Support NeoGAF

[Men_in_Boxes]

1st time playing on a new map:

18th time playing on the "new map":

I get the resistance to procedurally generated level design as a single player gamer. Those games are designed to be played once, and for players to feel what the artists want at specific points. Nathan Drake pulling back on some brush to reveal an idealic mountain range is going to produce an emotion that a new Minecraft seed can't.

That being said, multiplayer gamers are not single player gamers. We value entirely different things.

SP gamers value:
- Production values
- Passive narrative
- Being guided
- Predictability
- Ease

MP gamers value:
- Novelty
- Uncertainty
- Player choice
- Difficulty
- Social play

If the above is true, why are multiplayer studios still hanging on to a design philosophy from the single player era? We're in the GAAS era where games are not only designed to be played for years, but developers stay connected post release.

Luke Stevens just released a video where he said he played Marathon for two days. He played on the first map one day and the second map on day two. He felt the levels were so small that he knows them pretty much inside and out.

Bungie...no, PlayStation...no, ENTIRE VIDEOGAME INDUSTRY: Take your level designers out back, and pretend they're race horses with broken legs. It's about procedural generation now.

How much more interesting would Marathon have been if Bungie had a team of 30 people working on a modern AAA procedural generation system for 6 years?

This is a problem in multiplayer that no one is talking about. I play The Finals.

There's 7 maps in The Finals and each map has 10 PoIs that attract the most fighting. At a certain point, you learn the optimized strategy for each point. I play X character with Y load out so I really only have 6 or 7 ways to defend this point. Actually...I learned a lot over the last month...there's really only two or three viable defense points. Actually I played for 6 months...there's one ideal defense position.

If you play The Finals (any multiplayer game really) you quickly learn the geography and just slot into your position. You don't even have to talk to teammates anymore because the class I play always sets up in one or two locations.

Static level designers are a product of a bygone era.

 

[March Climber]

Are you saying roguelike PvP? Sounds very ambitious.

 

[John Bilbo]

Depends on the game. Not good for fighting games or Counter Strike.

 

[Men_in_Boxes]

Are you saying roguelike PvP? Sounds very ambitious.

What is PvP if not a roguelite? These little indie studios are making massive hits because they're figuring out how to keep their game fun for a long time. Valheim, Minecraft, Terraria, No Man's Sky.

Every time you play those games you have to "disvover" the new PG world.

Most PvP maps lose their luster because we've seen the same thing for months on end.

 

[StueyDuck]

I would prefer some form of server meshing and making MP interactions sparse.

If devs insist on BR and extraction games then make those moments rare and special.

Landing and getting instantly into fight with 30 sweatlords who just watch streamers all day and learn stupid new metas and exploits becomes worse and more brainless than playing against bots

 

[StreetsofBeige]

Would make maps much more interesting.

Even just randomly adjusting small amounts of the map like barrels and box placement, adding an extra floor to a building, turning a car wreck a different angle, weather, day/night, would add a lot to variety.

 

[Codes 208]

Halo gets around this by frequently updating the roster with forge maps, i think i prefer that over random elements in a competitive multiplayer game where map memory has no real advantage.

Buuut for co-op centric games I'd be totally down for that. One of the reasons why phasmophobia is one of my favorite games is because each game doesnt feel the same as the last because of randomized elements (different ghost, different cursed item, different bone location, ect)

 

[LordCBH]

Honestly I could see some games being pretty fire with something like that.

 

[Buggy Loop]

Eh I dunno

Part of my love for MP games was knowing the nooks and crannies of a map. Tribes with the slops, counter strike with all the sniper spots, etc

 

[Hawk The Slayer]

Especially for horror multiple games.

 

[IDKFA]

1st time playing on a new map:

18th time playing on the "new map":

I get the resistance to procedurally generated level design as a single player gamer. Those games are designed to be played once, and for players to feel what the artists want at specific points. Nathan Drake pulling back on some brush to reveal an idealic mountain range is going to produce an emotion that a new Minecraft seed can't.

That being said, multiplayer gamers are not single player gamers. We value entirely different things.

SP gamers value:
- Production values
- Passive narrative
- Being guided
- Predictability
- Ease

MP gamers value:
- Novelty
- Uncertainty
- Player choice
- Difficulty
- Social play

If the above is true, why are multiplayer studios still hanging on to a design philosophy from the single player era? We're in the GAAS era where games are not only designed to be played for years, but developers stay connected post release.

Luke Stevens just released a video where he said he played Marathon for two days. He played on the first map one day and the second map on day two. He felt the levels were so small that he knows them pretty much inside and out.

Bungie...no, PlayStation...no, ENTIRE VIDEOGAME INDUSTRY: Take your level designers out back, and pretend they're race horses with broken legs. It's about procedural generation now.

How much more interesting would Marathon have been if Bungie had a team of 30 people working on a modern AAA procedural generation system for 6 years?

This is a problem in multiplayer that no one is talking about. I play The Finals.

There's 7 maps in The Finals and each map has 10 PoIs that attract the most fighting. At a certain point, you learn the optimized strategy for each point. I play X character with Y load out so I really only have 6 or 7 ways to defend this point. Actually...I learned a lot over the last month...there's really only two or three viable defense points. Actually I played for 6 months...there's one ideal defense position.

If you play The Finals (any multiplayer game really) you quickly learn the geography and just slot into your position. You don't even have to talk to teammates anymore because the class I play always sets up in one or two locations.

Static level designers are a product of a bygone era.

This is actually a fucking great idea.

It would make multiplayer games a lot more interesting as every match would be a new experience.

Hopefully somebody has a go at this in the not too distant future.

 

[Loxus]

League of Legends tournaments can benefit from something like this.

 

[March Climber]

What is PvP if not a roguelite? These little indie studios are making massive hits because they're figuring out how to keep their game fun for a long time. Valheim, Minecraft, Terraria, No Man's Sky.

Every time you play those games you have to "disvover" the new PG world.

Most PvP maps lose their luster because we've seen the same thing for months on end.

Yea, what you're claiming you want is much, much more ambitious than a roguelite.

A PvP game where you can never get used to a map, because you will get a new map every time. It would mean every game would be more tense and slow, instead of CoD-like where people are running at 100mph to the best spots.

A new toolkit would have to be made to make sure that would make sure no map is too badly designed.

 

[Men_in_Boxes]

A new toolkit would have to be made to make sure that would make sure no map is too badly designed.

The great thing about GAAS is you can constantly tweak, add to, and improve the PG algorithm.

Your game design would certainly need to be built around it but landing on alien planets is going to result in sub optimal maps from time to time. That's half the fun.

 

[March Climber]

The great thing about GAAS is you can constantly tweak, add to, and improve the PG algorithm.

Your game design would certainly need to be built around it but landing on alien planets is going to result in sub optimal maps from time to time. That's half the fun.

Unfortunately I think your idea is so good that someone might discover this thread and steal it.

I consider this idea on the level of when the Virtua Fighter devs made VF Quest Mode. I've been hoping for them to eventually bring it back now that A.I. has been the trend lately.

 

[Redneckerz]

We have had several dozen games not even long ago that did this. First example that comes to my mind is Shadow Warrior 2.

The thing with procedural generation is that it isn't that procedural enough. Its still much or less the same map no matter how you slice it.

I prefer handmade maps and a singleplayer campaign and a no-frills multiplayer. I personally don't care about service games, or dogtags, or in-game currency, or tiktok skins. I know that i am not the primary audience for those games anymore (Because the primary audience is kids with too much time and money to spend) but as a mid-30's fella with game interest, that's where i am at.

My anchor of reference is stuff like UT99 and Quake 3. I know there are dozens of open source titles that mimic this, but please, one big budget title that does this would be great.

 

[March Climber]

We have had several dozen games not even long ago that did this. First example that comes to my mind is Shadow Warrior 2.

The thing with procedural generation is that it isn't that procedural enough. Its still much or less the same map no matter how you slice it.

I prefer handmade maps and a singleplayer campaign and a no-frills multiplayer. I personally don't care about service games, or dogtags, or in-game currency, or tiktok skins. I know that i am not the primary audience for those games anymore (Because the primary audience is kids with too much time and money to spend) but as a mid-30's fella with game interest, that's where i am at.

My anchor of reference is stuff like UT99 and Quake 3. I know there are dozens of open source titles that mimic this, but please, one big budget title that does this would be great.

I figured it might already be thing, but from your post it sounds like they aren't implementing it properly. Other than Shadow Warrior 2 what were some other PvP games that had it?

 

[ChorizoPicozo]

Procedural systems give devs a lot of bang for the buck and can extend gameplay exponentially... maybe they are not smart enough to implement them .

 

[Avid Reading Horse]

This game was kind of fun.

Stapled on randomly generated maps and an economy system on top of Siege gameplay. Random maps kind of work here because Rainbow has always had the concept of a planning phase, so being able to look at the map and strategize for a bit is fairly natural.

But Siege already had the actual solution to stale maps. Architecting and destruction. The maps are complex enough to where it takes hours of running around just to have a basic functional understanding of them. But then the real work of understanding of how to set up or attack a site begins. What kind of crazy lines of sight you can make through the floorboards. What kind of last second flank you can make through a wall.

I don't think random maps work that well for any kind of standard FPS objective based gameplay, because you don't get to leverage deep knowledge of the map to succeed, and that's one of the most satisfying things about a good FPS. It would probably be good for TDM and the like, just because it forces people out of their favorite camping spots... they'd at least have to rub together two brain cells and find a place to pitch their tent.

As for BR, could probably work there, that's all about adapting to the circumstances the circle thrusts upon you. It takes a long time to really learn a BR map, and it definitely makes you better at the game, but it doesn't really make the game better in the same way as learning a map in Siege does. Earlier days of say PUBG where you pull out your map, try to figure out what the best spot to go to might be, and how to get there was a lot of fun. Or desperately trying to outrun a late circle while trying to figure out what sad hay bale or minor feature of topography might keep you from getting crossfired from 5 different angles.

 

[Shubh_C63]

I am a novice MP guy but isn't mastery over a MP game comes from learning map and player habit patterns ?

A procedurally generated map would highly reset the skill to CoD level get seen first got shot first.

 

[Begleiter]

I am a novice MP guy but isn't mastery over a MP game comes from learning map and player habit patterns ?

A procedurally generated map would highly reset the skill to CoD level get seen first got shot first.

Procedural generation isn't the same as making perfectly random maps, they normally observe rules that can be themselves be learned. How much it rewards skill also depends on other mechanics present, radar being the obvious one.

Still a bad idea, mind.

 

[Fafalada]

I would prefer some form of server meshing and making MP interactions sparse.

MP is about cost effectiveness not breaking the bank. The whole idea of distributed compute scaling just isn't commercially viable, unless you can convince players to spend about 10x more than they do on today's games.

If devs insist on BR and extraction games then make those moments rare and special.

I mean there's extraction games like Hunt that try this, but it's still in predefined maps, so people adapt quick.

 

[Herr Edgy]

I thought this a while ago when playing Dark & Darker.
The atmosphere of the game is good, despite all the jank, but what sense of 'Dungeoneering' do you really get when you generally know where all the traps are and what room awaits you to your left?

Thing is, PCG doesn't need to be 'random slop', you can set it up to work as coarsely or as granularly as you want. Make the rooms by hand, connect them randomly. Doable. Just a lot of work.

 

[Killjoy-NL]

One of the most annoying things in PvP is not knowing the maps.

This is the worst possible idea to implement, unless you're going for a niche audience.

 

[Zuzu]

That sounds like it could be a really good idea for at least some multiplayer games even if it's just an optional mode in addition to maps which are more static and unchanging. I don't know why it hasn't become a widespread feature but perhaps too much randomness is too detrimental to multiplayer design. Maybe gamers want limited randomness alongside a healthy amount of predictability. So things like changing loot drops, changing weather patterns, perhaps some environmental destructibility add limited randomness without the entire map layout changing.

 

[ValiantInstance]

I'm 90% sure most of Black Ops 6's maps were designed by AI. So you tell me if that's a good thing.

 

[bender]

Designers have a hard enough time good maps with static layouts.

 

[Wildebeest]

Random content doesn't normally fly in multiplayer games because of the synch issues. Similar problem to complex physics simulations. A lot more would have to be done on server side to make the game fair and consistent and that means games feeling less responsive and being more expensive to run. If you are talking about random maps then for an ideal experience you are probably talking about something where the server would have to recalculate a "new" map before play then each client would have to download the whole map before they could start their game, each time they play. You can make that more efficient by moving from a full Minecraft terrain gen type system to a system where the server just shuffles up pre-made tiles that are already installed on the client. Something that simple can be done peer to peer.

 

[Sybrix]

Why is Counter Strike the most popular fps for the last 25 years?

People are playing 25 year old maps on CS2 now.

Great maps will shine forever supported with great gameplay. Most FPS games these days have weak gameplay.

 

[Sentenza]

I still can't tell if Men In Boxes is a very committed troll or genuinely incapable to nail a single topic.

 

[Wildebeest]

I still can't tell if Men In Boxes is a very committed troll or genuinely incapable to nail a single topic.

It is like he imagines himself as the champion of the MP gamer but has never actually played one.

 

[winjer]

This is a good way to get games that play, look and feel very generic.

 

[IAmRei]

Eh I dunno

Part of my love for MP games was knowing the nooks and crannies of a map. Tribes with the slops, counter strike with all the sniper spots, etc

agree, it will lost usual strategy if it's too random
i guess, it's better add more map, or spice it with new things
beside, there should be another genre for that.
usual MP are should be memorable for the gameplay side

 

[T-0800]

 

[TheUnicornGuy]

A tile based system like diablo might be a good compromise, big maps that are built from many smaller tiles. This would allow learning of geographic features without being able to learn the entire map.

 

[DKehoe]

Depends on the game. I could see there being ones that could benefit from it. As you've said, with something like a roguelike unpredictability is part of the appeal. But for most skill-based multiplayer game then consistency and balance are key.

Level design is a skill in itself too. Human designers craft maps with an intentional flow, sightlines, chokepoints, and balance in mind. If a procedurally generated map drops me next to a high-powered weapon and a bunch of power-ups while you have nothing useful nearby, that's pretty unfair. Good level design ensures a level playing field so that skill and strategy, not RNG, determine the outcome.

A football pitch or a basketball court have a consistent size from one game to the next. What changes, and what makes each game exciting, is how players use their skill and adapt, strategise, and perform within that consistent framework. Multiplayer games work the same way.

 

[Men_in_Boxes]

Why is Counter Strike the most popular fps for the last 25 years?

People are playing 25 year old maps on CS2 now.

Having a set map where all players understand them fully makes much more sense in a competitive setting:

Having an unpredictable map, where you're learning as you go makes more sense in "Heroes Journey" type games (like Marathon):