Is this right?

[Guest Guest]

deleted per user request

[Thomm]

Adam,

While you are at it, please test LOS from those prone troops to the ground below! The engine still has problems drawing LOS across sharp terrain edges, IIRC, because the LOS is drawn from center-of-mass and not from the head. Is this the case in your example?

Best regards,

Thomm

[Guest Guest]

deleted per user request

[Thomm]

While prone they can see up to about 10m in front and then they have no LOS until about the 300m mark.

Combine this with the missing gun elevation limits and you have some interesting programming challenges for the Afghanistan engine at hand ...

Let's wait for 1.20 and see!!

Best regards,

Thomm

[MikeyD]

I'm not sure I see the problem. You want the AI to distinguish between the top half of head and the prone lower body when calculating LOF? I suppose there should be a calculation about how loose the soil is on the top of the hill as well. Incoming rounds hitting low could skip on rocky ground and continue to the exposed head, I suppose. This could get awully complicated if we let it.

[Thomm]

I'm not sure I see the problem. You want the AI to distinguish between the top half of head and the prone lower body when calculating LOF?

This has nothing to do with AI, but with the fact that you cannot model a 'mountain ambush' scenario, where soldiers lay prone behind a terrain edge with only their heads and weapons protruding visually, but not for the engine, because LOS starts at the center of the body, which is well behind the edge. This is not imagined by myself either, it is a real problem that affects combat in mountaneous terrain.

I suppose there should be a calculation about how loose the soil is on the top of the hill as well.

Incoming rounds hitting low could skip on rocky ground and continue to the exposed head, I suppose.

I would not be surprised if this was already modeled. Again, unrelated subject.

Best regards,

Thomm

[Pandur]

i think its totaly unimportant if you can see only the head and weapon or half the body "in the game". every type of terrain got some hidden cover/concealment value and this should be figured in there.

about LOS extending from the center of mass of the "unit";

heck either we have 1:1 repres, wich cant be like in real world. you cant calculate LOS with a 50mm x 50mm(or less) action spot grid with computers we have now.

however i would take the 3-man counters we had in CMx1 again, but there arent too many wich would. so take it as it is, is my opinnion to it.

[Paper Tiger]

The LOS 'problem' is a result of ELOS. Irrespective of where their eyes appear to be in the game, for LOS purposes prone troops are at the same level as the ground they're in, kneeling troops one level higher, standing two levels higher.

This does tend to make it difficult for prone troops to spot enemy movements and ambush them as they can't see them. What you want is a tweak to the TAC AI to enable hiding units occassionally to take peaks, one unit shifts to 'Kneel' for a few seconds and then goes prone again. Funnily enough, I thought that the v1.10 patch enabled this behaviour. I remember seeing Mark Ezra's video entry for the Marines module that showed US Marines poking their heads up above a wall and then ducking down again. It looked a bit weird at first until I realised what they were doing.

So do the units in this test poke their heads up for a peak from time to time or are we just taking a 'snapshot' of the situation and judging the game behaviour from there? If they don't maybe we'll have to report this as a bug.

[Guest Guest]

deleted per user request

[Guest Guest]

deleted per user request

[Apocalypse 31]

This brings up a good point.

when you deploy a MG team from the top of a building, they normally go prone (looks like they're prone behind the top of the roof - as in, dont have line of sight on anything because they're up against the edge of the roof)

[Flanker15]

Indeed, I've seen BFC say that in CMx2 there are no invisible cover modifiers. Shots have to hit the character model and for things like trees to block shots, they have to hit the visible tree model.

I guess you could have the LOS of a prone infantry be drawn from the head instead of the middle, don't know how tricky that would be though.

[Paper Tiger]

It's my understanding that if a bullet intersects the soldier's 3d mesh in the 3d world, the soldier is "hit".

A bit like real life then

I know, that's not the point you're making.

If that is true, then kneeling soldiers are perhaps twice as exposed, more or less, than prone soldiers. If the soldiers need to be kneeling to fire it would seem they are more exposed than they ought to be.

It's still a LOS problem because in your screenshot and in post #3, they need to be kneeling to see more than 10m and therefore need to be more exposed than if they were hugging the ground. To fire, they need to be see their targets. I think there's a bug if they don't occassionally poke their heads up to spot because that would allow them to see that an enemy unit has entered their covered arc and execute their ambush. Otherwise, we might just be bumping into one of the game's abstractions.

[Guest Guest]

deleted per user request

[costard]

Should be a cover modifier in the negative for this situation - I think soldiers avoid being outlined against the sky for good reason.

[Pandur]

What you want is a tweak to the TAC AI to enable hiding units occassionally to take peaks, one unit shifts to 'Kneel' for a few seconds and then goes prone again.

iam not sure if i want this after all i keep spotting kneeling syrians buddy aiding someone and shoot em down, the other prone ones are much harder to spott naturaly. i would see this as unnessesary exposure. if they dont see, i move em somewhere else.

You're right. While prone they can see up to about 10m in front and then they have no LOS until about the 300m mark. While kneeling they can see everything

Adam

is this a problem? the old LOS system let you see through the crest and fire through solid ground, this was fixed. so really where is the problem?

soldiers tend to line up at the border of a action spot. means if this border happens to be a crest like in the picture they will crawl up to the "top" of the crest but wont, or rarely, go "over the top" to see down the other side. in front of any crest wich looks roughly like this will be a dead area you cant see if the troop are prone. although in this example 300m are quiet a lot dead space.

also i have a question, how did you get all the guys to kneel? i mean i never saw this "like" that befor. when i let em "quick" to a crest like this, most are prone and just a few are kneeling. when i let em crawl to the crest, all are prone, so i dont feel they are exposed too much in my view.

or do you mean the way prone soldiers position themselfs behind a crest is problematic?

[Guest Guest]

deleted per user request

[Pandur]

ok i see, now i have an idea. by the looks of it the "first" rough LOS check could be the problem.

we know that there are 2 LOS checks, the first to see if the tile can see the other tile, the 2nd between the few of the 64 action spots withing the 8x8 action spot wich are occupied by the soldiers, to determine who has los and who does not.

i hope this is right and i understood how steve explained it.

now by the picture of youres, the crest doest look straight, but jagged and the soldiers dont line up in a straight line.

when you use the "los/target" tool you can see it extends from the center of the unit but in the picture the center of the unit is likely behind the crest and not on top.

the picture makes it hard to see but i think this could be a problem. try a crest wich is as steep as this one but totaly straight.

if you have saved this, try posting a shot from 2D isometric like view with the LOS tool pointed somewhere. also did you had the sight totaly blocked in prone position or possibly a gray line or a reverse slope no aimpoint thingy?

to summ this up i think if the game doest belive the "unit" can see the ground, it doesnt check the single soldiers, where some of them could see there.

[hoolaman]

The LOS 'problem' is a result of ELOS. Irrespective of where their eyes appear to be in the game, for LOS purposes prone troops are at the same level as the ground they're in, kneeling troops one level higher, standing two levels higher.

This does tend to make it difficult for prone troops to spot enemy movements and ambush them as they can't see them. What you want is a tweak to the TAC AI to enable hiding units occassionally to take peaks, one unit shifts to 'Kneel' for a few seconds and then goes prone again. Funnily enough, I thought that the v1.10 patch enabled this behaviour. I remember seeing Mark Ezra's video entry for the Marines module that showed US Marines poking their heads up above a wall and then ducking down again. It looked a bit weird at first until I realised what they were doing.

So do the units in this test poke their heads up for a peak from time to time or are we just taking a 'snapshot' of the situation and judging the game behaviour from there? If they don't maybe we'll have to report this as a bug.

100% prone squads will often get blocked LOS on marginal ground like ridges or low walls due to the way the spotting system works.

Having them default to kneeling DOES make them slightly more exposed in game terms, but the benefit is all troops are always able to get LOS and fire their weapons. They will only do the kneel thing if they are behind a fairly sharp ridge that gives them some cover.

I've come around to the view that this is the best option, with a bit of tweaking using hide and unhide, you can choose maximum cover (and sometimes have LOS blocked) or choose maximum situational awareness (and expose yourself more). Its the same deal behind low walls, if you want your guys to remain out of sight you can guarantee that with hide and guarantee they will jump up to spot when unhidden.

[c3k]

My thoughts; it seems that LOS is from center of mass (ish), and adjusted for elevation of the soldier. Prone has LOS from ground level; kneeling from about 1 meter, etc. Here's where the LOS/LOF falls apart....

If my guy is standing, his center of mass is pretty much OVER his feet. If CMSF elevates that point by 1.5 meters, that point in space is pretty darn close to his chest/shoulder. About where his weapon is held. That's excellent. That's for outbound fire. Now, for INBOUND fire, anywhere on his "mesh" that an incoming round hits, it gets accounted for.

Again, that also seems to work.

However, if my guy is prone, his center of mass is NOT beneath his feet. It is under his belt buckle. That is NOWHERE near his weapon, head, eyes, etc. Yet, his head can (and will be) hit by incoming fire. He cannot spot the enemy, but the enemy can see him.

This is only true for prone positions.

Of course, your guys go prone when they approach the edge of a piece of cover. That's exactly the time you'd like the LOF/LOS to be accurately drawn from where their eyes and weapons would be.

I think the fix is to shift the center of mass (or however the game engine determines the LOS/LOF start point) to the point of ground where the HEAD is located over, rather than the belt buckle.

This would be a HUGE improvement....

Anyone agree or disagree?

Thanks,

Ken

[Guest Guest]

deleted per user request

[cool breeze]

The problem looks a little worse than it really is. see, the camera taking the screen shots is at man head height but the los heck which shows no LOS to the prone men is against the ground height. I wouldnt be srprised if the prone men do a fine job of seeing standing syrians within most of the 300 meter dead zone. Im not saying its not off a bit but I bet that makes it somewhat better than we are seeing now. The message it would say is "reverse slope target" Some day when we all got better computers Im sure it will work better. maybe next patch too THanks BF and friends here for working on it!

[Guest Guest]

deleted per user request

[cool breeze]

I know its failing against the terrain itself and that that is a problem but im saying that since the terrain is well below the camera the problem is a little exaggerated in the pics. put something the same level as the camera (like a standing man) and check the LOS to that and you may get different results. just in case I didnt make myself clear.

[Guest Guest]

deleted per user request