Discussion about vision. (if at all)

Demo added to the first post: viewtopic.php?p=687594#p687594
It's even simpler to understand than i thought. The mechanics behind it might not be so simple (and perfect sensor placement could become a thing for math majors), but the produced result is intuitive enough, especially if you had never played a game where radar and vision had been separated.

Not an issue any more, a single T1 tank possibly won't even show up at all on your screen, while a Galactic Colossus might even be tracked while still under construction on the other end of the map.

Tiered radar would have similar effects, but also requires LOTS of tiers to produce the same quality of information.

It has. I had planned to make almost the same topic, but you cut me to it by a few days while I was still preparing the data.

Looks great.
Though i wanted to discuss the impact of no FOW had on the gameplay .

Different side of the same coin i think.

the point of FOW is that it adds uncertainty. If you can see what your enemy is doing, it removes the element of surprise. i think it makes the game more difficult to win.

What a better radar system would do, is ensure that uncertainty still exists.

Not only the element of surprise, but it also means that you play with perfect informations. That means, that you can always build the perfect counter straight ahead, given that you have sufficient APM to observe your enemy close enough.

You often do have perfect information in round based games, like chess or other board games, but you also have sufficient time to analyze these informations while it takes also only a few rounds to execute a strategy.

In a large scale RTS, things work quite different, you have very little time for analyzing the massive amount data, but strategies take quite a long time to execute because you have very limited movement speed.

Playing a small scale RTS with perfect information can be fun, frantic, but fun. Trying the same in a large scale RTS will lead to a flood of useless information and enforces passive behavior because every aggressive strategy is going to be countered right away and with ease. Finally it only leads to massive encounters in open field and whoever has the better economy will win.

A good implementation of FOW and radar does not only hide the enemy strategy, but also aggregates the most important informations for you, so managing a whole planet is just as simple/difficult as observing a small area in early game.

This seems overly complicated, both for the defender and the attacker. It's hard for the attacker to know at what point they are spotted, and it's hard for the defender to know how much radar coverage is needed (It also means having 5 radars next to each other helps, and spamming radars doesn't seem fun).

Also, complexity starts to rise dramatically when you consider that the map is not on a 2d plane but can be an arbitrary, possibly multi-level 3D shape (and across many planets).

I think that intelligence needs to be non-perfect (omnidar was a cheap way out for countering stealth), shorter ranged but easier to deploy than Sup Com. Maybe some way of launching sensor drones that temporarily give radar in a smallish area but with a long launch range?. This balancing problem is tricky given satellites pretty much give you perfect intel on an entire hemisphere once they are launched.

As an attacker, you know you're spotted when they start shooting at you.

You know you're really spotted when they start hitting you.

That was a pretty silly thing. Vision already counters stealth, so adding in a second reveal was completely redundant. Omni was also necessary to reveal cloak, which turned out silly yet again when a giant laser appeared out of nowhere. Gosh, maybe I should follow the laser back to where it comes out of thin air and just shoot that! :lol:

Great. So you don't actually know what's going on until half your army is getting blown up.

Some people don't know what's going on even when half their army /is/ getting blown up!

Then again, traditionally the two opposing factions in a war don't tend to know exactly how much of their movements the enemy has intel on...

The purpose of the jammer is to hide larger units when a stealth field isn't sufficient.

Also since the jammer produces alot of noise the enemy won't know if there is an army there or not even though the enemy know that the player is using a jammer.

I kinda like how vision is handled in R.U.S.E.

The thing about sattelites is: they can see half the world, but half the world can see THEM too. And the other half can still fire missiles at them.

If they can fire a missile at your satellite without your defences being able to stop it, they can hit your *base* with said missile. Just park the satellite above your base/anti-missile and it becomes a LOS unit for half the world.

Just because they can hit your base doesn't make it a better target. You seem to be thinking about a big nuke to take out a satellite, but I'd think more along the lines of a small tactical missile. You can either take out an expensive satellite, or a single factory.

In such a case the satellite is a better target, and tac-missiles don't have to use the same defenses as nukes. Also, a satellite doesn't neccesarily have to give vision of half the planet.

Well, a satellite may give sensor coverage to part of the planet, but that coverage may be very low resolution. Even today's satellites have to get pretty specific about where they point the camera to get a high resolution image. I would find it reasonable to suggest that spy-sats need to use 'passive' sensing technologies, thereby rendering them less effective but broader ranged than a radar emplacement. This would make them hard to detect / take out with anti-sat missiles.

Sats could also have a vision CONE, an extremely long reach with a small field of view. That's what they do IRL.

The IRL ground equivalent is an observatory.

I love exterminians idea, that sounds awesome!

Only if your base is equatorial. If it isn't, you can't realistically park a satellite above it, orbital mechanics do not work that way

Found this topic after a search, I reckon a raytraced "actual" LoS would offer up some interesting variance on the default Fog of War concept. It's not like light has some arbitrary maximum travel distance, I can see Sirius fine in the night sky and it's 8.5 light years away, and I certainly have no issues seeing the horizon from the top of a hill.

So regarding the original question regarding removing Fog, I would suggest that instead of an on/off toggle, swapping the (now standard) Fog of War plus artificially limited circular LoS, for "actual" LoS (but not full map vision), would be my preferred solution. This is assuming that current hardware is capable of making those calculations without bogging the engine down too much, and also assuming that the devs are happy to build an engine that has both as options, rather than just sticking to one.

It does massively devalue radar; both sight and radar have similar occlusion characteristics (contrary to popular RTS implementations, radar does not go through mountains, especially iron rich rocks) so the main strength of radar is the detection of metal objects that are too small (or camouflaged) to be directly recognised in a field of view.

Afaik we know nothing at all how uber will implement FoW.

But we can safely assume that "real" LoS is not going to be the only type of intel we get, in case it will be used at all.

Why?
LoS on spherical bodies works terrible. Most units move very close to the ground, the diameter of the sphere is low and therefore is the vision range. You would barely have to worry about obstruction at all, the horizon for each single units would be only about a dozen steps away. Go forth, do the math yourself if you don't believe me.

You mean that vision range would very closely match what works best in typical RTS games?

Sounds good to me.