Map design philosophy

Ok I understand that the game will have some procedural generation of maps. With that understanding, I have some questions about Uber's philosophy of map design/generation:

1.As a design principle, will your maps tend to be path & chokepoint-based, like Starcraft/Supcom2, or more open like the first TA and SupCom1?
I ask because the rocky desert concept gives me the impression of constricted corridor based maps. They did it like this in SupCom2... and it was a big mistake.

2.Will you allow for the kind of maps that have rolling terrain (gradient slopes) like in TA, or will all elevation be step-based planes like all mainstream RTS?

This should really be in the General Discussion Sub Forum.

Anyways;

1] The Important thing is that there is variety, we don't know the entire ins and outs of the system Uber has in mind, but I don't have a problem with Desert Biome focusing on Corridor style setups.....so long as the other biomes aren't the same. Also consider that the images we've seen so far might not be based on the generation algorithm. There could be large rolling 'plains' made up of sand dunes as well.

2] everything seems to be point ot the TA/SupCom style, but keep in mind there isn't anything intrinsically wrong with step based planes, I even hop they'll show up in PA(in moderation) to represent built up installations or (abandoned) city themed maps.

Mike

So far, everything points to fully customizable worlds. Be on the lookout for a little bit of everything, at least until real gameplay can test it.

This is a very tricky one. Projectiles are meant to follow real paths, so rough terrain can mess up direct shots. The shots that end up blocked are seemingly random, so it can cause wildly unpredictable gameplay. On the other hand, it also makes ballistic units more desirable as a way of avoiding terrain issues.

Rough terrain could very well be a "feature" of certain biomes and future rage threads, but I expect it to be the exception rather than the norm.

That should be your signature

It should be in everyone's signature because it's true.

The first thing they said when they showed those images was that they were no real indication of how much height variation an ingame desert biome will really have, and were just concentrated with terrain features to show off the concepts.

So it's hard to conclude anything from it.

1.Well, I don't mind both. I just want to make sure that it's not going to be entirely one thing. I'd like to see chokepoints, but not as the norm. In fact for this desert setup, I'd like to see both huge open areas, as well as constricted rocky areas - all on one planet.

2.Step based planes.. May be "ok" only for specific kinds of geology (desert + strong winds vs rock). Not good for mountains or regular hill formations.

Idiots.

With proper terrain detection and Line of Fire calculations, like in Spring, the effect of obstacles stops being random and starts being tactical and strategical elements of the battlefield.

bobucle, "rage threads." Really? I fail to see how rage is a feature of this thread.
Is any thread that asks a critical question considered "rage" perhaps?

In what way would a projectile's path be "more real" in the setup you seem to be proposing? (i.e. no rough terrain, just "corridors"). If anything, it seems to be the opposite to me. The paths are more "real" when they are being calculated on the fly, taking all angles and factors into consideration. Not when the projectiles are predetermined to travel somewhere.

As godde explains, projectile paths is not a big problem for implementing rough terrains. In a nutshell, having a proper Line of Sight, and proper firing solutions, the kinetic and ballistic projectiles will travel and fly in dynamic manner. Some will hit, some will miss. Missiles will lose lock-on because an airplane is travelling too fast or at odd angles. An artillery shell will hit a hill, because you made a tactical blunder when examining the projected paths and the geography. It's not random, it was once touted as a feature in SupCom. Do you remember how Chris Taylor made us all excited about "emergent behavior" for SupCom? This is it.

If an artillery shell will hit a hill, as it tries to hit an enemy behind, it should not fire. The artillery should then fire at a target that is not blocked by terrain.
In SupCom and TA units on highground would sometimes fire at an enemy below them but because they weren't placed on the edge they would hit the cliff and maybe even damage a friendly unit infront of them.
Controlling this behavior is micro heavy and hard to predict. I'd rather have the unit determine whether or not it will hit before firing.

But that would mean that all the weapons in the game must have 100% accuracy and never miss. That is how mainstream RTS games work, and tainting PA with it would make it more boring. I can remember that most artillery pieces in TA and SupCom did not always hit their mark. What I mean is that the chance of the artillery piece hitting the hill may be just 10%-20%, but doesn't mean it won't happen. By aiming farther onto the flat area away from hills, there's a greater chance to cause more havoc with fewer misses. Could be the enemy placed some very strategic structure below that hill on purpose.
I hope they don't veer too far off into Starcrapt territory with predeterministic projectile paths, but rather semi-physics based like SupCom.

Let me correct my statement. If the unit has free Line of Fire it should fire at the enemy. If accuracy makes it miss or if the unit dodges the projectile is another question. If there is a hill in the way the unit should not waste its' shots shooting straight into the hill.

Yes I see what you mean now. I was thinking of a projectile traveling in a high arc, similar to a howitzer or mortar. With artillery, there rarely is "free line of fire." Especially not the monster artilleries we had in TA and SupCom. You could have a vertical projected firing-arc be steep enough on the way down that it grazes the hilltop on 80% of the shots, but misses and hits the hill on 20% of the shots. This would be in line with the accuracy of artillery pieces in the TA / SupCom games so far. I remember the shot landing very far from the target many times.

Like this.

The artillery piece would pick the best solution automatically, depending on your target. And then it would fire. I don't recall any artillery pieces in TA or SupCom telling me it can't hit my target because of some obstacle. But it may be a good idea.

http://www.collinsdictionary.com/dictionary/english/line-of-fire
I extend this definition to all types of weapon trajectories when I use this term in RTS.
So if an enemy can be hit by a weapon it is in Line of Fire of that weapon regardless if the if the projectile is lobbed or a direct(straight) Line of Fire weapon.
If an enemy can't be hit and is in range it is because the Line of Fire is obscured by obstacles.

So what mechanics can artillery with ballistic projectiles use?
1. Low trajectory: The howitzer in that picture.
The weapon can aim up to 45 degrees and the Line of Fire is obscured by obstacles
2. Interchangeable projectile velocities:
By varying the projectile speed a projectile could arc over small obstacles at closer ranges from which it would only to arc over from higher ranges because of angle of approach on the projectile. Artillery like the UEF Lobo in SupCom could change their muzzle velocity which made them able to arc over small obstacles even at close ranges.
3. High trajectory: Mortars for example.
The weapon can aim above 45 degrees. The weapon can fire above extremely steep obstacles at close ranges but at longer ranges as the weapon goes towards 45 degrees it can't fire above as steep obstacles.

When a weapon can be used both at Low trajectory and High trajectory everything within range can be hit from 2 angles of approach except at max range. If an obstacle is blocking the low trajectory path the weapon might be able to use the high trajectory path to hit the enemy from a higher angle of approach.
If a weapon has interchangeable projectile velocities there are an infinite large amount of angles of approaches except at max range. A weapon with these characteristics could chose the optimal path in this case probably being the fastest trajectory that minimizes the inaccuracy and travel time to target destination.
This was however not the case with the UEF Lobo in SupCom since it the projectile speed was set depending on the lateral distance(I think) to the target so there was only 1 possible angle of approach to 1 single target location. The Lobo could for example be unable to fire at high cliff from short range but be able to fire at that hill from longer range.
A ballistic weapon could have a mix of these 3 mechanics.
Anyway. This wasn't what I were suggesting.

I suggested that units should only fire at targets that are in Line of Fire so that they don't waste their time shooting into hills and cliffs.

I'm pretty sure it was possible in the Spring engine to force an emplacement to fire anywhere in their radius. If there was a terrain obstacle in the way it just meant the projectiles never hit their intended target. Mobile units however attempted to move to gain a clear line of fire before shooting.

It might have been like that at some point. But I think you need specific weapon tags in order for the weapon to fire regardless if the terrain is blocking or not now.

The maps are generated, so it would make sense to add a few variables for this.
Please make it moddable Uber!!

Variables that come in mind:
- number of mass/metal points, amount generated per sec etc
- elevation
- space between hills etc
- mountains between starting points (passable by planes yes no)

etc. In the end you want to be able to generate a maps like they existed in supcom, this allows different playing styles.

Extended on this it would be nice to generate neutrals and enemies depending on a number of variables (AI defend/attack/neutral, repair losses, attack after 20% of the units are destroyed etc)

I think it would help a bunch if the attack cursor would include some form of "likeness to hit" indicator. We've already seen the ones that show you roughly what kind of area shells will be landing in, now if we can just add in "odds of the shell not even landing in this area" by changing the color or something, you'll have all the info you need when selecting targets.

One thing that comes to my mind when speaking of randomly generated maps: How do you ensure the map is balanced for each player? What if one player starts in a crater, where he needs aircraft to expand out of, while the other player can expand with much cheaper ground units? Maybe the maps should be sort of mirrored? I'm not a fan of maps that seems too mirrored, but it definetly makes it fair for everyone. Anyone have thoughts about this?

Or maybe the battlefields will be pre-made by the community, using the editor and saving the solar system layouts? This way maybe some layouts will become "standard" for fair games.

I think the random algorithm will build the map, and a second one will scour the map for starting locations that would be fair to all players. That seems like the most logical way to go about it.

That, or generating fair starting locations is built right into the random map generating algorithm, so that any generated map can optionally include a number of fair starting locations.

Or, you let players pick their own starting location, although that might get messy if two players want to start in the same location or something.