It's painfully obvious when you have the Analysis mode on and the blue heatmap for biological results. You can see entire repeating patterns that makes it look like the planet has been decorated with floral wallpaper. Even on quite well-known worlds.
New Planet Tech is KILLER of Exploration (all terrain is tiling/repeating/not procedural/random)
- Thread starter BlackMaze
- Start date
Because it doesn't exist until the procedural key is run, and that's run during the loading screen when players jump in, or when FDEV run it themselves to generate the system. It's like asking why you can't see a drawing I haven't drawn yet.
Or these were the sorts of canyons they were talking about. There are a few steep, narrow bits there but certainly nowhere near Horizons extremes.
It's the ludicrously steep, narrow mountain that pulls me out every time I see it (there isn't one in this picture).
I know its procedural from a seed - question was - I wonder if fdev checks a few random systems to see what’s happening, would love to be able to run up a few randomly chosen systems to find the most mad ones.
I've visited 45,701 systems according to CMDR stats, how do you think you viewing a few randomly chosen systems will give you a better chance of finding a more mad one than me?
Frontier could clearly have a background process which runs the planetary generation algorithm for randomly chosen planets and then performs the computational equivalent of scanning the surface looking for things like highest point or deepest valley. But why would they I guess, certainly not just to prove a point. Folks here may recall an early Odyssey dev stream where Kay Ross spoke about the "baking" process they had to go through in order to position all the planetary ports and settlements. Essentially, although those features come with a certain flattened area of ground around them which can be "pushed" into whatever terrain naturally occurs at that point to make a suitable area for the base to sit on, it still has to pick a reasonable spot to do this (i.e. it can't just randomly pick coordinates for all the bases since these could be right on the peak of a mountain). From the sound if it this baking process was computationally very expensive (I suspect it took many days if not weeks to run) and they're clearly very reluctant to run it again (not to mention that doing so would move everything again). Until they commit to doing this they can't change the algorithm used to generate planet surfaces since doing this would completely change the surface of every planet (I think it's like a butterfly effect, change one tiny part of the algorithm and the end result is completely different - mountains where there were plains, valleys where there were hills).
Last edited:
I wonder if Fdev look at random systems to check out things.
Yeah - I try not to look to hard at Analysis mode as it pushes you out of ‘imurshun’
Not that there's anything wrong with that
Horizons and hand-placed, but it worked
It's the ones half-way up the slope that would be a mess.
There would be no advantage to do so, they can't possibly check more systems than the players themselves do, other CMDRS have visited in excess of 170,000 systems, I expect the combined number of systems visited by the top 100 explorer CMDR's would be in excess of 1 million. If you think FDEV by randomly looking at a few system could actually have a chance of finding anything madder than the things we have found, then that's pure fantasy.
Oh yes, would very much like to know what's going on.
That's a terrain stamp you can see on a lot of worlds, and while it can vary with canyon depth and width they don't come remotely close to what Horizons could produce.
From what I've seen your sharp, narrow mountain(s) is a similar stamp.
Horizons can produce realistic looking canyons and ridges (ie. wouldn't look out of place on earth) that could run for hundreds of kms and made me want to fly or drive to explore them, I've yet to see anything in Odyssey (myself or screenshots) that makes me want to do that.
Odyssey can look absolutely stunning, visually way better for screenshots than Horizons when it works, but geographical the terrain is really, really dull by comparison.
I was wondering if it was a stamp or whether there was a procedural generator for that type of terrain; I suspect the former, which is a pity because an algorithm to do the latter doesn't sound like it should be impossible. Is the exact pattern repeated? I've seen the same type of terrain a few times.
Have to say I didn't find Horizon's canyons particularly convincing or appealing; better if you're in to racing, sure, but I felt very much seen one, seen them all when it came to exploring even if they weren't the same one repeated.
I believe the large shapes are stamped, but what's in them (the geology/features - mountains, hills, dunes, ice cracks) is generated algorithmically. Maybe all of those micro features are also stamps, I'm not completely sure. And then craters are generated on top of all that.
![Benjumanji [VR]](https://forums-cdn.zaonce.net/data/avatars/m/42/42439.jpg?1626566342){kind=avatar}
Yes, the lighting and colours etc are nice, but all terrain just now seems flat, smooth and rounded; boring. It's as if the universe has aged and they've implemented a few billion years of erosion to planet surfaces. Pretty colours on pretty flat rocks.
1) Of course FD checks out a few random planets. Those proof-of-concept screenshots don't just suddenly appear on their harddrives.
2) They don't check out random planets to find 'madder things than we have found'. They do it as part of QA, to see if anything immediately broken or bugged jumps out. They dont know exactly what they are looking for, so it really has to jump out.
After this and other checks it is ultimately up to the player base to really go nuts on hundreds of thousands of planets and report issues. Some are easily fixed, some will take a long time, some never get fixed.
There's like, millions of already discovered planets out there, surely it would be far better to check out planets they can already view before and after the changes they make, it would seem silly to run the seed on random planets when I have scanned and mapped something like 300,000+ planets that they can view any time and get before and after shots. If they are looking for a particular atmosphere or environment to see what the changes make then just opening one with the correct conditions from the database of millions they already have would make far more sense than randomly generating system in the hope of seeing a suitable one.
For instance landable atmospheric planets with oxygen atmosphere are quite rare and they would have to run hundreds of systems to find one, and they might not find one after all. Why wouldn't they just do a search on their existing enormous database of planets to find a suitable one, running random seeds just doesn't make sense at all!
In QA, when you test a new procedure you will always, at some point, randomly sample when doing a general scan for [whatever]. Always. Why? Because you are checking for unknown problems, which can exist in contexts you'd never consider. At this stage there is no 'suitable planet', they are all equally suitable. At some point they may want to look at specific situations for a variety of reasons but a random selection is always part of any sane QA procedure whether we're talking about game development, the automotive industry or HEI QA.
Fuzzing - Wikipedia
This isn't that sort of situation, a random selection here isn't going to give you any different results, they could run millions of system seeds and see nothing different. Every system is generated by the system seed each time it's opened, so opening a known system is no different to opening a random system, just that they already have a database of what that particular known system should have in it. Opening a random system tells them precisely nothing, because there's no way to know what changes their adjustments have made to that system, whereas opening a known system and seeing what's changed will tell them something.