Procedural sub-biomes to motivate more surface exploration

[Yakky]

So I read a post in which Alshain commented that rovers, while neat in theory, are "incredibly boring" in practice because it takes too long to get anywhere interesting. He has a point, of course.

It got me wondering whether there would be an easy (i.e. simple to implement and not disruptive to current game dynamics) way to spice up planetary surface science to encourage more small-scale land based exploration. In the real world (e.g. Mars Curiosity Rover), a lot of interesting geology happens from checking out different features over a scale of meters rather than kilometers.

Here's what I came up with:

Surface biomes could be chopped up into procedurally generated sub-biomes. Instead of "EVA Report from the surface at Duna's Poles" you could have "EVA Report from the surface at [procedurally generated name XYZ] within Duna's Poles." Names and zones could be similar to those that are randomly/procedurally generated for the various survey and scanning contracts, and I would imagine a lot of the existing code that was built for those could even be recycled.

These sub-biomes would only be discoverable and accessible from the surface. From the atmosphere or from space, they would just show up as the regular biome.

This would motivate more surface-based exploration (rovers!) by offering up a few more science points to people willing to drive around a little. It also doesn't seem too hard to implement, and wouldn't really break any existing game feature.

It would be up to the devs to decide how much additional science to award for going to sub-biomes. It's probably too generous to award full science points for just driving a few km. I would suggest, as a starting point for further discussion/iteration, awarding 15-20% of the baseline science for each additional sub-biome that was explored. So for example, you land in a new biome and your first set of experiments gets the usual award. Then you drive/walk/hop to some other place in a different sub-biome, repeat your experiments (assuming you have the ability to repeat), and get perhaps 15% more science points. A third sub-biome would give you another 15%, etc.

Thoughts on this idea are welcome. One option, of course, would be for some intrepid coder to hack out a simple mod that does this. Anyone interested?

Edited July 30, 2015 by Yakky

[mythbusters844]

Agreed. It'd also be neat if you could actually get science from terrain scatter or even procedural surface features (ravines, caves, etc. that aren't Grand Canyon scale).

[Grumman]

I support the idea of randomly generating nearby "points of interest" each time you land a ship as optional objectives to encourage the use of rovers. They wouldn't have to only be science missions, either: sometimes it might just ask you to go plant a flag over there for some easy Reputation.

[Warrior_Me]

I am not sure if it is even neccesary to introduce a sub-biome mechanism. Maybe it would be easier to make the game automatically generate exploration contacts over time for locations near your rovers, with relatively low science rewards. Of course i would prefer it if those were tied to ground scatter or surface features, but those are just decorative right now, i think, so it would be harder to get that working...

[pincushionman]

How is this different from getting science from several different survey sites?

[Grumman]

How is this different from getting science from several different survey sites?

It's more reliable. Survey contracts aren't always offered, they aren't always for the right planet and they aren't always on the surface. These automatically generated survey sites aren't even a handwave - the landing is a good opportunity to get a better look at the general vicinity, which is in turn a good way to figure out which places are the most important to examine.

[Wjolcz]

I like the idea a lot. Though it could be done in a bit different way.

Generate points of interst around a planted flag. You plant a flag and the system generates contract-like points of interest in a few kilometer radius. You go there, take a sample, come back to base to analize it.

[tater]

I'm not against this, but there is already too much "science" WRT to the tech tree, etc. If something like this were added, I think all science values should drop. That said, a few points:

1. procedurally means "randomly," basically in KSP terms. I would make them less random, and predicated upon:

A. previously generated science for that particular body.

B. previously generated science in/over that geome (biome really needs to go, "biome" requires biology).

C. the presence and skill level of a scientist on the scene---more science skill allows them to discriminate between more of the same, and features of genuine interest.

D. possibly the completion of contracts related to that biome (including "Explore").

To maximize the science gains on your first crewed munar landing, for example, you should have to take data from orbit, over the biome, then previous landings (probes, since this example is your first kerbal there). Each would add the possibility for novel areas to be found near the landing site. Since the first mission is probably NOT a scientist, but a pilot, it will get less total science. Coming back to the same site with a scientist will generate more possible science.

This would be nice because it would encourage more diverse crews.

[Kegereneku]

[WILD RAMBLING]

I don't support randomly procedurally generated sub-biome (two separate thing here)... because we need REAL contents here !!

Procedural generation is all well&nice to fill up unimportant/decorative stuff like the reliefs, texture... but it is a sign of weak-design when it concern a deeply critical gameplay features that require complex consideration to please players.

After all...it is really a "point of interest" if a sub-biomes is as boring and grindy as every other biomes ? Nah! Procedural generation is how we got those contracts-description that no one probably read anymore and lack of attention to details is how we get brain-damaging contract (warning : drama at 120% here).

And so, I opine that if we really want to encourage surface exploration, like "Build useful heavy-rover and base", you have to create handmade masterpiece Area-of-Interest... around key reliefs, and visible from orbit. So you can actually plan on the long term rover mission and think your mission another way. Prepare ISRU and so...

Ideally it would be a different paradigm in gameplay...

- If exploring a dozen normal biome with cheap probes got you (ex) 20points

- Exploring one of those "point of interest" sub,biome would only give you 10points with a drone too cheap, but 50p with a rovers, 100p if you have a Scientist to reset heavy-science and 200p if you planned a rover to get to the equivalent Area-of-Interest that was put voluntarily 50km away with a hand-made "relief" for Rovers.

A zone you really want to explore yourself. And who know, maybe you could craft actual cave there ?(cave are for now incompatible with the tools used to make planet relief)

[/WILD RAMBLING]

[WinkAllKerb'']

Also like the idea of procedural biome / sub-biome.

[tater]

[WILD RAMBLING]

I don't support randomly procedurally generated sub-biome (two separate thing here)... because we need REAL contents here !!

Procedural generation is all well&nice to fill up unimportant/decorative stuff like the reliefs, texture... but it is a sign of weak-design when it concern a deeply critical gameplay features that require complex consideration to please players.

After all...it is really a "point of interest" if a sub-biomes is as boring and grindy as every other biomes ? Nah! Procedural generation is how we got those contracts-description that no one probably read anymore and lack of attention to details is how we get brain-damaging contract (warning : drama at 120% here).

And so, I opine that if we really want to encourage surface exploration, like "Build useful heavy-rover and base", you have to create handmade masterpiece Area-of-Interest... around key reliefs, and visible from orbit. So you can actually plan on the long term rover mission and think your mission another way. Prepare ISRU and so...

Ideally it would be a different paradigm in gameplay...

- If exploring a dozen normal biome with cheap probes got you (ex) 20points

- Exploring one of those "point of interest" sub,biome would only give you 10points with a drone too cheap, but 50p with a rovers, 100p if you have a Scientist to reset heavy-science and 200p if you planned a rover to get to the equivalent Area-of-Interest that was put voluntarily 50km away with a hand-made "relief" for Rovers.

A zone you really want to explore yourself. And who know, maybe you could craft actual cave there ?(cave are for now incompatible with the tools used to make planet relief)

[/WILD RAMBLING]

There is a good reason for "random," IMO. Replay. If the only really cool places are always in the same few spots, you know where to land every time.

So how about a slight variant...

For each region or group of regions depending on the body Squad makes, I dunno, 10 "areas of interest." Instead of tweaking the map directly, these will appear randomly within that region type in some fashion. Instead of just an area like the dumb survey contracts, they could be something like interactive "scatter" or the anomalies. No more than 2 or 3 of these would ever appear in a given career, so you might go a while before you see all of them. Some might be boulders that are substantially different than the scatter (all would be solid), others might be a sort of "decal" on the surface that changes the color (some sort of neat geology, perhaps ice in polar craters).

Your idea about awesomely done areas of interest is also a great idea, that's the kind of anomaly I'd like to see instead of what we have. Stuff that is plausible, the real universe makes some stunning topography.

[amorymeltzer]

For each region or group of regions depending on the body Squad makes, I dunno, 10 "areas of interest." Instead of tweaking the map directly, these will appear randomly within that region type in some fashion. Instead of just an area like the dumb survey contracts, they could be something like interactive "scatter" or the anomalies.

I think this is the way to go. Make some interesting and variable scatters, so that not every grasslands or highlands looks the same. They can be functionally the same but as long as they look different here 'n there, they'll be fun to get to. They can even be hardcoded (I would imagine) so that they don't change up on you (it would be weird to have trees and ice floes move around).

[Alshain]

Interestingly enough, Fine Print had rover specific survey contracts. As usual, when Squad implemented it in stock, they watered it down. Their justification was they wanted you to choose how to achieve it, but as you can see it doesn't work because it mixes space, aerial, and surface in single contracts. This is why I tend to be grumpy when Squad says they are implementing one of my mods in stock, almost all of the plugin based mods that are implemented end up inferior to the original for various reasons, and now Fine Print is gone.

I don't think the idea in the OP is a bad one but based on how hard it was for use to get biomes on anything beyond Minmus, I just don't expect it.

I do like the idea of randomly generated points of interest within a certain distance of your landing/flag planting. There was another thread that I suggested that for EVA's and the idea quickly got shot down by the people discussing it.

Edited July 31, 2015 by Alshain

[Tw1]

I would also prefer randomly, or at least procedurally generated points of interest. The kind of thing I'd like best is summed up in the first two posts of this thread http://forum.kerbalspaceprogram.com/threads/120014-Procedurally-generate-surface-objects-of-scientific-interest

You've nailed one problem with the current system, it's just too sparce. But it's also not that interesting to use. Something with semi randomness to keep things unpredictable, and a set up that has you watching your instruments, looking for changes would feel much more like proper Exploration. Plus, it'd work in sandbox - there's the challenge of finding stuff, despite you not needing to collect points.

Something like this is better than the survey contracts, because it'd be always available, and you're exploring, and discovering things by yourself, not taking some data because someone sent you, to something they identified.

The way KSP uses points is a problem. It makes it hard to add something that lets you collect science without re balancing many things, and if you add things which are kust for interest, they'll feel out of place, what's the benefit of visiting?

Career's features have been built around unlocking a finite list of upgrades, rather than sustaining and continually improving your endeavours in space. If it was about balancing your earns and gains over time, just the rate at which you earn is important.

You could have plenty of places to earn science from, as long as you're limited, some how, such as by life support, or how often you get paid. Or maybe the game could be balanced so you don't necessarily need to collect all the points possible and still come home with a gain, making player interest the limit.

Anyway, having some involved surface exploration would be a step from Rocket sim with challenges tacked on, to the full space experience it could be.

Edited July 31, 2015 by Tw1

[Grumman]

I'm not against this, but there is already too much "science" WRT to the tech tree, etc. If something like this were added, I think all science values should drop.

**** no. If anything, Squad needs to make up for the last set of nerfs, because the scarcity of science is the worst thing about the game right now.

[Yakky]

Wow, so many interesting extensions of this concept. I never expected it to generate this much interest!

Lots of discussion of how "random" (vs. repeatable procedural) the sub-biomes should be. My view there is they at least have to be consistent within a given game file. They shouldn't shift around from mission to mission, but it might be acceptable to have every new game have a different randomly generated set. That would keep things interesting.

Also could work OK to just have them be totally procedural (not varying from game to game) since realistically it's not that common to touch down in exactly the same place every time. When I go to a planet (or the Mun), the point where I land is generally not too carefully selected, or driven by practical considerations like my orbital inclination and where it's daytime, unless I need to link up with something that I've landed earlier. So in practice, randomness or non-randomness probably wouldn't impact the playability of the game much. I envision the typical celestial body having thousands of sub-biomes, so for most practical purposes there would be nearly infinite variability from game to game.

I'm sensitive to the number of things Squad has to try to balance in producing a game that works for a broad audience. I'm not a modder or insider, so I won't pretend to know how the conversations about implementing or not implementing different features have gone. At least we have an open architecture that allows mods to customize the game. I also proposed this idea in the "add-ons" section of the forum for that reason.

[tater]

**** no. If anything, Squad needs to make up for the last set of nerfs, because the scarcity of science is the worst thing about the game right now.

This is a joke, right?

If you don't have everything unlocked in career before even the first decent Duna transfer window comes along, you're doing something wrong.

How many Kerbin days from never launching a rocket to unlocking everything should it take? Nerving the science, aside from just generally making sense, in this case would have a net effect of nothing, or positive science as there would be additional stuff per "biome." If there are 3 new EVA things per "biome" then you really want 3X the science? I'd say they'd want no more than the total of normal, + 3 observations to be close to what it is fir the one set now.

Edited July 31, 2015 by tater

[Alshain]

This is a joke, right?

If you don't have everything unlocked in career before even the first decent Duna transfer window comes along, you're doing something wrong.

While I do agree there is plenty of science already, this sentence struck me as odd. If you choose not to collect all the science from all the biomes around Kerbin, the Mun, and Minmus before going to Duna you are not doing something wrong.

I very rarely finish the tech tree before my first Duna and/or Eve transfer. But, I don't bother with biome hopping to be a completionist.

Edited July 31, 2015 by Alshain

[tater]

While I do agree there is plenty of science already, this sentence struck me as odd. If you choose not to collect all the science from all the biomes around Kerbin, the Mun, and Minmus before going to Duna you are not doing something wrong.

I very rarely finish the tech tree before my first Duna and/or Eve transfer. But, I don't bother with biome hopping to be a completionist.

I don't, either. I play with LS, and aside from just bringing supplies, I bring at least hitchhiker for every couple kerbals for Duna (even more stuff for farther out), plus a kerbin reentry vehicle, often more. in 0.90 I unlocked everything in no time, way before Duna. It takes slightly longer in 1.x due to the lower science from contracts.

Brand new career with the kerbal planetary systems mod, and I have probes on Duna and Ike now, and 2 vehicles en route with astronauts (1 unmanned ahead with a hab and fuel to leave in orbit). I have a few of the 550s left to unlock in the tech tree, so it's not totally unlocked. Wanna say it's around day 255 of year 1.

So 1 year ago, no kerbal had been in space. Before the year is out, Kerbin will have exhausted all possible improvements in spaceflight technology, and will have landed kerbals on both moons, and Duna. There are a couple of permanently crewed bases on both moons, with transfer stations above (I cycle resupplies and new crew so the crew can become useful skill wise).

So the question is, how many days should this take? Sputnik launches, and before the year is out cosmonauts are on Mars? By the time Kennedy makes his speech, there are men en route to Jupiter? Or is it 4x shorter because of kerbin days?

Edited July 31, 2015 by tater

[Kegereneku]

There is a good reason for "random," IMO. Replay. If the only really cool places are always in the same few spots, you know where to land every time.

...and there's better reasons we don't do thing at random...

It might sound counter intuitive but Game-maker actually work to turn uncorrelated logic-function into a sublime & deeply intertwined gameplay.

Nothing else to say about the rest of your message as it's within the things that change from Black to White depending of balance and for which Grey is masochist boredom. (aka "we can't actually satisfy everybody")

Sometime you really have to DUMP ideas, put them unto fire and send the ash away in space... like the idea that "random = infinite novelty".

This is a joke, right?

If you don't have everything unlocked in career before even the first decent Duna transfer window comes along, you're doing something wrong.

It's actually sound design.

You have to go to great effort (lot of Mun/Minmus landing and retrieval) to actually advance much in the Tech-Tree.

By allowing to finish the entire Tech-tree with only the Mun/Minmus or another planet you are giving the ability to play a whole scenario centered around a Moon or Planet.

This one seem childish but "Balanced Progression =/= 100% Completion Grinding" many game-maker do that error.

So the question is, how many days should this take? Sputnik launches, and before the year is out cosmonauts are on Mars? By the time Kennedy makes his speech, there are men en route to Jupiter? Or is it 4x shorter because of kerbin days?

I'm afraid it is actually realistic.

We will have largely gone beyond KSP tech-level by the time we 'need', let alone 'do' Surface-Base on the Moon. All colonization/Life-support experiments can be done in orbit (yes all, we would barely need sample to reproduce Mars soil) and we will never exploit economically Space-Resources before we have technology which would sound OP in KSP.

It's an unrealistic luxury in KSP to be able to send crewed mission to another planet.

And in term of gameplay, many know that a Tech-Tree independent from Science, but based on Money and Experience would be more realistic.

However, reproducing reality isn't how you encourage Exploration (opposite to spamming satellite contract and sucking Subvention).

Edited August 1, 2015 by Kegereneku

[tater]

...and there's better reasons we don't do thing at random...

It might sound counter intuitive but Game-maker actually work to turn uncorrelated logic-function into a sublime & deeply intertwined gameplay.

Nothing else to say about the rest of your message as it's within the things that change from Black to White depending of balance and for which Grey is masochist boredom. (aka "we can't actually satisfy everybody")

Sometime you really have to DUMP ideas, put them unto fire and send the ash away in space... like the idea that "random = infinite novelty".

Once again, we are talking past each other.

The first thing to remember is that career mode, and the game in general are NOT going to change massively. I'd like it to happen, but the probability of that happening is ZERO.

Given this, we have to work within ideas that have a non-zero probability of being added.

"Biomes" are a thing. All the worlds need to be improved, so your idea of making interesting areas via "art" is not impossible, so that's a good idea. Scatter exists, as do "anomalies," so perhaps those mechanisms might also be used.

I like your idea of detailed areas of interest... describe one to me, just an example. Is it possible that some of these areas of interest, as detailed as they might be are not really site-specific? (by site I mean the exact location on the map, not the type of terrain. Ie: interesting feature X only occurs in lowlands, but it need not occur only in a certain lowland area)

Duna might be a better example, but for the Mun, we have scatter rocks, sort of like this, only rounded for some inexplicable reason:

(the guy in the picture, Jack (Harrison) Schmitt, taught a unit on lunar geology I took a rather long time ago, as you might imagine, his slideshows were epic for a geology class ).

Isolated outcropping like this that are different than the usual scatter would be a simple, random addition that would provide something to do on EVA. Once you land on a body it might have a % chance of generating a rock that is different from other scatter XXX to YYYY meters away from the landing site (far enough that it might look like normal scatter at a glance). Say just angular boulders for the Mun, instead of the rounded scatter. These would have colliders, too.

Your idea would be great for things like Rilles, domes, or even a kind of material collected on the above Apollo 17 mission, dark mantling deposits. Those are macroscopic features and by definition need to be done to the terrain at large. What I am talking about is the sort of small-scale novelty that only gets spotted once you are on the surface itself. Think novel deposits that Curiosity finds pretty commonly. Geologists know what to roughly expect, but then Curiosity sends an image os layered sediment visible due to erosion, for example, and it's exciting. Invisible from orbit, obvious at EVA height. These features are not truly random, which is why we can use the "biomes" to chose appropriate small-scale additions. Areas that formerly held water on Duna would get exposed sedimentary stuff that a scientist could spot, for example.

It's actually sound design.

You have to go to great effort (lot of Mun/Minmus landing and retrieval) to actually advance much in the Tech-Tree.

By allowing to finish the entire Tech-tree with only the Mun/Minmus or another planet you are giving the ability to play a whole scenario centered around a Moon or Planet.

This one seem childish but "Balanced Progression =/= 100% Completion Grinding" many game-maker do that error.

I was merely making an observation that this is the case. I wasn't saying to nerf the science to the point you needed to go to Jool to make a new rocket engine (dumb tech tree), just that by adding additional large chunks of science the whole tree might get unlocked far faster than it does already. In general, KSP needs a time-based mechanic. I suppose many play by warping any given mission to competition, to get time passing. I'm against planetary science buying technology, anyway, but there is no way that's gonna change, sadly.

I'm afraid it is actually realistic.

We will have largely gone beyond KSP tech-level by the time we 'need', let alone 'do' Surface-Base on the Moon. All colonization/Life-support experiments can be done in orbit (yes all, we would barely need sample to reproduce Mars soil) and we will never exploit economically Space-Resources before we have technology which would sound OP in KSP.

It's an unrealistic luxury in KSP to be able to send crewed mission to another planet.

I don't disagree, actually, I have said many times that the entire tech tree is basically concurrent, and spans a gap from the late 1950s to the early 1960s

And in term of gameplay, many know that a Tech-Tree independent from Science, but based on Money and Experience would be more realistic.

However, reproducing reality isn't how you encourage Exploration (opposite to spamming satellite contract and sucking Subvention).

You are "preaching to the choir."

I think that the best possible game from a replay standpoint would have the entire solar system randomized each game, with detailed planets that are interesting, and the science is redesigned to gain useful information some of which is required to be able to play well (determining the important data about planets needed to safely fly there and return within the game, instead of consulting a wiki page---the data that requires spaceflight, that is, the baseline would be what you'd know from ground-based astronomy).

Edited August 1, 2015 by tater

[cantab]

Well, at the moment I think the contracts do a reasonable job of making rovers useful. In particular there are some that generate several survey sites all close together and all on the surface, those are the ones that are obviously aimed at rovers.

As far as science goes, well I think the whole system could be a lot more interesting. Adding science instruments to work with ground scatter, along with generally improving said ground scatter, could do pretty well. Even if it would boil down to "hunt for an interesting rock" most of the time. I also think there should be experiments that depend on moving through time and space. For example the barometer would benefit from taking readings at different altitudes, the seismometer would want to be left stable on the surface for some time, and perhaps a new experiment would want you to drive around with it.

[Kegereneku]

Once again, we are talking past each other.

I like your idea of detailed areas of interest... describe one to me, just an example. Is it possible that some of these areas of interest, as detailed as they might be are not really site-specific? (by site I mean the exact location on the map, not the type of terrain. Ie: interesting feature X only occurs in lowlands, but it need not occur only in a certain lowland area)

The concept of "Areas of real interests" hinge around human creator actively thinking about how to actually make best use of the game mechanics available, opposite to "program an half-assed generation of 99% similar content without any foresight, hoping it turn good by chance". Procedural generation is a tool here, not a goal.

# If you generate sub-biome in every biome you are only increasing the grind and globally making static probes inferior to rovers. Gathering science become a Rover-exploit or a matter of luck.

Aka : "will I have fun here ? Or did I wasted my awesome Manned-Roving-Base on a boring place 100km away from any fun ?"

# On the other hand, if you make Area-of-Interest at specific place, carefully balance science rewards between biome & AoI, make them visible from orbit...

- Static probes stay useful on normal biome.

- Small rover become useful within AoI without having to drive for hours to already covered biomes

- Long manned mission become fun because Kerbal become much more interesting than rover (imagine climbing stuff !")

You wanted an example : Imagine Duna

- One area of interest is at the equator, deep in the basin (large regions of dark soil), it's a "magnetic anomaly" (no that's not a cheesy monolith, it just generate concentric sub-biome), if you look closely, you realize that the anomaly is at the center of an absolutely flat place over a great distance, like... a runway.

- One 'cave' is 50km away North, at the edge of the depression/ravine. It is ISRU friendly.

- Next one is closer at the top of the depression/ravine it's just a big noticeable rock... if you look closely you'll notice a "road" to the 'cave'

- Still toward North is a mountainous crest. One huge point of XP at a place hard to land or drive to.

- Another one is 30km away, it is basically Olympus Mon. If you look closely you'll notice something like a roads.

- Next one is 30km away at the limit of the polar cap, it's a large crater with something at its center (all interesting things are visible from orbit like survey after a scan except forever), it is also ISRU friendly.

- Last one is at the pole.

...if my calculation are right, you now have a line for an awesome Rover-trek from the equator to the pole with refueling place for rocket/shuttle, base, and so much science you can train your scientist here (meaning regular travel)

The best in the above ?

It's more or less already exist, except visible from orbit and considered as something apart from biome, with Asteroid-like big rock you can climb on without needing 16G of RAM.

...I safely bet that we all consider knowing in which biome you are without doing a report to be a missing feature.

I think that the best possible game from a replay standpoint would have the entire solar system randomized each game, with detailed planets that are interesting,

From my point of view it would be the WORST idea ever for KSP. You wish to pursue a goal with a process that defeat it entirely.

As said, "randomized" don't necessarily correlate with "interesting", nor with increased replayability. In fact it can lessen replay.

You can easily randomly generate 10 generic & unsatisfying planets or 1 absolutely marvelous planet desperately surrounded by 9 dull-planets... leading inevitably to "hey let's player customize each system>planets>biome>sub-biome !" then "I'm lazy, let's allow to download system in case a moder do a cool one for free".

I'm blaming Minecraft and cognitive bias for the belief that infinite & random generation is a evolution for anything.

[Mr. Scruffy]

I have been pondering this a little after having read the thread and playing for a bit, again. I was thinking, that maybe we dont need artificial interesting spots, but could do with what we have, if we just use a couple of more parameters for the experiments. For each geome there could be altitude as well as attitude ´records´. Like: Your first experiment in the given geome pays the usual (maybe compensated for the addtional stuff, as: ) and the altitude and attitude (slopyness of terrain) get noted. Now, you can conduct additional experiments at the same geome, but at higher/lower altitudes/angles, paying you science according to the difference of each to your ´record´ (highest/lowest already conducted experiment of this type in the geome).

In effect, after landing and conducting the standard experiments as we know them, you´d want to drive around to cover a bit of altitude difference, go as high above and low below the landing site as you find reasonable to maximize your science rewards, and investigate this slope over there, while you are at it.

A third parameter could simply be "distance to landing site" rewarding just simply driving as far a way as you want to repeat the experiment on the landing site.

Have each of the parameters modify the reward for the repeated experiment for upto 100% max. The extra 100% for altitude would have to be split between at least two (additional, after the first) experiments as at least one experiment would have to take place at the lowest point of elevation of the geome and another at its highest (if you want the full 100% that is - you could chose to go with less, of course). Same for attitude, if the first experiement is not conducted at absolutely level terrain and the second standing at 90° (the full 100% will be very hard to get i guess, or the max could be set to <90°, say 60° ). The bonus for distance could be fully ´harvested´ with one additional experiment by moving the rover to the threshold of the bonus at, say, 10(?) km distance.

EDIT: I think this will also give you a good choice of difficulty: Landing sites with differently sloped terrain and a lot of altitude differences tend to be hard - but would be rewarding.

Edited August 3, 2015 by Mr. Scruffy

[tater]

The concept of "Areas of real interests" hinge around human creator actively thinking about how to actually make best use of the game mechanics available, opposite to "program an half-assed generation of 99% similar content without any foresight, hoping it turn good by chance". Procedural generation is a tool here, not a goal.

# If you generate sub-biome in every biome you are only increasing the grind and globally making static probes inferior to rovers. Gathering science become a Rover-exploit or a matter of luck.

Aka : "will I have fun here ? Or did I wasted my awesome Manned-Roving-Base on a boring place 100km away from any fun ?"

# On the other hand, if you make Area-of-Interest at specific place, carefully balance science rewards between biome & AoI, make them visible from orbit...

- Static probes stay useful on normal biome.

- Small rover become useful within AoI without having to drive for hours to already covered biomes

- Long manned mission become fun because Kerbal become much more interesting than rover (imagine climbing stuff !")

You wanted an example : Imagine Duna

- One area of interest is at the equator, deep in the basin (large regions of dark soil), it's a "magnetic anomaly" (no that's not a cheesy monolith, it just generate concentric sub-biome), if you look closely, you realize that the anomaly is at the center of an absolutely flat place over a great distance, like... a runway.

- One 'cave' is 50km away North, at the edge of the depression/ravine. It is ISRU friendly.

- Next one is closer at the top of the depression/ravine it's just a big noticeable rock... if you look closely you'll notice a "road" to the 'cave'

- Still toward North is a mountainous crest. One huge point of XP at a place hard to land or drive to.

- Another one is 30km away, it is basically Olympus Mon. If you look closely you'll notice something like a roads.

- Next one is 30km away at the limit of the polar cap, it's a large crater with something at its center (all interesting things are visible from orbit like survey after a scan except forever), it is also ISRU friendly.

- Last one is at the pole.

...if my calculation are right, you now have a line for an awesome Rover-trek from the equator to the pole with refueling place for rocket/shuttle, base, and so much science you can train your scientist here (meaning regular travel)

The best in the above ?

It's more or less already exist, except visible from orbit and considered as something apart from biome, with Asteroid-like big rock you can climb on without needing 16G of RAM.

...I safely bet that we all consider knowing in which biome you are without doing a report to be a missing feature.

Again, I agree completely that areas of actual geologic interest should be taken advantage of, 100%. Heck, I think squad should do a better job of creating such areas in general.

Change scale, though, and ask a simple question: Could an astronaut or probe ever find something of interest at a landing site that is not apparent from orbital photography? Yes, or no? If the answer is "yes," then you are either suggesting that Squad manually scatter stuff on this tiny scale all over every world very carefully, scatter it randomly over every world on a level that you'd find them at a reasonable rate---OR, it could be scattered randomly only in areas where the player has a craft capable of possibly spotting these tiny "anomalies" of interest (these would be actual geology).

BTW, if you said "no," you are simply wrong, it's not a question of opinion. Curiosity, for example, finds things that were not visible from orbit practically every Sol. In the image above of the Apollo 17 site, that rock would have been barely visible in some high-res imagery. The LRO flying now can image Apollo sites with a resolution of ~1m, and the Lunar Orbiters of the mid 1960s did about the same. Ranger managed 0.5 m resolution on the last image before impact. You could tell boulders were there, but not if they were interesting for a geologist.

The first 2 methods require literally millions of new, interactive scatters be added. The 3d method need only throw possibly a couple per landing site, likely 5 orders of magnitude less stuff for the game to track.

There is a place for random, and it's on THAT scale size I am talking, NOT major features, visible from high-res orbital photography.

From my point of view it would be the WORST idea ever for KSP. You wish to pursue a goal with a process that defeat it entirely.

As said, "randomized" don't necessarily correlate with "interesting", nor with increased replayability. In fact it can lessen replay.

You can easily randomly generate 10 generic & unsatisfying planets or 1 absolutely marvelous planet desperately surrounded by 9 dull-planets... leading inevitably to "hey let's player customize each system>planets>biome>sub-biome !" then "I'm lazy, let's allow to download system in case a moder do a cool one for free".

I'm blaming Minecraft and cognitive bias for the belief that infinite & random generation is a evolution for anything.

The purpose of KSP is really to explore a solar system with a space program. Such a randomized system would have to be done in the context of changing the game such that there is "fog" about what the other planets actually looked like. Orbital elements are calculated with precision from ground-based astronomy, and some data about atmospheres is know pretty well (spectrometry). That said, the resolution of the surfaces would be pretty low. The first Mariner flyby utterly changed what people knew about Mars, for example. A random system need not have planets generated 100% randomly, BTW, they need only be placed and scaled randomly. Imagine the Kerbol system, but squad makes 10 more planets, and twice as many new minor planets/moons (tiny is size a, huge mons are size z)---they can be made at any detail you desire. A randomized solar system would simply alter the orbits, and pick 6-10 as "planets" and give each a chance to have moons based upon size (smaller planets get 0-3 moons of size a-g, medium planets might get sizes a-n, and gas or ice giant planets get moons of sizes a-z. All can be as designed, or scaled by a factor of 1 to 6 (size, and orbital distances). The scale factor would be capped in difficulty settings, since 6X is about like 6.4X RSS. Kerbin would stay the same (or maybe scale from 1-3X so standard parts would fine (3.2 kerbin is great with nothing but stock parts)), though it could have the orbital distance chance somewhat (within kerbol's habitable zone).

So we have a random system that the player must explore to know about (again, better science instruments and "science" in general to meaningfully unlock the secrets of worlds, taking images (camera part, like scansat) to make a map view at whatever resolution, for example). The planets are just as well constructed (ideally better) than they are now.

Randomization is involved, and it is unambiguously better for replay, with none of the negatives you mention (I granted all your positive improvements, so you'll need to explain how not having a wiki page for every world makes exploring less cool---the game will generate all that data as a function of what science you do (atm height, composition, etc, etc).

Edited August 3, 2015 by tater