The balancing process

[Tynan]

I've been reading feedback pretty intensely in the last few days. Most of it's positive but some players are writing concerns about things like animal taming maintenance or turret maintenance. I figured it might be worth offering some of my thinking on the balancing process. It could enrich the feedback process if you guys knew *why* I was making some of these changes. I can't guarantee I've got everything right (yet), but I can guarantee there is a thought process behind every change.

The first thing to note is that nothing is final yet. The whole reason we do unstable builds is to get feedback so that we can fix the problems. So if there's something you don't like in the unstable build, don't worry too much - if it does turn out on broad testing to be a bad idea I certainly intend to change it. And in fact I've already adjusted quite a bit in the few days since unstable came out. It's best to not get sad about something that may not even happen.

The second thing is that theorycrafting is really dangerous. Theorycrafting is when someone just reads or thinks about a game, imagines how it might play, and gives feedback based on that without actually playing it in significant depth. The problem is that games are frightfully difficult to imagine and hard to predict from a description. Even professional game designers with 15 years' experience can't theorize accurately at how a game design will play. I can't! So we use tons of coping mechanisms (constant playtests, short iteration cycles, unstable builds for feedback) to escape from our own mental incapacity. So it's best not to get sad about something you've just imagined - it may not turn out that way at all in real play.

The theorycrafting point I think is especially important on something like the animal training maintenance. Consider this: We all know they need maintenance now, but how much maintenance do they require? There's a huge difference between needing to re-train each animal every 4 days and needing to do it once every 60 days. But from the changelist, nobody can tell this since no numbers are written. Which means that theorycrafting about this change requires simply inventing a certain balance point - which could be off by more than 10-fold! From this alone, any imagined outcome from this must be suspicious since there's a really good chance it's off by 10x or more in terms of impact. Even those who think training maintenance is a negative change might be okay with it if it was 10x milder than they're imagining.

The actual intent with this change is specifically to make it so that super-swarms of attack animals are still viable and still powerful, but require commitment. In B18 you can have 100 attack boars for almost free. They feed themselves automatically by eating grass. They haul stuff for you, rescue your people, fight your battles with zero risk to colony or colonists, provide meat and leather (even when killed in battle), reproduce themselves for free. All this can be done for the price of training each (free) boar once. It's an insanely OP strategy in B18 to the point of being quasi game-breaking.

Animal training maintenance is quite mild; it should be barely noticeable at "normal" animal herd sizes and even if you have a mega-swarm it just means you need a few dedicated beastmasters to keep them all together. A few dedicated beastmasters is still a a small cost for the benefit of a mega attack animal swarm, it's still a bit OP compared to the core strategy of straight-up gunfighting. (Though I still plan to watch for more play stories about this and see how it really plays when someone tries it, the balance can still shift either way.)

---

Regarding how I'v approached balancing the game, here's one of the ways I see it. The way Beta 18 was, we can imagine there are 7 player strategies. Label them Strategy A, B, etc. What we had was this:

B18
Challenge level: 6
Strategy A strength: 9
Strategy B strength: 8
Strategy C strength: 6
Strategy D strength: 5
Strategy E strength: 4
Strategy F strength: 3
Strategy G strength: 2

A few observations about this:
1. People love strategies A and B. They're super strong! They always work! They give you what you want, which is victory.
2. People don't even think about strategy F or G. These are newbie traps. You try them once, get your ass handed to you, and never touch them again. Bad for newbies, irrelevant for everyone else.
3. Much of the game mechanics are wasted. Since only 3 or 4 strategies are even viable, we've got whole game systems supporting strategies EFG which aren't really being engaged by players.
4. There's not much choice. If you want to do really well you pretty much have to use A and B. If you want to survive at all you can do a few more things, but you have to force it.
5. Strategies A and B are really easy, so you don't really have to engage the game much to play them. Not much risk, not much drama, not much thought.

Overall it's not a great situation. But how to remedy this?

Well, we could power up the challenge level to 9. Then strategy A would be nice and challenging, solving problem 5. But we've now totally obsoleted all the other strategies even more. There's even less choice; problem 4 gets way worse.

It's impossible to power all the strategies up to 9; there are inherent constraints in the game that make this impossible in some cases. E.g. if one strategy is "open field melee combat", it's almost inherently symmetrical between player and enemies; there's no elegant way to make this favor the player more. There are other constraints like, "does it make sense thematically" or "is it intuitive", etc. All these constraints are the fundamental challenge in balancing.

What I've tried to do is rejigger things so it's a bit more like this:

1.0
Challenge level: 5   <--- reduced slightly
Strategy A strength: 6    <--- nerf but still OP
Strategy B strength: 6    <--- nerf but still OP
Strategy C strength: 6    <--- the rest are unchanged
Strategy D strength: 5
Strategy E strength: 4
Strategy F strength: 3
Strategy G strength: 2

Some observations on this:

1. The old strategies that everyone loved are now nerfed! But...
2. The whole game challenge level is lower to compensate, which means...
3. A bunch of previously useless/newbie trap strategies are now viable.
4. It's still not perfect because it can't be due to the abovementioned constraints. We can't freely turn these dials. Some strats are still better.
5. But, overall, there's more choice, more variance. The player can, for role-playing, situational, or personal preference reasons, succeed in more ways. There are more ways to design your base, more strategies, more variant stories.

Basically what I'm getting at is that sometimes good game design really does require nerfing stuff that players previously liked to do. But if you evaluate the game from the point of view of a new player, instead of from the POV of someone who had a valuable strategy taken away, it's obviously a better game.

---

Looking at other specific cases:

---Turret maintenance is a targeted resource sink for late-game killbox-heavy colonies. The idea is that killboxes remain perfectly viable, but they are now an *economic* solution to *military* problems. Which I think is interesting.

Another goal with this was to minimally affect colonies who use fewer turrets. This is why I didn't just debuff the turret straight up. I want turrets to be useful even when there's just one, but without some other cost there's no way to do that without making turrets OP in large numbers.

So the turret maintenance is quite cheap, and takes several battles to even kick in for the first time.

Another goal was to make it possible for us to put in more powerful turrets (autocannon turret) without totally breaking the game. The mini-turret was already OP in B18, there's no way we could add an autocannon turret and keep it balanced without some sort of structural disadvantage. So turret maintenance allows more powerful turrets, further emphasizing the economy <-> combat relationship.

Finally there's a high-level issue with late-game colonies getting super ridiculously rich. Turret maintenance forms a long-term late game resource sink. Also note that the ship is a lot cheaper to build now, which opens space for this.

-Like I mentioned, animal training maintenance is a targeted change specifically to bring the "mega animal swarm" strategy somewhere in the neighborhood of a reasonable level of effectiveness.

It also addresses the late-game resource overflow issue. You can eschew turrets and instead us an animal mega-swarm, but now you need some good handlers and a good amount of food to keep all those animals trained and healthy. Again, perfectly viable, but no longer trivial.

And now, since the overall challenge level is lower, some more basic strategies should become more viable. I'm talking about things like "build sandbags and just fight them in a gunfight" or "draw them indoors and melee their asses".

Other changes relate to that too. For example, armor is now a chance-of-damage-cancel instead of a damage reduction. This means there are less wounds, but the wounds you get are significant. But, medicine is spent per wound, so this reduces time and medicine spend tending wounds, which on the econ side makes straight-up combat more viable. It also means that if you can get some really awesome armor, sending melee fighters to actually fight should be more viable since there's a real good chance you can win without getting hurt, as opposed to previous builds where you might win but you'd have a bunch of damage-reduced (but still bleeding) wounds - possibly on your eyes or brain.

There's a million more relationships like this too.

----

There's lots more to write. In fact I could probably do a book on this (har har) but I figured this is enough for one day. I don't think everyone will agree with my thinking, but I figured everyone would at least understand the portion of it I'm capable of writing down here. Please do keep on with the feedback!

[ChJees]

-snip-
Other changes relate to that too. For example, armor is now a chance-of-damage-cancel instead of a damage reduction. This means there are less wounds, but the wounds you get are significant. But, medicine is spent per wound, so this reduces time and medicine spend tending wounds, which on the econ side makes straight-up combat more viable. It also means that if you can get some really awesome armor, sending melee fighters to actually fight should be more viable since there's a real good chance you can win without getting hurt, as opposed to previous builds where you might win but you'd have a bunch of damage-reduced (but still bleeding) wounds - possibly on your eyes or brain.
-snip-

In practice this means the Feral Nailgun from Rimsenal no longer will lobotomise my fighters as long as they got armor!

Though it looks very odd to see normal clothing deflect attacks.

[Tynan]

Though it looks very odd to see normal clothing deflect attacks.

That was one concern about the damage deflecting system, for sure. I decided to leave it in for now for simplicity's sake and because it's hard to think of a better alternative that doesn't jam complexity into the game for very little benefit. Besides just zeroing out armor for soft clothing, I suppose (which also isn't satisfactory).

[Alenerel]

That was one concern about the damage deflecting system, for sure. I decided to leave it in for now for simplicity's sake and because it's hard to think of a better alternative that doesn't jam complexity into the game for very little benefit. Besides just zeroing out armor for soft clothing, I suppose (which also isn't satisfactory).

I have 2 points:
- Maybe make the cloth damage reduction only
- Even tho your long post has a point, I cant help but to feel that it was an excuse to plug in your book.

[Mihsan]

That was one concern about the damage deflecting system, for sure. I decided to leave it in for now for simplicity's sake and because it's hard to think of a better alternative that doesn't jam complexity into the game for very little benefit. Besides just zeroing out armor for soft clothing, I suppose (which also isn't satisfactory).

What if different layers of armor could work in different ways? Like bottom layer softens damage while medium/top layer deflects it.

As alternative: change graphics/sound for it so devilstrand shirt deflecting damage does not act like bullet just hit metal wall.

[Syrchalis]

I really don't think it bumps up complexity too much to just SPLIT armor and clothing into two different categories, it even makes sense in a way. Good clothing (think hyperweave) can reduce how severe wounds from bullets and other weapons are - kind of like silk or Kevlar. (Silk could let you survive a duel with guns back in the day when those were a thing).

And armor can outright protect you from injury, at least on the body-parts it protects.

More importantly - what does this mean for the player?
Clothing:
1. Avoid the "naked" debuff
2. Help with temperature
3. Reduce wound severity on basically every body part
It makes perfect sense, because even low quality, bad material, worn-out clothing does #1. Good material, any quality clothing (wool) does the #2. But for #3 you need high quality, rare material clothing.

Armor:
1. Protects vital body parts

So if you have important pawns -> Get them armor asap. While clothing is more of a slow upgrade process from "not being naked" to "actually helps with combat".

[Jibbles]

This might be silly.  I've always wondered if you had the time to play several games yourself for each version, and how long your colonies typically lasts?

I know you got your reasons for all the changes you do. I usually question those changes if I don't agree with them but all I can do is speculate.  So I appreciate that you took the time to share your perspective on some of the changes you made for balancing.  Yes, you should get on that book sometime 😉

Sorry, not sure how to word the this topic below so apologize if I ramble. pretty tired hehe

Could you give some insight on how you'd expect colonists skills/traits/health/capabilities to be generated or acquired and why?
I've seen skill decay.  I'm more interested in the time it takes to level up skills, and if those stats make significant changes.

In previous versions, my colonists skills took very long to improve without neurotrainers.  Pawns being incapable in multiple fields were frequent too; making micro a bit annoying. Incapable of violence is a neat one but...I swear, I come across them way too often. I would build ships for them if I were to come across them once every 2 games IF that was the case.  I had no regrets in stripping them in the lake they crashed in to bleed and drown in b18. This does not match with how I treat my colonists, and how I imagine their colony to behave.<--my main issues

We  have all that going on, but stats in skills don't guarantee much, or not consistent in previous versions.. Such as furniture quality, shooting etc.  Then we have other things affecting those jobs they're supposed to be good at, such as sight, manipulation etc.
My main point is that I feel things are getting overloaded in this area.  I imagine you try to make all these stats to intertwine so the pawns feel organic, but my experience is that most of them feel the same... I mean, if a pawn's stat is around 14 then I immediately look at their age, most of the time they were old af or almost useless.. The incapables are taking over, and I'm getting disinterested with my pawns cause of that.  I wonder if you made any changes in this update, or if there are any plans to do so.

[Tynan]

- Even tho your long post has a point, I cant help but to feel that it was an excuse to plug in your book.

FWIW, I make basically no money off my book. There's zero chance that the effort to write that post was worth the cup of convenience store coffee I might earn from related book sales. This is about RimWorld and the RW community.

[jecrell]

I just wanted to thank you for taking the time to write such an informative look inside the design process. I thoroughly enjoyed reading it. I have always been envious of your designs, and I hope for more articles like these in the future.

[DariusWolfe]

Hey Tynan,

I personally really appreciate when you take the time to write stuff like this out. It makes me more likely to trust that you have reasons for the choices you make, even when I don't understand them, and it helps me to understand them. It also helps me to know what to concentrate on when I'm playing, so I can give more informed feedback.

One note is that I'm personally going to keep theorycrafting; I think it's likely everyone will, but for me, I often have long periods of time to browse the internet but not to play, so it helps pass the time and keeps my engagement high. But since you pointed this out, I'm going to start explicitly labeling any theorycrafting I do, as well as the actual play feedback, so that when you do come across my comments, you know what to ignore and what to read.

[lancar]

When it comes to armor, why not just use both systems at once? It's fairly common in RPGs that a piece of armor has BOTH deflection to avoid damage completely and some damage reduction to soften the impact when a hit does actually land.
It'd make picking materials to craft armor more interesting, too.

[Ser Kitteh]

I'm not gonna pretend to understand everything Tynan just wrote. I'm not a game designer, he is. But I appreciate you explaining your reasonings, even if the minor details escape me.

I also think that we should point our problems in the 1.0 thread, instance of clogging this thread about what we think we should do. I trust Tynan would do what is best for Rimworld, even if I myself don't agree with it.

[Falcon_the_Slut]

Thank you Lord Tynan for your thoughtful post and insight.

[TheMeInTeam]

This is really insightful and I appreciate it.

I'd like to note one inconsistency - early in the thread you correctly point out how dangerous theorycrafting is.  It gets especially more dangerous as more mechanics interact with each other too...definitely not easy.

On the flip side, it's not so easy to assign numerical values to strategy A vs C or G for example.  Lots and lots of players relied and likely will still rely on turrets. 

As a Randy/extreme/tribal/1 pawn start player on B18, I wouldn't grade turrets as a 9 strategy.  I'd have placed them closer to "newbie trap", something ~4 or maybe a 6 if used as a mobile distraction.  Research burden, power burden, mathematically not much better than steel deadfalls in terms of steel consumption : raid size defense.  There were (and remain) options with substantially less resource cost and, if you micro well, equal risk.  That was true early (years 0-1), mid (years 2-3), and late (years after 3) game.

When I see a turret nerf, I see a 4 strategy getting nerfed as if it's an 8 or 9 strategy (I have similar feelings about mountain vs field bases, perceived strength vs actual strength under decent optimization is massively different).  While the compensating nerf to raid size in general might still leave it as a ~4 strategy, the fact remains that perceived strength vs actual strength is pretty disparate.

There is a possibility my own estimates are off in terms of numbers, in fact this is likely to at least some degree and there are almost certainly better players out there than me.  However, there is evidence to support my position in this case: experience using alternative strategies that cost 1000's less steel/dozens fewer components to defeat the same raids with the same damage taken on the game's highest difficulty level, consistently.  This evidence is why I have some confidence in the assertion despite the B18 ubiquity of turrets. 

Melee remains a serious challenge to fix in my mind.  Shield belts faster/easier certainly helps, but the fact remains that on high difficulty where you're badly outnumbered there are multiple raid types where it's not viable (mechanoids, sieges, to a lesser extent sappers)...and the fact that all raid types can be handled by good gun play micro at all difficulty levels with high consistency puts this in an awkward spot.  This seems particularly hard to solve - I can't even imagine a solution to this that won't break things even at a theorycraft level, and that's usually more optimistic than actual implementation when it comes to one's own suggestions!  But I'm more of a mechanical system breaker than creative, so perhaps someone else can do better.

[CrowSR]

I've not played that long in 1.0 but I can already feel that the deflection on cloth items needs to be changed a bit.
Nothing feels more frustrating than having a colonist with armor and an autopistol get downed by a tribal raider with a club and magical cloth duster that deflects BULLETS