System scale as stock difficulty option

[pandaman]

I seem to recall this mentioned or suggested before a while ago,  though a quick search turned up blank, so please move or remove if appropriate.

Ok I know there are mods that already sort of do this, but my thought is for an option in the game creation/setup menu to include different scaled stock systems as part of the difficulties settings. 

In its simplest form it could just offer something like - Small (current stock size), Medium (2× scale) and Large (4x scale) the only difference between them would be the sizes of the orbits and the planets all scaled up accordingly.

Ideally scaling up a planet or moon won't change the heights of mountains etc. and it may well make sense to keep the atmospheres the same heights etc too (so that a whole new set of values don't need to be learnt), but I don't know how all that is worked out and whether or not it would cause other issues.

By not going much bigger than this in scale the stock parts should all be adequate as they are, so the extra difficulty is purely as a result of larger bodies further apart, and therefore genuinely more difficult.

An extension of this concept could be to set the system and planetary scales independently, so that we could have say x2 planet sizes at 4x distances etc. but that may be a bit too much. 

I have no idea how easy or practical this could be to implement, but on the face of it, it could be a viable way to increase the difficulty and challenge without actually adding or taking away stuff to artificially make it harder.

[tater]

I've said the same in various threads, and it makes worlds of sense.

Honestly, scale is by far more of a difficulty modifier than the other stuff.

Easy = current scale.

Normal = 3.2X.

Hard = ? (6.4X is doable with stock parts, but rough in the stock game (no KJR, etc).). I think you are right that 4, or maybe 5X is possible.

I played the 365 mod for a while, which was 3.2 planet rescales with 6.4X distances and liked it.

[SchweinAero]

I like this suggestion because I'd love to experience those ISS style orbits where the planet looks enormous and practically flat underneath. But allow me to illustrate one problem with rescaling of terrain:

(This is 3.2x by the way.) If absolute heights are not changed, the terrain will look mushy at high scales, spiky at low scales, or both. If you can tolerate it visually, no problem, but it might also make spaceport and easter egg placement harder.

[pandaman]

@SchweinAero

Cool illustration, thanks.

That's exactly the kind of issue I would be concerned about.  I just don't know if it would be 'right' to have all the hills, mountains and valleys scaled up too, especially at 4x the mountains West of KSC would be a real obstacle for  shuttle landings, but then again they are 4x the distance away too so the glide angle would be the same.  I suppose it would depend on how exactly the height maps are done and how easy it would be to adjust them to compensate. 

@tater and anyone else who's used a scaled up system.  How was it done on those?  Were the hills stretched like in SchweinAero's images, did they expand the distances between them and retain the steepness without actually making them higher, or were they just scaled up?

[tater]

The terrain is a little less spiky, generally. You can control the vertical scaling independently (look at sigma dimensions as an example for an easy interface). 3.2X seems to be the sweet spot for rescaling existing worlds, retaining their visual appeal.

3.2X is not enough to make a huge difference in gameplay, however. From a difficulty standpoint, I think rescaling combined with life support makes a substantial difference.

I could see fixed rescales at each size, and perhaps it requires new world models built at this scales.

From a gameplay standpoint, I would want to see the scales introduce meaningful design challenges. For example, the Mun. Given available parts, I think that there should be multiple, plausiblely efficient mission strategies---Kerbin orbit rendezvous, direct ascent, and munar orbit rendezvous (or combinations). Any Mun landing where a staged lander is not a reasonable design option is not different from stock. Right now, a tiny craft with a mk1 pod is trivial to get to the Mun, anything else is roleplaying, or overkill.

[Wjolcz]

I never really liked how you can see the curvature of Kerbin when 10km up. It just doesn't look any good. I've seen comments that the 6.4x rescale feels just right and always wanted to try it out, but I'm still waiting for the will to play the game to come back. Wouldn't mind if the 6.4x rescale was a permanent thing with stats of the stock parts bumped up a bit to compensate for the rescale.

 

[tater]

6.4 with some part mods is awesome. SSTU, for example. 

Edited August 10, 2016 by tater

[klgraham1013]

I agree.  A much better means of difficulty than what we have now.

[CSE]

I love the larger scales, and especially the difference style of launching to orbit, where sideways becomes vastly more important than up. But I think it will always be difficult to offer rescaling as a stock setting because some of the scaling decisions are subjective choices with no single right answer.

• For orbits, should you scale the SMA by the same amount as the planet radius? If so, do you set Kerbin's day to a round number of hours, as many scaling mods do, and give up spin/orbit resonances? Or do you accept a fractional day length (good luck remembering your synchronous orbit period)? Or do you apply different scale factors to the planet radius and the orbit SMA (my preference, but moons look different in the sky because they're closer or farther away; it's a lot less visually impressive with them further away)? These are all valid solar systems to choose.
• For the terrain, Earth is very flat over large areas; all the steepness is at very small horizontal scales, smaller than an upscaled Kerbin's terrain would have. Kerbin rescales face a choice between Earth-like steepness, which looks 'right' from ground level and low altitudes, or Earth-like heights, which plays 'right' with launch and re-entry.
• For the atmosphere, stock Kerbin's atmosphere is given much more thickness than Earth's to improve the game experience and balance. It's possible to calculate a realistic rescaled atmosphere but, for scales smaller than 10x, most players will prefer some padding; how much is a matter of personal preference, and perhaps which mods you're using (FAR and/or RealHeat). The atmosphere scale combines with the terrain scale to affect whether you can fly over or parachute onto mountains.
• For career mode, the rocket equation makes costs very sensitive to the amount of dV required. Balancing this across several scales would be difficult. It would also depend on the choices for day length and atmosphere scales.

So although I play almost exclusively in a rescaled Kerbiverse, I don't think there's a single set of scaling choices that suits everybody (except in RSS), and that makes it hard to make it a stock feature. Beautiful feature though it is, IMO scaling is better as a mod feature.

[tater]

@CSE, those are good points about rescales, but the balance issues don't matter in this case, and the terrain and atmosphere can scale independently.

Balance doesn't matter in the sense that the only question we need ask is, "Is 3.2X with the settings chosen (atmosphere height, etc) harder than stock?" If the answer is yes, then we have our medium difficulty. If the answer to "Is 5X (or whatever testing shows works with stock parts) harder than 3.2X?" then our hard is harder, and we're good.

I would have the atmospheres become closer to reality with each rescale as well (i.e.: I'd not scale them at all I think).

Again, the goal is to make it harder. I have played career with mostly stock parts (I added bigger parts made for the stock game like SpaceY) on rescales up to 6.4X, no other changes (FAR, etc, obviously) and it works.

Day length is the only issue that I always have with rescales... dunno about that one.

[Snark]

Yah, the problem is that Kerbin and other planets have terrain that's shaped "appropriately" (i.e. like real-world terrain, as viewed through a playability/game-artistic lens)... it just happens to be wrapped around a much smaller sphere.  There's not really any way to preserve that when you scale, without something giving way.

I don't see any simple answer to this.  A few possibilities that spring to mind:

• Have only two or three fixed size options (no sliding scale).  Have a complete re-implement of each planet for each scale (i.e. an artist at Squad builds it from scratch, with appropriate landforms).  Advantage:  everything works nicely and looks great.  Disadvantage:  big labor investment (I've never built a planet from scratch, but I'm guessing it's quite a task).  Also, would put a fair amount of onus on planet-pack developers to follow suit, though that's less of a concern.
• Have planets be generated procedurally, so that changing the planet scale would automatically adjust the way landforms work.  This can certainly be *done* (look at what No Man's Sky does)... but again, it would be a chunk of work, especially if you want to try to build a system that smoothly meshes the ability for an artist to decide overall look with procedural generation.  Large development task.

Then there's the question of how many people would actually use it.  I have no idea what the numbers would look like, here, so I really can't say.  I know that I myself am quite happy with the scale and balance of the current system (I've gotten used to it, it's like an old friend), so I'd never use that option even if it were available.  Given the size of the development task that it would be to implement such a thing, it's a legitimate concern whether there would be enough demand to make the investment worthwhile.

My own preference:  if Squad has some time budgeted for solar-system development at some point, I'd far rather that they spend the time in adding new planets/moons, and/or refining the terrain of planets that are already there (see all the discussions about "old Duna" versus "new Duna"), and/or revamping the code base to make axial tilt possible, rather than implementing a scale-picker.

[tater]

I'd agree completely with the idea of making all the extant worlds better first. 

That said, I personally try stock with new updates, and sometimes for testing mods (the only reason I'm playing stock size at the moment is for testing SSTU). Otherwise I have decided I really dislike the stock scale, it makes space seem less majestic to me... Remember that stock Jool is about the size of Earth as a reality check. I like the sense of working to get to LKO only to see I'm only barely in space  .

I also like rescale for replay as I need to figure out new solutions to new problems.

[pandaman]

I was rather hoping that it would be a relatively simple process to scale up the system by just adding a multiplier to the diameters or something, then a couple of tweaks, like adjusting height map limits to keep the mountains in check, which could be done 'on demand', but it seems as it's not quite that simple.  I would gather that the only real way this will work is by effectively having scaled up 'copies ' of the planets ready to go. I guess that increasing system size would be much simpler though, as none of the bodies themselves would be different, but I doubt that would increase difficulty enough on it's own to be worth it.

If it's going to be that much work then perhaps a better solution may be to have scaled up versions of Kerbin, so that the 'home' planet is always the same in all, but scale.  Then have alternative sets of new planets for each scale.  That would be a LOT of extra work, but it would have the benefit of giving new places to explore as players get more proficient and feel ready to 'go larger'.  That I think is certainly out of scope for stock for now, but as additional 'official stock expansion pack(s)' then it may well be a viable option.