player has a goal, instead of just a crippled colony.
- If thirty megascarabs are swarming through the defenses, pop up a radio message from Charlon urging the player to abandon the base and escape to the north. Now the player has an option, instead of just being doomed.

But these only work if threat scaling readjusts properly. If there's any permanent time-based increase to threats, then a temporary reprieve doesn't really help. The colonists' hedonic treadmill of increasing expectations and joy tolerance is already punishing of any setback.

[dearmad]

Most players assume that a lost fight or crippled colony is game over. You can blunt some of the frustration by priming the player to keep playing:
- If colonists are kidnapped, pop up letters to represent the ransom quests. Now the player has a goal, instead of just a crippled colony.
- If thirty megascarabs are swarming through the defenses, pop up a radio message from Charlon urging the player to abandon the base and escape to the north. Now the player has an option, instead of just being doomed.

But these only work if threat scaling readjusts properly. If there's any permanent time-based increase to threats, then a temporary reprieve doesn't really help. The colonists' hedonic treadmill of increasing expectations and joy tolerance is already punishing of any setback.

Well thought out and a good approach, imo. I already did this for myself in my own imagination as I played, but I'm sure others could use some prompting.

[anticrisis]

1. Should the game have a such a thing as bad luck outcomes that's not induced by some obvious, non-pressured, voluntary player decision? Or should I make a universal design standard that nothing bad ever happens unless the player actively induces it or makes some clearly-traceable mistake to cause it?

The best stories have foreshadowing, so a player can tell himself, "Oh, that's why that happened. Should have seen it coming!" For example, a pyromaniac is bound to eventually start a catastrophic fire, even if his first crises saw him attack potted plants. That's pretty perfectly foreshadowed, because there's an escalating consequence to not keeping him happy.

High-consequence RNG feels bad when it seems completely out of the blue. Maybe there's a way to build up to high-consequence outcomes? For example, treat "character causes friendly fire damage" as a bad consequence. The first time they're fighting, he shoots his buddy somewhere relatively inconsequential. This foreshadows that maybe he's not that great a shot, and we should be careful. The next time, it's worse.

2. Should I just ignore some classes of player feedback as simply not linking up with what RW is? Are some players worth leaving alone to try to make a game that's different from the usual assumptions? Even if it leaves them pissed off because they intepreted a story generator as if it were a skill test?

My experience with the game is that it requires and rewards skill, such as build order, planning, work assignment, etc. So I think this question is too polarized. High skill games create a sense of mastery, which is undermined by too-severe random effects. RimWorld is a high learning curve and highly rewarding game. When RNG is manageable and somehow "fair," I enjoy it best.

3. Should players be able to consistently avoid losing people/resources even at high difficulty? At any difficulty?

4. Is there a way to set expectations (relative to the whole game, or relative to a given difficulty level) to encourage players to accept some degree of randomness to game outcomes? Or will they always reject this randomness and demand to be rewarded in accurate proportion to their skill/effort?

Lower difficulty should have a slower ramp-up to high-consequence RNG events. Easy to Medium could be framed as "random events will throw wrenches into your works but will be relatively easy to overcome." Hard (or hardest) mode could be framed as "random bad stuff will happen and it will be incredibly hard to deal with."

[mcduff]

The best stories have foreshadowing, so a player can tell himself, "Oh, that's why that happened. Should have seen it coming!" For example, a pyromaniac is bound to eventually start a catastrophic fire, even if his first crises saw him attack potted plants. That's pretty perfectly foreshadowed, because there's an escalating consequence to not keeping him happy.

High-consequence RNG feels bad when it seems completely out of the blue.

I think this is important, and a big difference between "I've ballsed this up" and "wtf?"

Some RNG is fine, but I think a scaling mechanism for consequences rather than just "haha you're hosed for no reason now" feels better.

[mcduff]

Most players assume that a lost fight or crippled colony is game over. You can blunt some of the frustration by priming the player to keep playing:
- If colonists are kidnapped, pop up letters to represent the ransom quests. Now the player has a goal, instead of just a crippled colony.
- If thirty megascarabs are swarming through the defenses, pop up a radio message from Charlon urging the player to abandon the base and escape to the north. Now the player has an option, instead of just being doomed.

But these only work if threat scaling readjusts properly. If there's any permanent time-based increase to threats, then a temporary reprieve doesn't really help. The colonists' hedonic treadmill of increasing expectations and joy tolerance is already punishing of any setback.

Well thought out and a good approach, imo. I already did this for myself in my own imagination as I played, but I'm sure others could use some prompting.

In general I think one of the things that isn't quite there yet is the balance of "some good, some bad" events. I'm not sure whether the move to get around "exploits" of the mechanics has left some events being more bad than good, with rewards out of balance to the risks.

I feel this way about traits too, tbh. I think that the traits, character backstories etc in general could do with a pass to make them so that the ones that give a nerf also give a buff, and vice versa, so you might get characters that are more difficult to play with but you never get ones that are just straight up *bad*. Chemical fascination, for example, has been called an absolute nightmare, and it is, but there are things that could be done to it to make the endless cycle of bingeing and withdrawal less of a ballache.

[zizard]

Indeed, the game has almost no punishment for playing glacially, waiting around for only good pawns. So accepting something "below ideal" is simply a long term loss. It's almost like a punishment for lack of infinite patience. Perhaps the lowest risk play is to be extremely picky with pawns while focusing on research, since AFAIK there is no research point contribution to raid strength.

[Ramsis]

I would like to take a moment to remind everyone to take it easy and relax a smidgen. I'm seeing a bit of anger cropping up here and there and want to mention everyone is entitled to their opinions on our board and we support all feedback good or bad.

[Aerial]

The core issue, to me, is the colony death spiral.  I play on middle difficulties because I want a survival story but I don't want to have to be a meta gamer or consummate micromanager just to survive raids or manage mental breaks.  I have about 1000 hours in the game, over the past 3(?) alphas. 

Right now it feels like *everything* in Rimworld is there to kill you and nothing you do will really help (much).  Previously, raids were kind of simplistic and you could build a defense setup that would soften them up a lot and make the outcome of the battle likely to be in your favor.  I'm not sure why this was a bad thing, at least on lower/middling difficulties.  It took time and resources to build, but it felt satisfying to finally be "safe" (except for the sieges and poison/psychic ships).  And being able to achieve that relative "safe" provided an emotional pad for me as a player that made a lot of the negative mood issues (like the pawn being mad that they ate lunch out in the field instead of at a table) and injury issues (like having an armored pawn's eyes scratched out by a squirrel) more palatable. 

Now, with the raids seemingly structured so that you'll never be safe and largely have to risk your colonists' life and limbs for every conflict, those bizarre negative moodlets and RNG killer squirrels just become contributing factors to the death spiral.  They're discouraging rather than silly/stupid "what's the X thing that ever happened to you in Rimworld?" story fodder.  That saddens me, because it was the mix of survival and silliness that drew me to Rimworld. 

More positive events would really help.   Getting rid of some of the really unfair-feeling events would also help (slaughterer mental break, rampant illnesses on caravans, for example).  But mostly, I, as a player, need to feel like my little colony is more than one bad break away from total destruction every single time I sit down to play. 

Sorry, that wasn't directly about RNG, but I don't think the real issue is RNG itself (which is largely a good thing - Randy Random is popular for a reason) but instead about how easily bad RNG can destroy a colony despite the player's best efforts. 

[Ser Kitteh]

What exactly is "unfair?" It's a word that means a lot of things to different players as far as video games are concerned. I somewhat understand sudden plagues, especially during caravans when you haven't much in the way of meds. But again, with good meds and the odd pop of penoxydiline, these are all preventable. Ah, but neutroamine is something very much RNG based, and you really can't rely on trade caravans to bring them so that's understandable too.

It's quite likely, that a player can set up a proper medicine production line before the first year ends but they can't do anything with it because one resource is pretty much always out of reach. You can plop your colonists in a cryptocasket too, but that can spell the end of your colony if you aren't prepared either.

[sadpickle]

I will savescum if a raid goes horribly wrong and a large portion of the colonists are wiped out. For me, it's time invested in a colony. I tend to build large and set long-term goals. If that all ends abruptly, or I get badly gimped, because some raider had a noob tube (Doomsday launcher) then I am not entertained. I'm just enraged.

It's not all bad things. I had a colonist get brain damaged and spent a long time debating what to do. She was the fiancee of another, original colonist, and I knew in time he might break it off. But I opted to put her on luci and she got better later that year. Then they got married. It was pretty great, and memorable. This sort of emergent gameplay is highly satisfying. There was a setback, and a cost incurred to recover, but in the end it worked out.

[EvadableMoxie]

What exactly is "unfair?" It's a word that means a lot of things to different players as far as video games are concerned. I somewhat understand sudden plagues, especially during caravans when you haven't much in the way of meds. But again, with good meds and the odd pop of penoxydiline, these are all preventable. Ah, but neutroamine is something very much RNG based, and you really can't rely on trade caravans to bring them so that's understandable too.

It's quite likely, that a player can set up a proper medicine production line before the first year ends but they can't do anything with it because one resource is pretty much always out of reach. You can plop your colonists in a cryptocasket too, but that can spell the end of your colony if you aren't prepared either.

No one event is really unfair.  You can beat any one event in Rimworld with proper play.  So yea, plague is fine, when everything is going fine and you have doctors and a hospital and medicine. Sure.

It's when the enemy one shot your best soldier and as a result you barely held off a raid but your best doctor lost an arm, half your bedrooms burned down, and another colonist broke from observing corpses and being injured and decided to go wildman and you don't have 6 taming to get him back.  THEN you get hit by a plague. Now you're struggling with manpower issues to fix all the broken stuff and keep everyone fed and treated.  There's no way to get back to where you were before by the time the next raid comes.  This raid is also ugly as a result.  Sure, you hold it off but a pawn died and another lost a leg and your colony took even more damage. Then you get a manhunter pack, a mad boomalope and a toxic fallout before the next raid.

That's what the colony death spiral is and what we're talking about when it comes to RNG. It's not that singular bad things can happen. Any of those one events, in a vacuum would have been fine.  It's the way that things beyond your control happen that lead into more things beyond your control and how each one just makes you a bit weaker and a bit less able to deal with the next one until your colony inevitably dies.  You can see this coming, know it's going to happen hours before the final event ends you, and yet still have absolutely no way to prevent it. 

[Triade]

You can see this coming, know it's going to happen hours before the final event ends you, and yet still have absolutely no way to prevent it.

This is the quintessence. The stories aren't fun when they are predictable, and currently it's very predictable that when an event (or multiple, parallel events) did sufficient damage, you will lose. Doesn't help that the "sufficient damage" is randomly happening. Events and stories in real life have much more to them than the consequence. Most events happen because of multiple reasons that might only become noticeable in hindsight. This component is in most cases completely missing in Rimworld.

Basically, the "why" is sometimes non-existent.

[Robc]

RE: Lost hand discussion.

So I start a new tribe, roll for a decent starting crew and a day 1 social fight has 1 pawn bite off the hand of another.

It's an interesting isolated case, because here you reject the "lost your hand on day 1" situation as simply bullshit. Which is reasonable from the "skill test" game frame. After all, if you lose because of random events, it's a pretty shitty skill test.

But in a story, losing your hand day 1 is actually a really common sort of thing to happen.

Juxtaposition of a negative event to the game timeline is really important.  Early - player has little time or emotion investment, decision = restart.  Later - figure out a way to manage without that hand -or- reload.  The player restart decision does not mean that you have a faulty game design for story telling, it is a natural consequence of the time (economic investment) it takes for human connection to the unfolding story.  Real life story telling does not allow for the economic analysis of whether a restart is a good idea, its not possible.  We cannot achieve true story telling because we cannot prevent the player from periodically calculating his or her investment in each story.

The reroll decision is an identical economic analysis.  Game story telling has to cope with the player doing economic analysis on events that they perceive to be significant.  Players do these calculations all the time (early, mid, late) without realizing it and occasionally they decide to restart or reload.  Every player is going to calculate their cost/benefit differently and make their own decision depending on their own perception of individual events.  Players in 'Flow' or deeply immersed will not stop to do the analysis very often and this is the holy grail, some players will almost continuously analyze the economics of events because they are significantly driven by the achievement motive. 

We can achieve something like conditional storytelling.  The player continues to invest in the story and does not reload or restart as long as no event triggers the economic analysis to be done and/or doesn't trigger the analysis to 'fail'  ie "That shouldn't have happened, it shouldn't have been part of the story, I want to start again."

An option to mitigate restarts is to tilt events toward the positive early, or to limit perceived 'catastrophic' events early and tilt toward more balanced or negative events later.  RNG is replaced with a progression influenced RNG, I expect many will hate this approach... it smacks of the skill based progression you were describing and that is done well by other games, so don't copy them.

Option B, which I like better.  Continue the education process that describes RW as a world in which you will occasionally be punished by ill timed RNG events.  Establish a player culture (which you have largely already done) of rewarding stories of overcoming unbelievable odds.  The hero's journey is alive and well in RW precisely because of the event dynamic that causes pawns to lose hands etc., but not all prospective heroes get to tell their story, some die.  Reroll, rinse, repeat.  Players will by all means take their frustrations to the forums and point vehemently to their own economic analysis of the situation... many will holler but then they will reroll, rinse, repeat.  And that's ok, no actually its really good.

I think that you have invested insufficient time in your "you lost" screen and it contributes to the problem, I think that few of us read it anymore.  Collect a few specific sentences about the unfolding drama that has been played and share them with the player on the you lost screen... help the player to celebrate the tragic loss.  Finish the story as an amazing tragedy!!  It's really really ok to lose especially at the higher difficulties.  This screen in particular is your opportunity to help remind the player that they really prevailed against some amazing odds, help the player to remember all the events that they overcame along the way... don't just drop them with a message that essentially says, well you sure sucked there didn't you... you lose.  Specifics about this particular player's story and this particular ending are important in your losing screen, yes this is an investment of limited coding time and few if any other games do it.  RW is different, this would be a really good investment imo.

By the way, while I am ranting, I think advising players to adjust their difficulty level when bad things happen and it frustrates them is not a good option.  You are inadvertently challenging their past decision making and their economic analysis.  As you well know, people tend to go way out of their way to defend their past actions and current beliefs.  Many, in fact most veteran players, already understand the consequences of their past decision and don't want to be reminded of it, especially when they are fuming about that very thing. 

I don't know that any solution to the dilemma here exists, but one option that comes to me off the top of my head is to describe your game tiers as story genres.  The more difficult tiers are "short story" generators that are very likely or almost certainly going to end in tragedy, that's the point, this is the horror genre, or the tragedy genre, the hero loses.  Ask the players to tell us about their tragedy!  Help them tell the whole story with your closing screen.  And have the player revel in it, make them excited to read the closing screen and what they accomplished.  Man that was fun right up until the end. 

For the easier levels these generate longer stories with many setbacks but the player has at least some hope of rescuing the princess.  This is the adventure genre, the hero might just save the day.  And you have the base building genre which includes many of the other traditional story telling themes.  Use your closing screens to help tell individual stories about the play through, invest that development time.

My suggestions are not design suggestions but player communication suggestions.  I think you are pretty close to an optimal design already.  Spend your design time making sure that individual events don't become 'nonsensical' as one commenter put it... because the nonsensical events immediately break immersion and cause players, even the most story oriented players, to do an economic calculation.  Identifying and eliminating nonsensical individual events is a difficult enough task without questioning your larger objectives (as you seem to be doing with the OP).

Now on to your questions:

1. Should the game have a such a thing as bad luck outcomes that's not induced by some obvious, non-pressured, voluntary player decision? Or should I make a universal design standard that nothing bad ever happens unless the player actively induces it or makes some clearly-traceable mistake to cause it?

Stories have unexpected events, so should the game.  RNG is another word for 'unexpected.'  Keep them.

2. Should I just ignore some classes of player feedback as simply not linking up with what RW is? Are some players worth leaving alone to try to make a game that's different from the usual assumptions? Even if it leaves them pissed off because they interpreted a story generator as if it were a skill test?

Note the frequency of the feedback, but stick to your knitting.  You have a good design concept that won't please many.

3. Should players be able to consistently avoid losing people/resources even at high difficulty? At any difficulty?

No.

4. Is there a way to set expectations (relative to the whole game, or relative to a given difficulty level) to encourage players to accept some degree of randomness to game outcomes? Or will they always reject this randomness and demand to be rewarded in accurate proportion to their skill/effort?

Expectations are re-calculated by players after each time they do an economic calculation as I describe above.  It is impossible to keep players in immersion all the time, but you should try.  When they do break, and do a calculation, how frequently do they decide that this is an event they have to work through and how frequently is it an event that forces a re-roll or a re-load?  It's hard to collect this data from the forums and it might be hard to design a play test.  Don't try to know the economic value that each player puts on the consequences of an event (should I re-roll, should I reload?) because I think it is hopeless to find a consistent pattern.  Make sure that on whole most players choose to come back and try again so that they keep playing for a long time.  Concentrate all your effort on removing nonsensical events to limit the number of times that players even do an economic calculation at all.  Accept that some players, very vocal ones perhaps, are so achievement oriented that they will do an economic calculation far more often then your average story oriented player, and that's ok, you just can't really design for it imo

Secondly, follow through at the end of each game with a reminder that the player just took part in a story with a __________ ending (tragic? or whatever)  There was story value in the play through however it ended, but you need to help the player to remember they had fun even as they are frustrated/ agitated/ or otherwise put off by that particular ending.  The value of the story just told should not have been lost.  I think you are missing an important opportunity here.

Robc.

[Raikoin]

Signed up just to throw some ideas/opinions on this. So as a quick disclaimer, I've only skimmed this thread but I'd like to throw in my take from the more recent changes and the general trend I've felt over the couple of years I've been playing.

So, generally I play at the 'Rough' or 'Intense' difficulties and would often keep a colony going longer instead of rushing for the ship. Natural this means I also tend to build up wealth and what not past what you actually need, obviously I've stopped this as of recently now that hording large amounts of silver/items/whatever is often an unavoidable death sentence (across the board in terms of difficulties from my own testing). I also tend to play on the 'Cassandra Classic' storyteller and have always shied away from 'Randy Random' for a key point; I enjoy the risk management element of games and find situations far less fun if the risk element is forced on me with zero input which Randy is supposed to do more. Effectively I'm fine dealing with the outcome of taking a risk and rolling the dice if I chose to roll said dice. The 'chased refugee' event is an excellent example within Rimworld of this where you get a basic bit of information and decide to take the risk or leave it. There is a decision and an aftermath you must deal with which may or may not be worth the reward but you can choose whether or not to risk it.

As more content got added to Rimworld there was a slight skew towards certain things can potentially be game ending if it comes up at either the wrong time and some of them were not something the player had any real control over. For example infestations can be avoided by not digging into a mountain and short circuit effects minimised by not leaving batteries connected. This might limit you in terms of access to resources or a power safety net but that's part of the risk/reward management. Raids are one of the odd ones here since the scale of a raid is basically a punishment for amassing wealth/resources/whatever past what you can defend but the raid itself rolls so many dice that they can nullify planing. Similarly 'Cold Snap' is a non issue a lot of the time but if poorly timed could cost you several colonists, or even the colony if coupled with other non-serious events.

As I play more of 1.0 and test each set of changes (not every minor patch obviously) I feel more and more like RNG swings are getting to high to account for and makes any form of 'permadeath' runs luck based, or at least 'winning' is semi-random. The game has hit a point where it is possible to play perfectly in every measurable way but have no chance of winning, or more accurately, your planning, fail-safes and accounting for future issues suddenly counting for nothing because the game said so. Notable recent cases for myself include; an enemy 1 shooting skill pawn out in the open downing 1 and killing 2 pawns behind cover within 7 shots where none of the pawns attacking them (shooting skills 9, 8, 0) landed a single shot, and the odd case of a mad rat downing two pawns (who both ended up with infections), one of which was armed with a knife, which were both attempting to fight it in melee. Both of these effectively ended an early colony not because I hadn't planned, not because the game was being unfair, but because RNG had decided that. The fact that this can happen at any point in a run means that I've found myself wonder less 'what's coming next, I hope can I deal with it' and more 'what's coming next, I hope I don't random have people die'.

These swings don't happen every time but they've happened frequently enough recently that the game feels unrewarding as a player. I suppose it feels more like the game is telling a story to me and if I try to participate in the story my hands get slapped away for changing how the game wants the story to play out. It's not about planning in order to overcoming challenges as opposed to hoping that the random dice rolls that have been forced on you have a good outcome. Basically I don't feel like I'm playing one game so much as playing small segments between major events which might randomly wipe me out despite my best efforts.

Plus it feels like I'm slowly having my options reduced. Things are being added and tweaked to stop me playing how I want to play and play in one specific style. I can't say for certain but it does feel like any effective playstyle or strategy that's not 'the one' we're expected to be using has been treated as an exploit to be patched or otherwise stopped, or at least crippled enough that it's not worth trying.

So TL:DR:
RNG has a place in games and is a key element of Rimworld. However, large swings coupled with the RNG being forced on the player to often makes the game devalue player actions and appear unfair. Personally I feel like Rimworld is headed a bit to far towards RNG being dominant and would like to see more options to take the risks rather than RNG results being handed to the player. This isn't easy to do for many events but not all events need to be addressed. Similarly events that do nothing but randomly annoy the player and inconvenience the player should have a 'counter' that can be invested in. For example heaters and coolers let us offset the extreme temperature events at the cost of resources and power which is great and something you can work towards as a plan or safety net, this gives players the illusion of a choice despite the event happening regardless. It might still bite you but there are steps you can take/have taken to lessen the impact.