Jump to content

Question about maps!


Recommended Posts

1 hour ago, alienman said:

Got a few map repeats in my current ongoing playthrough. So, I'm just wondering. Will the game be like Xenonauts 1, not having randomized maps?

And if that is the case, how come not? Technically too complicated?

Randomised maps are fully supported by our map editor (and were in Xenonauts 1 too), but in general the final maps produced don't look right. You can see this in Xenonauts 1 where modders and mappers tried to make random maps and the only ones that worked fairly well were randomised wilderness areas like forests. Any real world areas that contain man-made areas are constructed logically by humans and it's very obvious to a player when that coherence is missing.

Some maps can support a degree of randomisation fine, for example the town maps you see in terror missions have consistently-sized "blocks" where theoretically an office could be replaced with a shop of similar dimensions to give some randomisation. But this doesn't allow full randomisation, as for example the gas station building can't be placed where you would place an office because that needs extra road connectors to allow cars to enter the gas station. This is an example of human logic - people would notice if a gas station spawned and there were no gaps in the pavement to access it, because it looks wrong. This is where random maps usually fall down and the better the graphics in the game are, the worse the problem is.

The reason we don't do stuff like the office / shop swapping is because then you could randomly get a map generated where every building is an office - instead, I designate a particular space as being an office or a restaurant or whatever and there's usually 2-3 somewhat different layout variants of that building the game can pick. This is just personal choice; I'd rather have to create more static maps than risk any map being able to generate a version of itself that looks dumb.

The game should have a system that tracks the maps you play and ensures you don't get repeats unless there's no other maps to pick, but I've not tested it thoroughly and it's possible it needs some fixes before it works properly.

  • Like 1
Link to comment
Share on other sites

On 11/22/2022 at 4:10 PM, Chris said:

The reason we don't do stuff like the office / shop swapping is because then you could randomly get a map generated where every building is an office - instead, I designate a particular space as being an office or a restaurant or whatever and there's usually 2-3 somewhat different layout variants of that building the game can pick. This is just personal choice; I'd rather have to create more static maps than risk any map being able to generate a version of itself that looks dumb.

This is a personal design choice of course but I really disagree with this one. Map variety (the lack thereof) was one of the biggest issues in X1 as originally released. It's a major issue in X2 now, even though more maps will surely come. But even now in X2 (and I haven't played that much), I know what the entire map is like on turn 1. Oh, terror mission and I walk out in front of a gas station? Okay, that's the one with a restaurant on the next block and the roof of the grocery store that offers great firing lines. Woods map with one car in front and a workshop to the right of the dropship? That's the lumber yard map, trains up in the corner provide good cover and another workshop building on the opposite end is tricky to clear.

X1 was the same initially, I knew the farm and industrial maps like the back of my hand.

The problem with hand-crafted maps is that they take a lot of time to create and you need many to get the replay value up. Xcom-style games are very replayable but that replayability requires map variety. And it's not easily addressed by adding another two maps per biome, you need an extra dozen maps per biome to make any difference.

To me, the possibility of "every building is an office" sounds great! Why couldn't I have one fight in a dense office tower district?

I would consider the solution that X:CE adopted. In X:CE, there are the base game maps, the wonderful community-created static maps, but also random (actually procedural) maps. The key being, random maps are only picked when there are no other maps to choose from, there's also a repeat-tracking system (it's actually a bit probabilistic IIRC but that's details). With that, map repetition is not a thing in X:CE, you're very unlikely to encounter any repeats within a single campaign. 

Procedural forest/desert/arctic maps work wonderfully. Yes, the procedural maps can suffer from illogical placement in urban areas and it's visible with examples like a gas station with no entrance as you mention. But I find that to be a far, far lesser violation of my immersion than heading to the third terror mission in the same city block, where I blow up the same gas station again before turning to aliens in the same diner as last time.

By the way, Firaxis learned this as well. Their first Xcom had static maps, which may have had a minimal degree of randomization with how cover was placed. That was consistently pointed out as one of the game's main weaknesses. Their Xcom 2 went for procedural generation, and that was very well received despite the occasional weird placement of stuff.

  • Like 4
Link to comment
Share on other sites

Solver, first I suggest you play the actually Version [V.24.7B] to Day 320 and try out the full Maps you can play [Base-Defenses, Base-Attacks, Terror, different Bioms etc.]. I did it and must say the Maps have many surprisses.

Like Alienman already said, all Developers of such Games have the same Problem. Either you get unplayable 08/15-Maps (Computer-Generated) or excellent Handmade Maps (like in XCOM 2, Xenonauts 2 or UFO ET) with the later Oportunity of dynamic Generation for them (like in UFO ET Gold-Edition / UNI-Mod, XCOM 2, Phoenix Point, Battletech).

An X:CE-Modus you won´t get for Xenonauts 2 about the complete difference in Programming to the Predecessor. But an light Modding Modus is planed for the Final Stage of Xenonauts 2 or after the Game is in the Main-Programming for playability completly finished.

Link to comment
Share on other sites

26 minutes ago, Alienkiller said:

Either you get unplayable 08/15-Maps (Computer-Generated) or excellent Handmade Maps (like in XCOM 2, Xenonauts 2 or UFO ET)

You can't laugh endlessly at the same joke.

A silly, but new joke is more interesting to listeners than a good, but old joke (a joke heard many times). It's the same with maps.  In old and repetitive maps, there is nothing new and unexpected for the player (there is no intrigue).

Edited by Komandos
  • Like 2
Link to comment
Share on other sites

I played V24 to V24.6. Wasn't able to get to the time limit but after day 200 I was definitely playing on the same map, like the lumber yard, for the third time in the campaign. The maps are great. But I'm pretty sure I've seen all of them, at least on some biomes, there are a few maps I am very familiar with by now and I haven't even been able to get to the endgame yet.

  • Like 2
Link to comment
Share on other sites

Komandos, I don´t make Jokes in that. I played the Game to Day 320 in that Beta-Version yet and have seen much surpisses in the Maps which were not half bad, esp. with the Alien-Advantages which come in now Step by Step.

Solver, like said, the Devs work still on the Maps, Geoscape, etc. There they aren´t finished yet and we will see some more Adjustments in the Early Access Phase later on. For now the Maps must be Bugfree as possible and fully playable. Then the Devs can think about bringing in the Feature Dynamik Map Generation like it´s in XCOM 2, UFO ET-Series, Battletech etc.

When the Game goes in Early Access-Phase evtl. the Maps are complete finished in Bughunting / Playability. Then we will see more Dynamik Maps from them.

Edited by Alienkiller
Link to comment
Share on other sites

3 hours ago, Solver said:

This is a personal design choice of course but I really disagree with this one. Map variety (the lack thereof) was one of the biggest issues in X1 as originally released. It's a major issue in X2 now, even though more maps will surely come. But even now in X2 (and I haven't played that much), I know what the entire map is like on turn 1. Oh, terror mission and I walk out in front of a gas station? Okay, that's the one with a restaurant on the next block and the roof of the grocery store that offers great firing lines. Woods map with one car in front and a workshop to the right of the dropship? That's the lumber yard map, trains up in the corner provide good cover and another workshop building on the opposite end is tricky to clear.

X1 was the same initially, I knew the farm and industrial maps like the back of my hand.

The problem with hand-crafted maps is that they take a lot of time to create and you need many to get the replay value up. Xcom-style games are very replayable but that replayability requires map variety. And it's not easily addressed by adding another two maps per biome, you need an extra dozen maps per biome to make any difference.

To me, the possibility of "every building is an office" sounds great! Why couldn't I have one fight in a dense office tower district?

I would consider the solution that X:CE adopted. In X:CE, there are the base game maps, the wonderful community-created static maps, but also random (actually procedural) maps. The key being, random maps are only picked when there are no other maps to choose from, there's also a repeat-tracking system (it's actually a bit probabilistic IIRC but that's details). With that, map repetition is not a thing in X:CE, you're very unlikely to encounter any repeats within a single campaign. 

Procedural forest/desert/arctic maps work wonderfully. Yes, the procedural maps can suffer from illogical placement in urban areas and it's visible with examples like a gas station with no entrance as you mention. But I find that to be a far, far lesser violation of my immersion than heading to the third terror mission in the same city block, where I blow up the same gas station again before turning to aliens in the same diner as last time.

By the way, Firaxis learned this as well. Their first Xcom had static maps, which may have had a minimal degree of randomization with how cover was placed. That was consistently pointed out as one of the game's main weaknesses. Their Xcom 2 went for procedural generation, and that was very well received despite the occasional weird placement of stuff.

I have offered my help to Chris and would be willing to map few dozen high quality maps for X2, but it seems there are a few technical and legal hurdles that I think are unsolvable to make it happen unfortunately.

Edited by Skitso
  • Like 1
Link to comment
Share on other sites

I agree with Solver, whatever immersion or gameplay value is lost with the randomness of procedurally generated maps it is more than made up by the fact I won't be able to immediately do an MTV Cribs style walkthrough of the map once I step out of the dropship.

I think some form of procedurally generated maps would be a great thing to test in v25 or 26.

Link to comment
Share on other sites

Legal issues would be dealing with Unity licenses and commercial plugins I assume? I loved your X1 maps @Skitso and I'd certainly be happy to help with technical issues for 3rd-party mapping after the Early Access builds, if there's something I can do to help that.

I realise now I hadn't even mentioned map variety in my V24 feedback - this is because I had assumed (and still assume) that the lack of map variety is a known problem that will be addressed. I'm sure there will be more static maps but my assumption was also that procedural maps would be used to some extent.

  • Like 1
Link to comment
Share on other sites

@Solver Yeah, I think the issues are mainly related to unity licenses.

 

What I've gathered, @Chris is well aware of the issue of the game having too few maps ATM. I'm sure they are doing at least something to make it better, but the amount of work required there is huge, to be frank.

I would probably be able to produce maybe 2-3 maps per week after learning the tools. If I remember correctly, it took maybe 3-5 hours per level in X1

Edited by Skitso
Link to comment
Share on other sites

Yes, and that's exactly why X:CE ended up with (semi-)random maps eventually. For long-term replay value, you simply need too many maps for manual creation to be viable. Randomized maps are the only way to achieve that. Even limited randomisation, such as replacing buildings or rotating pre-defined square submaps adds massive value. This kind of game is at its best when you cannot predict what's going to happen and rotating a grocery store 180 degrees is a significant change because suddenly the lines of sight and fire are all different. So I firmly believe that X2, like any xcom-style game, needs 30-40 quality map templates that are then randomised at runtime, giving you thousands of possible maps. With purely static maps, you'd need a team with several map designers working on maps for 6+ months to provide a decent level of variety.

And for several biomes, procedural maps have no in-game disadvantage I can think of. Forests, jungles, deserts and polar areas don't follow much logic, they can have any layout and the huts / cabins / vehicles, if any, can be more or less anywhere. Whatever mistakes could happen are certainly less counter-immersive than me recognising the specific jungle map right away so I remember where the impenetrable trees are.

Link to comment
Share on other sites

A hybrid style might work as well. For the biomes that have little to no issue with randomization, they can be random, while biomes that are more difficult to randomize like cities can have more handcrafted maps to compensate for the lack of randomization. I’m not sure how well that would work in practice, but it might be a potential way of combating the problem.

Link to comment
Share on other sites

I think the Devs will make it better then in UFO 1 ET with the Randomisation which come in the 2009 / 2010 Gold-Version and in XCOM 2 from Firaxis.

But first the remaing Bugs and other Problems in the Ground-Fighting Maps have to be fixed completely, which are still there. They are very seldom, but still there. V.24.7B is in that a big Step forward in the remaining Bug- / Problemfixing / Map-Reworking / Map-Upgrades, but it´s not finished completely.

There are 2 Examples I had in my 2 Gameplays to Day 320 in that Version:

1. Terror-Maps: You throw Grenades outside or in an Buildung and it work like it should. In the Alien Turn the Aliens throw 1 Grenade in an Building and then you get an light CTD.

2. Other Maps / Terror-Maps: Sometimes an friendly Soldier / Civlian Helper or an Structure get still the purple Coulor.

That have to be done first with other Things the Devs are doing in the Map-Reworks / Map-Uprgades before we can think to go the next Step. With that remaining Problems in the Maps we can´t go in the Early Access next Year. And especialy Think about Randomisation.

Edited by Alienkiller
Link to comment
Share on other sites

Surprisingly, I have to agree with Solver on this one. Nothing is more fun killing than knowing exactly what the map is going to be like, and with that you dont really need to pay attention anymore, and that lessens the fun.

During the development process of X-Division we took on what needed work the most, and for maps it was the terror missions. Because we added so much aliens the relatively small terror maps felt more cramped than usual. One of the things the team did was create different leveled versions of the same building. Eg:

Unbenannt.thumb.png.068242c08cc5b4518d7a2dff180942e5.png

Additionally we simply alternated the vanilla layout. That was a very easy copy and paste job, as we simply copied one level as textdata and pasted it one level higher. Given that terror missions are quite hard, the added variety felt like a good reward for taking on a difficult mission. Keep in mind though that we also forbid the aliens from starting inside of a building, vanilla does not contain any spawn points on higher levels, and the more aggressive AI meant that the player never really had to search for the aliens (thus we also removed the reveal after 20 turns mechanic, and later added it with 40 turns back in again).

 

However we could do better.

Quote

 

Think about this.

We make a box. We dont know exactly whats going to be in the box, but we call it battleship_map. Its going to contain our battleship submap (x1). We put 2 more boxes into the box, one in the upper half, called "upper box" and one in the lower half, called "lower box" (x2). The upper box will contain the battleship in an hovering state. So we put in the exosceleton of the battleship. We dont know exactly with what to fill it yet, but the exterior is the same in all cases. Then we think about the interior. On an abstract level you realise that all that the player is doing is traversing from level to level, until he reaches the end. However, since we have teleporters we are not restricted to topolocigal access, which would only limit us to two cases, entering the first level and going upwards, or entering the highest level, going downwards. The battleship has 4 levels. We can enter any given level, and make any traversal possible. We set the properties of the upper box to have 4 alternatives, one entry teleporter for each level. Each alternative has again 3 more choices of where to put the next teleporter, and each of them has 2 more choices, and then 1 choice. That makes 4 x 3 x 2 x 1 possibilities to traverse the interior of the battleship map, or 24 possibilities (!4). Now lets think about the number of layouts for each level, since we only have 4 levels to make layouts for, and not !4 layouts. I propose 5 alternatives for each level. You might think its hard to come up with meaningful designs for 5 layouts, but we can have a guideline for making them. There are pistols, rifles, shotguns, heavy machineguns and snipers. We create the layouts by trying to give one weapon type the most advantage in the environment. So for instance for snipers we give good cover on both teleporter sides, and no obstacles inbetween. Shotgun (X-Division Shotgun) advantage is lots of little rooms. Given the space requirements some rooms will look more similar to each other, like a rifle room and a pistol room, but small meaningfull changes are still appreciated. That gives us 5 layouts for 4 levels (x4 x5). So much for the upper box.

The lower box will contain teleport pad zero, that leads directly into the battleship, wherever the "first floor" is. It will have 8 variations (x8). It will also contain an alien watchtower, in 8 small variatons, 4 fully functional ones in 4 directions, and 4 shot apart ones in 4 directions. Though i won´t count them into the final score, because its rather small details, so only (x1). It will also contain a reference to a small scout, scout or corvette underneath in 4 of the 8 cases, though the number of variations for that is depending on the native variations of the UFOs, so are undefined at this point (x?).

In total that gives us 1 x 2 x !4 x 4 x 5 x 8 x 1 = 7860 different variations to play a battleship map for only designing 4 x 5 + 8 + 8 = 36 submaps.

When i took a look at illunaks submaps using submaps, i felt that the ulterior goal of that functionality was to enable procedurally generated submaps.
https://www.goldhawkinteractive.com/forums/index.php?/topic/12993-linked-submaps-submaps-using-other-submaps/
The most suprising thing about this is that it actually works as it is supposed to. However, some small native Xenonaut bugs ruin the fun of actually fully utlising this system. In detail ground tiles with level 0 get mixed up upon loading, which is even present in vanilla or XCE maps/coding. It doesnt outright break anything, but it makes it very hard to visually communicate to the player where for instance the teleporter pad is. Just imagine walking below a battleship, with nothing but grass below, but one random grass tile teleports you inside. There are workarounds, but i would call it essential for this to work.

X-Division uses a lot of battleship maps in the late game. Battleship, Bomber Battleship, Terror Battleship, Dreadnaught, Terror Dreadnaught and motherships. All linked to battleship submaps. So i thought that it would be a lot of fun to have procedurally generated battleship maps.

When i read illunaks post i also felt a deep connection to how strustru talks about concepts in C++ ( coming 2022, soon in your compiler). We make a function. We dont know what the function is about in detail, but we will just call it main(). This will contain what we want the program to do. Inside of that function we put another function. Again, we dont know exactly how it will work out, but we want the function to print strings. We call the function print(). And so on and so forth ... . Thats basically how the compiler library has been built up.

So i would really like to design some procedurally designed maps in A game, where i can go wild with my imagination. I feel like it will help me develope similar skills to being a programmer, while also creating a product that other people can enjoy at the same time. Xenonauts 1 doesnt have the functionality, X-Division has the need, and Xenonauts 2 doesnt have a public map editor.

These are my thoughts on it anyway.

All of this is based on illunaks contribution of allowing submaps to load other submaps. This technical aspect alone makes a lot possible. Want to make an asset which contains all other boxes ? We can do that. Want to make another asset that loads said box mix with 20% possibility ? Just put and add 2 of the box mix as a file, and 8 empty ones.

I am familiar with all the map packs and how they work (mostly) and its not an understatement to say that without the lovely added chaotic and unpredictable maps X-Division would have never come to be. Before people would add additional unit content if no map variety would have been present the gameplay loop would have quickly burned out. Nothings more fun killing than "Been there, done that, seen everything". Imagine playing 200 hours of X-Division on the vanilla maps only, and you will see how much core the mapping scene is to the modding community.
Unfortunately, Illunaks contributions came way too late. By the time the functionality was added, all the other maps had already been written, and no map content which utilises this system has been published today. The mapmakers moved on, and thats just that.

It came around that that somebody actually implemented this system, and improved on other aspects of vanilla maps too, focusing on the battleship map. Here is showcase of how that could look like:

1142599909_X_COMUFODefensedesign.thumb.png.238e1e990f8d84dc635c1a58f68cff94.png

More images of actual implemented variations can be downloaded here: https://drive.google.com/file/d/1bY-FRNoOYGcEpiQ4ILPaWGlFoagedOtU/view?usp=share_link

Here i talk roughly 2 hours about proper game design and explain the ins and outs of the system specifically for maps.

And the presented idea visualised in a CAD:

835102956_BoxShowcaseExplained.thumb.png.c4f11b148614eabc239c6c6c8a3563ef.png

 

This was all done with the functionality that XCE and Illunak provided and is already possible in X1. Now this shouldnt be an argument because it is possible in X1 it should be possible in X2, just showing that the concept removed from the implementation works. Put to the test the procedural generation of battleship maps increased the loading time by 50%. That sounds long, but if you realise you will play at least an hour on such a map its usually worth it. However as Solver once noted, the increased loading time fundamentally shows that something wrong happens in that system, and it should really not take more than 300 ms between a procedurally and non procedurally map generated. Performance is still a factor, and the faster a map loads the happier the players are, as you break immersion less with lower loading times. However, the system will have to be build with procedurally generated content in mind, otherwise you will just end up with scaling issues like in X1.

 

I strongly advise follwoing points:

  • Hire somebody that actually cares about procedurally generated maps. Whether you like it or not, randomised maps are THE factor for the longevity of a game, and a healthy modding scene. Just take a look at Wacraft 3, the public map editor kept the game alive until the game company literally killed its own modding scene, and spawned entirely new genres, like DOTA. Some programmers just really dont like to do procedurally generated maps, others do. If you cant do it yourself, hire somebody who cares for randomised maps. You will have the highest return on gains on this.
  • Make a public map editor available on day 1. Before anything else can happen for a game, there need to be maps. That means before any other modding can happen, you need the possibility to increase the map variety by the community. Just like Illunaks technical achievements were unmatched, as unmatched was the lack of actual modders who used it. By September 2015 the community scene mostly moved on to other things. So that people feel attracted to the vanilla game, and the modding scene, you need a public map editor. The more map variety there is on day 1, and/or the possibility to make their own, the longer you will attract players. The longer you attract players, the higher the chance for them to buy the next game, the higher the chance they really like your content, and the higher the chance they will start to add to the modding community.  Because people WILL move on, the question of it is just "when". Providing a bigger variety increases the time window people spend on your creative content, and therefore increase the time window you can paint yourself in a positive light.

 

All of this is a statement about the biggest bang for the buck that the IP can get, not necessarily how possible it is to achieve it. I have not been filled in about the details of the legal issues for a public map editor, or the situation GI is in. But again i have to stress that the least amount of effort is required to implement the system I presented. The modularity of the system is fully accessible, how much effort the map creator spends on any particular asset is a free choice, assets can be reused on other maps ( additional to the "one map, different biomes" idea), and the gain of the system is exponential. By creating only 36 small submaps you can create 5 figure variations of meaningful and good looking maps.

Sincerely, Charon

 

PS: I had to chuckle at @Komandos post again, he put it, as other people would say "most expertly", in very few words.
PPS: I do love handcrafted maps, but as Solver said, simply turning an asset in all 4 directions can provide immense value, with little to no effort. As a matter of fact most 4x variations i produced are the same asset rotated in 4 directions.

Edited by Charon
  • Like 4
Link to comment
Share on other sites

The map variation is something we'll be looking at in Early Access, the goal for the Early Access launch is simply to have enough maps that people can do one or two playthroughs of the campaign without encountering many map repeats. I think we're on track for that (at least if we can get the script that handles it working properly).

As to what the solution is to support hardcore players who want to be able to play the game over and over again, I'm not sure. There's several avenues we can explore:

  • Continuing to expand the existing selection of maps and add more variation to the existing areas of randomness within those maps.
  • Giving the map editor to community mappers, assuming we're able to overcome the licensing issues we have with various Unity plugins and suitably seperate it out from the rest of the game. Unfortunately at the moment any mapper would need our entire codebase and a bunch of dev knowledge to use the map editor.
  • Implementing some kind of fallback randomised map, while being aware that the final result is going to look worse than a normal map.

The map editor is much more flexible in X2 with regards to data; for instance it already has (and uses) the ability Charon outlined where sub-maps can contain other submaps and so on. Ultimately the results are just a factor of how much time people are willing to put into asset creation and how much you're willing to compromise the end product to generate randomness.

Sure, XCOM2 has randomly generated maps but there's a reason their maps are all either wilderness maps or unfamiliar sci-fi cities where the player has far less feeling for what looks and feels "right". And while you can rotate maps and so forth to add more variation, all our maps are designed to have clear routes which ensure your sight lines are blocked and means you have to split up your team as you come out of the dropship (e.g. a block of trees or a building placed in front of the dropship ramp, etc). A rotated map wouldn't necessarily have this, so might be less interesting to play if you could just keep all your soldiers bunched or clear large areas of the map just by looking in that direction straight out of the dropship.

Anyway, yeah, I suspect we'll end up introducing more randomisation in Early Access - but I do think you'll end up with worse maps as a result, so I'll be wanting to keep them seperated from the normal maps so that only the heavy users end up seeing them.

  • Like 3
Link to comment
Share on other sites

Current maps feel quite flat. I'd like to see more two or even three story buildings for more verticality and sense of exploration. Certain enemies should prefer higher grounds (roofs, upper floor windows and balconies). Higher ground should also provide bonuses for accuracy(does it already?)

Needing to carefully clear upper floors from ambushing snipers would definitely give more flair and suspension to the missions. Especially more important missions like cleaner building where Xenonauts need to retrieve intel would feel a lot more intresting if you'd need to infiltrate higher floors for the intel.

Urban area terror missions would cerainly also feel more dangerous when multi store buildings could be full of aliens shooting from second and third story windows... etc. 

Edited by Skitso
  • Like 3
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...