Should Ludeon release Alpha 17 as refinements, or wait for new content too?

Started by Tynan, February 08, 2017, 01:05:49 PM

Previous topic - Next topic

Should we release Alpha 17 as just a refinements-and-fixes update, or wait to have new content as well?

Release refinements, and then new content in two updates
497 (38.7%)
Wait until there is significant new content to release in one update
787 (61.3%)

Total Members Voted: 1282

Tynan

Thank you for the thoughts so far everyone!

I just wanted to address one thing that's a bit off-topic, but whatever.

Quote from: b0rsuk on February 08, 2017, 01:38:59 PM
- colonist coming from mining site / growing zone empty-handed
- colonist automatically cleaning hospital before operating
- colonists avoiding jobs in rooms that would disturb sleep
- sunlamp automatically turned off at night
- leftovers from deconstruction are hauled away from ALL nearby wall squares before construction is started
- colonists going to remote construction site empty-handed instead of with new batch of materials
- zero warning (not even barking dogs!) when there's a predator coming
- colonists prefering raw food over nutrient paste (it seems to occur when there are several going for paste at the time ? Or just distance ? Makes playing "no sowing plants in the ground" all but impossible)

This isn't an open suggestions thread, but just to note why none of these are changing. There are two main reasons.

1. They're defeated by balance. All of these are things that are essentially of the form, "add AI complexity to make AI perform more optimally and make the game easier". But, if we do that, we'll just have to make the game harder in other ways to maintain the same balance point. e.g. colonists automatically turned off at night? Well, now we just double the sunlamp power consumption to maintain the same balance. Ultimately we end up in the same place, with a bunch of new AI complexity and a bunch of new problems.

2. Any solution creates new problems which are often worse than the original issue. Of course it's easy to say something like, "colonist automatically cleaning hospital before operating". But if we do that, people will soon get pissed off because the patient died while the surgeon was busy cleaning the hospital instead of performing the damn operation. And now we have to add in some sort of player override control, or AI decision making, that can decide when to clean or not clean, and so on and so on. It's an endless regress of AI needing to solve the problems that the last solution created.

The same goes for something like, "sunlamp automatically turned off at night". By whom? When, exactly? What if you don't want it off? How far should they walk to turn it off, and can you control that? What does the UI for configuring all this look like, and what new problems will that create?

We can ask the same battery of questions about almost all these notions.

RW's approach to AI is to set up simple mechanistic rules that the AI follow, let them be suboptimal, and balance the game against those suboptimalities. I've seen other developers try to solve every tiny problem and it's just an endless cycle of problem creation and solving, with complexity increasing the whole time. And ultimately it's pointless, since all those optimalities just get balanced out per point 1.

The one main argument about these kinds of solutions is that they eliminate the incentive for micromanagement by "doing the micro for you". But, there is a point beyond which the downsides of trying to do this outweigh the advantages. Any game like RW is going to have places where players can micro-manage to get micro-benefits. Like many inherent problems in complex systems, this isn't something that can be totally eliminated without the cure itself destroying the whole system. All we can do is try to find an optimal balance where we keep the problem manageably small while also avoiding the huge costs of being too militant in trying to solve it.

I would be really open to specific suggestions about changes to make (in another thread in the suggestions forum) - but it's a lot more helpful to say *exactly* what changes you want. Remember the AI isn't people, it needs to have specific razor-sharp rules for every situation; it can't make intuitive decisions and doesn't understand anything about what's going on. It would be great to try to think through exactly what you want, and what problems that would or would not create, and present the solution instead of just the problem. Then you'll really be engaging in the design process! (and again, please do not make suggestions in this thread, they'll get removed as off-topic).
Tynan Sylvester - @TynanSylvester - Tynan's Blog

Birdy

I'm a new player, but I think making one bigger bugfix+feature patch is more fun! Thanks for the great game so far.

Delgar3

I voted to wait but that's only because I want to selfishly wrap up my modded series before the next alpha. :)
My Youtube Channel: Click Here!
Noob Plays Rimworld: Click Here!

Ouan

I'd rather wait for a larger more dynamic update. I find they make for a greater impact when they come at you and add greatly to replay than smaller, less obvious tweaks. Except of course when boomrats were reskinned. Blew my mind, small, deadly, but now you can tell.
Damn Pirates!

hoffmale

From my experience as programmer (and modder/player), a fast iteration cycle is usually beneficial for both player and developer.

Players get to play less buggy versions of your game faster and have more time to learn new mechanics without being overwhelmed by other new stuff.

Developers get faster bug reports (which can usually be fixed faster, since the relevant code is still fresh in memory) and continuous feedback that gives early warnings when steering into a wrong direction.

However, it puts more work onto modders as the game (and very likely modding interfaces) also change more often. One can try to give some basic consistency guarantees (like semantic versioning tries to), but even then mods will have to be re-tested and sometimes bug-fixed even for small additions.

There is no "best of all worlds" solution to this problem (sadly!). I really don't envy Tynan for having to choose this :/

TheMeInTeam

If you are close to beta or release, I would recommend a faster bug-fix prioritization.  Those things have a habit of causing unforeseen outcomes or to bring the planning fallacy to the fore, which could then stack on even more from new content and pressure the quality of a later/release build.

However, if that kind of milestone is still a decent ways off, it would make more sense to hold off and do the update later.  Two patches gives more iterative feedback/ability to correct than one, but it's less efficient in pure developing.

Absent that being known I'd lean towards a bigger release later.  I'd love to try to break some new systems, but I want to see this game grow as much as it can long term.

Adalah217

Quote from: hoffmale on February 08, 2017, 02:46:44 PM
From my experience as programmer (and modder/player), a fast iteration cycle is usually beneficial for both player and developer.

Players get to play less buggy versions of your game faster and have more time to learn new mechanics without being overwhelmed by other new stuff.

Developers get faster bug reports (which can usually be fixed faster, since the relevant code is still fresh in memory) and continuous feedback that gives early warnings when steering into a wrong direction.

However, it puts more work onto modders as the game (and very likely modding interfaces) also change more often. One can try to give some basic consistency guarantees (like semantic versioning tries to), but even then mods will have to be re-tested and sometimes bug-fixed even for small additions.

There is no "best of all worlds" solution to this problem (sadly!). I really don't envy Tynan for having to choose this :/

If it really is just some additional content, most mods will have no problem with it. I think spreading out the releases into 17 and 18 will ultimately be better for modders as they will be able to see the optimizations (which really break mods) earlier rather than later, and then updating for content won't be too much of a problem.

Thyme

I'm indifferent as well. I've worked around some changes like diseases with mods, so I'm fine. Updates are always nice, but it should be done neatly. If you feel like an update needs new stuff, go for it. Whichever your choice is, I know it'll be great. Also, I'm not yet done with my first A16 run and want to do the Sea Ice Community challenge before A17 comes out! DRM-free tho ;)

PS: Have fun in Venezuela!
I'm from Austria. If I offend you, it's usually inadvertently.
Snowmen army, Chemfuel Generator, Electric Stonecutting, Smelting Tweak

Kirinya

I am in favor of waiting a longer time, although it is not a strong opinion, because of the mod breaking problem.

I love Rimworld, and I think it gets even better with the incredible modding community, which provides a whole lot more replayability than Ludeon's small team could implement in the same time.

Goldenpotatoes

I personally wouldn't have an issue with a purely refinement update to polish up some issues/balance, but I don't see the need to rush to get it out if there's potential for some new content along with all that polishing. The fact you posted an update on A17 should be more than enough to let people know you're still working and not to worry about abandoning it.

99victor

New features tend to bring back players who have become bored with the current state of things. Since this game is still in development, you would want the most people possible testing after a new release.

+1 Wait

...and thanks for letting our voices be heard!

eadras

I've noticed a lot of new players being discouraged when their colonists are hunted and killed by wild animals.  This could be addressed by simply tagging predators that are hunting humans as manhunters (with the accompanying popup message)...  Because they are hunting people.  Perhaps on higher difficulty levels such as Intense or Extreme, the popup warning could be disabled, for those who like the challenge.

FreakyNeo

I voted for two updates instead of one, since I like a more polished experience over a one with more features. I'm not a modder but if you're just doing some large scale bugfixing and rebalancing, I feel it shouldn't take the modders as long as it usually does to update their mods. And long wait times might look bad to people who aren't as involed with the game and only wait for Steam to update their game to remind them playing it, I certainly had that experience with other games.

marcus29ra

I think branching it off would be ideal, players can get accustomed to the changes. Non-Mod players can use it freely and people who need their mods just stay on the old branch

kosh401

Quote from: Tynan on February 08, 2017, 01:05:49 PM

The disadvantage is that it'll break mod compatibility, and the release itself will consume some developer time, slowing long-term progress slightly.

The advantage of waiting is that we break compatibility less, and avoid the chaos/time consumption that an update brings.


Ultimately those points make me vote "wait." It would be nice to have the refinement bug and balance fixes sooner, but we're in alpha and that's just the nature of the beast. The game is in a fun and playable place right now and virtually all of the major mods are updated, so I say wait to include more of whatever content you have planned for us to test/enjoy/suffer through in your next update. Cheers