24 replies to this topic

[Last Of The Brunnen-G]

Since we get new maps recently i want to point out that this is a lost cause. Sooner or later the players will know every new map well and some sort of meta behavior will occur. Most of the times it means that everyone runs into the center, even if it's not domination and dies quickly.

I think that procedural generated maps would be much better, because they would force players to think more about their general tactic in the unknown enviroment.

With MW5 PGI has shown, that they are in general able to make procedural maps, even if they are not very good. With some smart brains behind the idea there could be really good maps.

[Horseman]

Does CryEngine 3 even support that?
That aside, remember that the map generation process would have to be either fast on user-grade computers (good luck with that) or the maps would have to be generated server-side and then downloaded (which increases load times whole hell of a lot).
I don't think that's viable. Random placement of objectives might be. Randomizing existing maps by loading them from several pre-designed combinations of geometry might be. Hell, just a few different versions of the same map with minor changes (ex: Crimson where the mountain isn't blocked, or the tunnel is partially collapsed etc) might add unpredictability.

Edited by Horseman, 29 June 2021 - 09:04 AM.

[pbiggz]

Procedural does not mean no work.

Often, it requires more.

[Heavy Money]

Yeah CryEngine 3 isn't gonna work for that.
Modular maps with some randomization might work though, but its still probably a larger amount of investment than is worth making. If they're going to put a lot of money into the game, it'd be better to make MWO 2 in a modern engine.

[pbiggz]

You can get alot more mileage out of existing maps with dynamic objectives, respawns, and weather conditions. They'd require a bit of engineering time but would definitely be worth the trouble.

Imagine Terra Therma, Caustic Valley, or Hibernal Rift where whole map is choked in sensor garbling black ash fall, or a raging thunderstorm on viridian bog, or canyon network. Imagine vitric forge where the nasty O-type star hanging over it is in a flare state.

I wont say these would be easy changes, but they'd absolutely be worth the time to do. Combined with rotating objective locations, and dropdecks, you could make MWO quick play a much more healthy game mode.

[MechNexus]

Yeah no, this'd be terrible. Procedural maps would be an incoherent mess full of chokepoints and twists, or be repeats of polar highlands - and if you try and shackle the generator too hard, they'll probably come out samey.

Even if it was possible in cryengine, hard pass. Procedural maps are almost never done in competitive multiplayer FPS for a reason - Ace of Spades' first version is honestly the only one I know of that does this, and that only worked because of the digging & building mechanics.

Edited by MechNexus, 29 June 2021 - 09:55 AM.

[FinnMcKool]

you would probably do just as well if you randomized the spawn points, never knowing where the other team is at the beginning of the match would mean more need for scouts, different areas the battle would take place,

the best funniest matches where always the "Odd" ones, those that for some reason didn't go the same as the rest, the battle was off to one side or an other on the map

always spawning directly opposite at farthest ends of the map always leads to people running to the center, you know the general location of the other team, your just racing to get there first

[Monkey Lover]

They can't even do something as simple as random spawn points.

[Horseman]

Monkey Lover, on 29 June 2021 - 12:53 PM, said:

They can't even do something as simple as random spawn points.

I believe at one point it was mentioned that the dropship positions are fixed, yeah.

[Daneel Hazen]

I think about this a lot... maybe so in the future.

[LordNothing]

pbiggz, on 29 June 2021 - 09:06 AM, said:

Procedural does not mean no work.

Often, it requires more.

and the work involved is usually programming, amd mwo gets very little coder time.

however there are a lot of good offline procedural generation tools. the game engine need not know how to generate terrain, you generate geometry in 3rd party software and then import the geometry/textures into the map maker (assuming cryengine editor lets you do that).

Edited by LordNothing, 29 June 2021 - 06:33 PM.

[MW Waldorf Statler]

LordNothing, on 29 June 2021 - 06:31 PM, said:

and the work involved is usually programming, amd mwo gets very little coder time.

however there are a lot of good offline procedural generation tools. the game engine need not know how to generate terrain, you generate geometry in 3rd party software and then import the geometry/textures into the map maker (assuming cryengine editor lets you do that).

Thje Cry 3 engine has this auto generating of Terrain as feature ...problem ,most you have more Work to bring this Terrain in a Playable Version,as build it from Hand ( 90%impassable Mountains for example,objects in the Terrain.)

In MW5 im seeing by many Maps Problems like Modules thats have Holes to the generated Terrain ,or hovers a little over the Terrain , and thats mit Multiplayer bring massive Problems with glitchs for example, or you drops in Objects or Terrain while the Droprange from the Module in Collision with the generated Terrain.

[pbiggz]

Games that use procedural generation tend to spend a solid majority of their development time on the world generators.

I promise you this is not the catch all we are looking for and I say that as someone who used to think this was a good idea, then learned about what goes into it.

Monkey Lover, on 29 June 2021 - 12:53 PM, said:

They can't even do something as simple as random spawn points.

Random is a buzz word. We want mixed spawn points. Random from a set of pre-defined spawns just doesn't sound as snippy though, but I think its firmly within the realm of possibility.

[Thorqemada]

pbiggz, on 29 June 2021 - 09:06 AM, said:

Procedural does not mean no work.

Often, it requires more.

Yes and No...

Yes - the initial work is obviously more
No - the follow up work is way less and in the long run pays off

Ofc if done right...

Edited by Thorqemada, 01 July 2021 - 02:56 AM.

[MW Waldorf Statler]

how manyMultiplayer Shooter have good looking ,not generated looking procedual Terrain ?! Ok ,Ark have a Genaerator (and many generated Maps looks good, many looks terrible) and Ark only genereted Natural Terrain without each settlements or itys(thats a own big problematic Theme bei PT)

[My Lord and Saviour Jesus Christ]

Last Of The Brunnen-G, on 29 June 2021 - 08:49 AM, said:

With MW5 PGI has shown, that they are in general able to make procedural maps, even if they are not very good. With some smart brains behind the idea there could be really good maps.

You admit that MW5's ProcGen isn't good so why ask for it in MWO when good map design is even more important since its PvP?

[pbiggz]

Thorqemada, on 01 July 2021 - 12:24 AM, said:

Yes and No...

Yes - the initial work is obviously more
No - the follow up work is way less and in the long run pays off

Ofc if done right...

I majored in game development, I know what I'm talking about.

[Thorqemada]

Well, i despise premade maps that you can memorize and look for the best way to play in YT Videos and How Tos.
I rather play an average procedural map that has a layout i have to explore and find a on the fly strategy/tactic than a good map i have played 1000 times and sleep away while doing so while my mussle memory takes over my conscience.

And that there are no good procedural maps only shows how the promises of the past all have be left abandoned with AI getting worse, gameplay getting worse, UI getting worse, even manual mapmaking getting worse and the only creative things come from modders that often have no academic education.

Yeah, every now and then, every once in a decade there is a game that is in some things better than what we had in the past but then none of them meets what was predicted/promised back in the 90s.

The latest show of inability was Cyberpunk and while i had my fun and had no issues i am lucky bcs that game was never on my radar as i ever expected it to be a shooter game and i usualy do not play such games yet it turned out to be an action adventure but really, it failed to fulfil its promises like a desert fails to be an ocean.

And that is what most games are about today, they cant even code a good football manager anymore (aside of the Sega one and i use 14 or 15 realism/difficulty mods to make it a challenge) and that is mostly data based number crunching with next to no creativity involved (looks at "We are Football").

Everything is in decline, it may be a grafic hog (soon™ not anymore) but the gameplay is in steep decline usually.

Edited by Thorqemada, 02 July 2021 - 04:13 AM.

[MechNexus]

Thorqemada, on 02 July 2021 - 04:04 AM, said:

Well, i despise premade maps that you can memorize and look for the best way to play in YT Videos and How Tos.
I rather play an average procedural map that has a layout i have to explore and find a on the fly strategy/tactic than a good map i have played 1000 times and sleep away while doing so while my mussle memory takes over my conscience.

And that there are no good procedural maps only shows how the promises of the past all have be left abandoned with AI getting worse, gameplay getting worse, UI getting worse, even manual mapmaking getting worse and the only creative things come from modders that often have no academic education.

Yeah, every now and then, every once in a decade there is a game that is in some things better than what we had in the past but then none of them meets what was predicted/promised back in the 90s.

The latest show of inability was Cyberpunk and while i had my fun and had no issues i am lucky bcs that game was never on my radar as i ever expected it to be a shooter game and i usualy do not play such games yet it turned out to be an action adventure but really, it failed to fulfil its promises like a desert fails to be an ocean.

And that is what most games are about today, they cant even code a good football manager anymore (aside of the Sega one and i use 14 or 15 realism/difficulty mods to make it a challenge) and that is mostly data based number crunching with next to no creativity involved (looks at "We are Football").

Everything is in decline, it may be a grafic hog (soon™ not anymore) but the gameplay is in steep decline usually.

[redacted]

Procgen tech is just fine - however, the way it works just wouldn't suit a multiplayer FPS. Procgen works best at large, expansive spaces that are pretty to look at, but it struggles a lot at making something that's fun to fight in, which is when they require human assistance.

Edited by Armchair General, 02 July 2021 - 06:52 AM.

[Last Of The Brunnen-G]

Thorqemada, on 01 July 2021 - 12:24 AM, said:

Yes and No...

Yes - the initial work is obviously more
No - the follow up work is way less and in the long run pays off

Ofc if done right...

There are many different levels of procedural generation. On one end we have games like Star Citizen, which generate whole planets (mostly empty). On the other end we have something as simple as settler of catan, with a modular pieces that can be combined in different ways.
What i have in mind is a map generation that uses different modular pieces that are combined with world parameters and texture layers. Very simple top bottom aproach. Here an example:

1. solar system generation (this just gives basic map characteristics like gravity, radius, moons and suns visible, day night cycle)
2. Biome generation (gives the overall setting, which terrain tiles to choose from)
3. map layout generation (uses the tiles from step 2 to generate a map with choosen characteristics (urban, rural, etc.)
4. Season/wheater generation (gives a layer of snow or makes gras more green, etc.
5. Building generation (fills the tiles that are supposed to have buildings with buildings of a specifiy theme and gives them thematic visual enhancements like faction related flags, signs and paintings)