Spring RTS Engine

More or less appropriated from Gabba's suggestion for Supreme Commander here. Simplified so that it is manageable within Spring.

It is a slightly modified LOS system, which I think could work better then the current one, without the need for Fog of War, as well as allowing for deeper strategy, and a somewhat more believable LOS design. (with less Krogoths appearing from what appears to be two feet infront of you)

Basically, it involves giving the units themselves a "visibility value" (probably in the FBI) which determines how easy they are to spot. Or, as Gabba puts it, a "camouflage value". This visibility value is combined with the standard LOS value designated to units, to determine whether an enemy unit is visible or not.

The gist of it is that a small unit would have a low visibility value, while a large unit would have a very high one. This is because a Krogoth is far easier to spot, from further away, then a flea, or a Peewee.
This fixes some realism issues with large units, where immersion is somewhat suspended as enormous units pop in and out of LOS range at what appears to be a really short distance.

I think such a system would work better in Spring, where it looks like FoW implementation is nowhere in sight, and the LOS overlay is too abrasive and slow to update to play with constantly. In this way, large things will appear further out then small things, which makes sense. It makes the whole unit spotted thing a little more dynamic then the current system, which can be a little odd at times.

It also improves strategy, because it means that light units are not just quick, they are also difficult to spot, so flanking with them, or using them as a commando raid becomes far more reasonable then trying to sneak a Krogoth or an ATAT around your enemy. Also remember that in Spring, until units have been spotted visually, targeting them is very innacurate. This gives more strength to smaller units, as they are harder to get a visual lock on, making it more difficult to target them. This, in turn, could add credence to using patroling scout units to spot for your forces, a strategy that seems somewhat neglected of late.

Finally, this value doesn't just have to be related to size. What it also does is effectively give mod makers a "sliding scale" of cloaking/camouflage. Instead of having "yes, this units is cloaked, and therefore invisible", it becomes far more advanced in the level of control people have, so that a unit can be given a level of "camouflage", which makes it harder to spot (has a low visibility value), but is by no means a cloaking unit. Things like commandos for SWS/AATA, sniper bots in other mods, etc.

Gabba goes on to suggest some more complicated things, where the visibility value is more of a variable then a constant value set in the FBI. Things such as Visibility Values changing according to whether the unit is mobile or immobile (good for ambushes), what terrain they are in (forest, fog, etc), and more, however, I've simplified it into the above to make it more relevant, and more reasonable to implement into Spring.

What are your thoughts?

i really like it.


Someone make a patch

A very, very good idea. Especialy for infantry, who's greatest armour (in real life) is their ability to remain undetected.

First off: I hate this for OTA. However, one could simply make it work fine for OTA by having the default be 1.0 and forget that it's there.

On the surface, this seems redundant with cloakradius, but cloakradius is a fixed radius, while this "visibility" concept correlates with the sight range of the spotter. That's got some merit. The catch is that any mod that uses this will have problems with usability - players won't be able to rely on knowledge of the view ranges of their units, because view ranges are so variable.

This could be very handy for stealth-oriented mods, as it means that long-sighted units are useful for spotting low-visibility units. So unlike in OTA, where peepers/towers are no better at seeing a cloaked unit, here we get units that make a good "anti" for sneakers.

To extrapolate this further, one could have terraint types (or even field-generating features) that lower the visibility of units around them, so that one could have a representation of shooting from cover. StarCraft and the BattleTech boardgame both have the concept that, when under trees/etc. your units are harder to hit. Here, the lower visibility provided by the trees could push the units into radar-targetting, making said units likewise hard-to-hit.

The basic, simplest concept for making it map-oriented would be to use terrain and indestructible features (so you have some indestructable tree-looking features creating a "forest" area, and units inside the forest are low-visibility".

Thing is, what you lose in terms of exact knowledge (of sight radii) you can potentially gain in terms of intuiviness.

Me likes. Pxtl has good points, though.

t'would be awesomeness. wouldn't mess with old mods such as AA either as you can have a scale of say 100
where 50 is 'visible in LoS'
100 is always visible (superweapons in C&C games)
75 is 'visible outside LoS'
0 is always invisible
25 is 'visible within LoS'

Then you could have another variable
SeekSearch = 0/1/2/3
(default is 0)
1 makes those with longer vision ranges better at spotting units inside their LoS
2 makes those with longer vision ranges better at spotting units outside their LoS
3 makes those with longer vision ranges better at spotting units inside their LoS and outside their LoS

Might work if it was based on a unit's metal cost, but also defineable elsewhere? Similar to how mass is done currently. I don't want to break too much from the TA formula though.

Caydr wrote:Might work if it was based on a unit's metal cost, but also defineable elsewhere? Similar to how mass is done currently. I don't want to break too much from the TA formula though.

I wouldn't base it on mass at all - I'd assume a visibility of normal (call it 1.0) and simply have a tag for a multiplier (2.0 = visible at twice view range). That way TA compatibility is maximized without needing to add a visibility=1.0 to every unit to maintain TA-style gameplay.

And it could be relatively easily "grandfathered" into TA-based mods. Take AA - start by giving the L3 mechs larger detection ranges, see how that works out. The work down from there, through the larger L2 units (the tanks) to the smaller ones and the L1 units (pretty much everything else), then handle the really tiny stuff like the Flea and Dragon's Eye.

NOiZE wrote:i really like it.


Someone make a patch

I Agree

1) I love this idea - it should be done, I think, as the original poster suggested, and should use a separate tag which functions as a multiplyer in the LoS calculations, and defaults to 1 if not specified. Also, I would suggest that units with low visibility but no stealth should appear as a radar dot if they are within the spotter's LoS but not within 'visible' range, whether or not they're not actually covered by radar.

2) Not sure if this would be a great idea for AA - it would probably require a lot of arm/core rebalancing (core get gimped by larger, heavier, more visible units) to work properly. Still, might be worth looking into if you're willing to spend time on the balancing.

Soulless1 wrote:1) I love this idea - it should be done, I think, as the original poster suggested, and should use a separate tag which functions as a multiplyer in the LoS calculations, and defaults to 1 if not specified. Also, I would suggest that units with low visibility but no stealth should appear as a radar dot if they are within the spotter's LoS but not within 'visible' range, whether or not they're not actually covered by radar.

2) Not sure if this would be a great idea for AA - it would probably require a lot of arm/core rebalancing (core get gimped by larger, heavier, more visible units) to work properly. Still, might be worth looking into if you're willing to spend time on the balancing.

For (1) I'd simply make units have a radar range equal to their sight range. Being unable to detect small units within your sight range would only be a problem if the game makes heavy use of jammers - and most mods don't go for AA's obsession with jamming. No need to add a special "I know units are in my sight range but I don't know what" radar blip, since not everybody would want that feature.

Make the values customizable on both the spotter's side and the target unit's side. Each unit would be able to have a "sightrange=x" and a "visibility=y" value. Viewer's sightrange (as a percent) times viewee's visibility (as a percent) equals total modifier to the ordinary spotting algorithim. For example, if a unit has visibility=0.5 (is spotted from 1/2 the ordinary distance) and an opponent unit's has sightrange=2, the first unit is spotted from the ordinary distance.

This would allow for specialized observer units i.e. watch towers, radar towers, and would allow things like planes etc. to have more realistic chances to spot. Not everything is equally good at spotting enemy units.

This would be excellent for A&A Spring.

no one thinks my suggestion is simple enough to be implemented???

Min3mat wrote:no one thinks my suggestion is simple enough to be implemented???

To be honest I didn't really understand your suggestion

The thing is, Caydr, that we've already broken the TA LOS formula to an extent by not having a fog of war system. Without standard fog of war, the TA LOS system doesn't really work.

And I wonder what the system strain would be if such LOS calculations were implemented? I can't see it being particularly stressful, but I realise that LOS calculations can be quite complicated. What would the stress be, firstly with a constant value set in the FBI, as I have suggested, and then the more advanced option, if it was a value that could be modified according to the terrain. ie: if there is a tree nearby, divide Visibility Value by half, etc.

Warlord Zsinj wrote:The thing is, Caydr, that we've already broken the TA LOS formula to an extent by not having a fog of war system. Without standard fog of war, the TA LOS system doesn't really work.

And I wonder what the system strain would be if such LOS calculations were implemented? I can't see it being particularly stressful, but I realise that LOS calculations can be quite complicated. What would the stress be, firstly with a constant value set in the FBI, as I have suggested, and then the more advanced option, if it was a value that could be modified according to the terrain. ie: if there is a tree nearby, divide Visibility Value by half, etc.

Correct me if I'm wrong, but isn't the "fog of war" gameplay there? I mean, it's not visibly represented (unless you live in the fugliness of L-view) but it's there. We all still think in terms of "fog of war" - the idea that "I can see any unit in X range of my forces unless it is cloaked".

I have a quick question. Warlord, why don't you post here very much anymore?

I mean its not like every post you make is as insanely brilliant and fantastic as this one, but I mean...we miss you. Or at least I do, if no one else does

Indeed I would enjoy seeing this in Spring. It would allow me NOT to have to go into spectator every time I wanted to take movies like this one:

http://video.google.com/videoplay?docid ... 4578430159

Ps. Was origionally created for Cybernations.