Improvement suggestions

[Thewood1]

I would also say that I no longer use area fire for MGs in a lot of cases.  The above changes also seems to have given some real intelligence in target selection and not letting the MG become target fixated.

[Bulletpoint]

11 hours ago, Kaunitz said:

(hedges everywhere) led to very satisfying engagements in which firing almost never gave away the position of a unit to the enemy. Of course, in such a setting, you simply need to area-fire, and the AI cannot make use of it in a way an actual player could.

Problem being that you can't really properly area fire into wooded areas in this game, as you need the fire order to "click" into a specific square, and the trees block LOS. When you do get LOS, the trees will still block nearly all your bullets. Especially when you do 3 trees per tile as in your forest edge example.

 

[Kaunitz]

@Bulletpoint

That's true. Actually, finding a LOS into the wood from the outside is a bit tricky, but not that difficult, especially when you elevate the inside of the wood a bit (to prevent the "reverse-slope, no aimpoint"-problem). 

The second issue that you mention is more problematical. Indeed the woodland-edge stops many (not all) shells and bullets. So, firing into the edge-zone is very easy, firing into the "interior" of the wood, however, is difficult. Trees are concrete structures invulnerable and impenetrable for tank rounds in CM.

However, this problem does not directly affect the main idea of placing wall-tiles (in the form of hedges, boage) instead of foliage-tiles (trees, bushes) to provide concealment for troops at the wood-edge. It affects the visual appeal though. The dense trees at the edge are primarily an aesthetical necessity. I have placed them here to create a gapless canopy of leaves - cover and concealment was not my main intent. I can replace those trees with dense bushes that have no trunks (so they don't stop fire into the wood). But this doesn't look nearly as good as my current solution. What I would need is "higher" bushes (as tall as the smallest "G" tree). 

Inside of the wood, area fire is not a big problem, as I use a low density (only 1 tree per square) wood-layout with very tall trees (no treetop-foliage obscures LOS on the ground level).  

Edited May 29, 2017 by Kaunitz

[Bulletpoint]

5 hours ago, Kaunitz said:

I can replace those trees with dense bushes that have no trunks (so they don't stop fire into the wood). But this doesn't look nearly as good as my current solution. What I would need is "higher" bushes (as tall as the smallest "G" tree). 

I actually suggested that some time ago, but it was not welcomed by the community  

I agree with you completely that the edge of the woods should be a "green wall". Many map makers actually place a lot of bushes outside for this reason, and it works so-so to block LOS into the trees, but of course it looks mainly like what it is - some trees with bushes outside.

 

[Kaunitz]

I've been trying to delve into the whole terrain/LOS aspects of the CM, but there is not a lot of info available on the forum for obvious reasons. I can only experiment in the editor. Here is how I believe the game determines LOS (when you use the target-command):

There are hard and soft LOS-blockers. 

Hard LOS-blockers instantly block any LOS unless the LOS passes over the blocker's hitbox:

• trunks of foliage-tiles (even low bushes have very small trunks)
• wall-tiles (hedges, walls, bocage): you can imagine a hedge to look just like a wall for this matter.
• buildings
• the ground-mesh itself/hills

Soft LOS-blockers seem to decrease the LOS quality square by square. Once the LOS-quality reaches 0, your LOS is blocked. This is what an official post suggests (see below): 

• ground: Certain ground types like tall grass or forest deteriorate your LOS if your LOS actually cuts through the hitbox of the tile [I do believe there is some kind of hitbox involved, otherwise why would I get the reverse-slope-LOS and the difference between kneeling and crouching stance in my experiment (see below)]. Once your LOS-quality reaches zero, it seems as if the hitbox of the ground-tile is treated like a hard-LOS-blocker. (see experiment below)
• foliage: according to the official post, foliage acts as soft blocker that deteriorates LOS-quality (if the LOS cuts through the hitbox of the foliage). I assume it's also based on action-squares, not individual points within the squares (like the hard blockers)?

So, from my understanding, foliage and ground decreases the quality of any LOS cutting through their hitbox. I assume that LOS-quality affects spotting chances. When LOS-quality reaches zero, LOS is completely blocked. Hard blockers immediatly block LOS.

 

Experiment: Looking through soft blockers?

Since it is an easy experiment, I tested how troops can see through certain ground-types (like tall grass, forest). Note I'm not talking about foliage-tiles (actual trees), just the ground-types that you paint over the map. So, I took a perfectly flat map and a MG team in prone stance and let them look through different types of terrain. Here are the results:

• In tall grass, they had a clear LOS for a distance of ca. 11 squares. From square 12 up to square 102, they had a "reverse slope, no aim point"-LOS. LOS was completely blocked from square 103 on. 
• In forest ground (light or heavy doesn't matter here) LOS was clear up to 12, reverse slope until 42, then blocked. 
• Ordinary grass or dirt: clear up to the end of my testing range (square 125+)

The same experiment with kneeling soldiers:

• In tall grass: clear up to square 41, reverse slope for the rest of my testing range (square 125+).
• In forest: clear up to square 19, reverse slope for the rest of the testing range (square 125+)
• Ordinary grass and dirt: clear

How to interpret these results? Here is my hypothesis (see picture 1). A different hitbox-shape might explain the different result for forest and tall grass (picture 2).

What I don't understand, however, is the difference between kneeling and prone stance. If the kneeler is higher than the ground-tile-hitbox, then why is there even a reverse-slope on his LOS? If he was lower than the ground-tile-hitbox, on the other hand, then how come his "clear LOS" reaches out farther than that of his prone collegue? 

I'd also like to try out what happens when there are patches of different ground-types involved. But things get very complicated here. 

------------------------------------------------------

Official post on this topic: 

 

 

 

Edited May 29, 2017 by Kaunitz

[Bulletpoint]

44 minutes ago, Kaunitz said:

What I don't understand, however, is the difference between kneeling and prone stance. If the kneeler is higher than the ground-tile-hitbox, then why is there even a reverse-slope on his LOS?

Because his target might be prone, which means at some distance, the LOS from kneeling guy to prone guy would pass through enough tall grass to block the LOS?

[Kaunitz]

Oh, that would make perfect sense. Thank you! It's the "if you stand up I can see you because I'm taller than the grass"-effect (see picture). If this sketch was true (remember this is all purely speculative) then allowing area-targeting on reverse slope spots would be a nice addition - simply because noone is safe from suppression and bullets behind a LOS-blocking wall made of grass . 

It's probably still all totally wrong.  

 

 

 

Edited May 29, 2017 by Kaunitz

[Kaunitz]

double post, sorry

Edited May 29, 2017 by Kaunitz

[Kaunitz]

A quick tree-experiment. It seems as if LOS in trees can be explained by the model quite well. One could regard trees (foliage-tiles) in the same way as ground-tiles, just with a different LOS-blocking value and a differently shaped (taller) hitbox. Note the strange irregular pattern (half-LOS-squares). I had noticed a similar effect when I had my unit look through ground-tiles before. That diagonal LOS-extension seems to be consistent.

PS: Density 2 and 3 lead to the same results for tree A. But I assume that LOS still deteriorates faster in density 3. So, you're more likely to spot someone on a LOS-square in density 2 than in density 3.

The one thing I would really be interested in is whether the LOS-blocking effect of terrain increases with distance AND/OR whether the square in which a unit is positioned is ignored for calculating the LOS of that very unit. Increasing the LOS-blocking effect of terrain with distance would result in cover-advantages. If you're close to a terrain piece, you can see through - if you're far away, you can't. It's a keyhole-effect.

Certainly the game does either of these two - otherwise cover wouldn't really work as both units in a spotting duel would always have the same chance to spot each other (even if one is in a wood and the other in the open). The game engine certainly does produce cover-advantages  - very clearly e.g. when wall-tiles are involved (e.g. a unit behind a hedge certainly ignores the LOS-blocking-effect of that hedge). It would be interesting to know if ordinary terrain (ground and foliage tiles) can also deliver a spotting-advantage or whether foliage (unlike walls) always works in both ways. I need to run some tests.

If we assume that the game ignores the terrain in which a unit is placed for calculating that unit's LOS, then this is how you would gain the greatest cover-advantage for unit A against unit B:

• maximum cover in the square in which the unit A is placed (ground-tile with high LOS-blocking value - e.g. heavy forest; foliage tile with high LOS-blocking value - e.g. density 3 "low" trees; wall-tile - e.g. hedge). This cover is ignored by unit A but not by unit B.
• Any terrain in between unit A and B works in both directions, i.e. it reduces the LOS-quality of both units, reducing their spotting chance. It can be used to fine-tune risk/security for unit A. Note that the LOS needs to cut through the hitbox of the terrain (i.e. it needs to cut across a certain height of the ground) for terrain like high grass to take effect. Note that the stance of the unit matters for this purpose (prone/kneeling/standing)! E.g. you could design a map so that unit A has excellent concealment against unit B's position when it goes prone.
• Unit B should be in the open.

 

Edited May 30, 2017 by Kaunitz

[Bulletpoint]

1 hour ago, Kaunitz said:

 

 

Trees are placed randomly within squares, so depending on where the actual trunks go, that affects LOS. Sometimes there will be "alleys" where there are no trees, leading to LOF corridors in woods.

Also, are you sure that the position of that single soldier in the square doesn't matter? I think it does.

[Kaunitz]

I found that for LOS-purposes, the exact location of trees within the square does not matter at all. This is also what the official post (linked above) clearly suggests. For LOS, a square that has three trees on it is a square of tree XY/density 3. You can drag your LOS through treetrunks without any problems (until you reach the end of the (half-)square...). When it comes to bullets, however, it's an entirely different matter, as individual tree-trunks block bullets. 

But until now, I admit that I also had the impression that individual treetrunks affected my LOS (target-command). But now that I gave it a detailed look, this might have been a wrong impression. I was clearly able to drag my LOS through treetrunks (as long as I had LOS in that square!), and the LOS was blocked or clear based on the layout of squares (see picture above), not of the individual treetrunks. You can place a single soldier directly with his face pressed against a tree - he can still see through (but he might not be able to fire).

For some reason my screenshot cannot not capture my mouse-cursor and the LOS-line. I've added it (red line). My LOS cuts through the tree, but I had a clear LOS to the end point. This also works with distant trees. LOS (unlike LOF) is not blocked by individual treetrunks. As a sidenote: I was a bit surprised that this team was able to see though the low bocage wall-tile (the dark green bush). A quick test showed me that bocage acts as a soft blocker. It seems to drastically decrease LOS-quality, but it doesn't automatically block LOS. If you're close enough (so that your LOS quality is still high enough to "penetrate" the bocage-square, like it is the case here), you can see through - but your LOS-quality will be bad "behind" the bocage. In fact, a prone soldier placed 1 square away from the first hedge has enough LOS-power to see through 2 hedges and 4 squares of ploughed field!

As for the position of the spotter within the square, It is interesting to note that when you draw a LOS, it does not start at the center of the square of the unit. Rather, it starts in relation to the individual soldiers' positions within that square. For example, in a unit of two men, the LOS always starts at the center of an imaginative line between the two soldiers. It's also interesting to note that sometimes a grey LOS-line comes up, suggesting that at least one weapon in the unit cannot draw a LOS (LOF to be correct!) to a particular point in the target square. So here the game clearly takes into account position of a weapon and the exact spot targeted in the square. It's hard to tell where the game draws the line between LOF (detailed resolution, using individual points within a square; "cover") and LOS (based on squares; "concealment/spotting"). 

 

NOTE TO SELF: It seems as if fllavour objects don't block LOS, but they do stop bullets. So it should be possible to use unsuspicious flavour objects (e.g. rocks) intentionally to improve the cover of a fighting position (better for my nerves than fiddling around with the terrain mesh and elevation-edges...).

Edited May 30, 2017 by Kaunitz

[Kaunitz]

Okay after some more testing I think that you are correct that trees do work in a more detailed way. I can drag LOSes of very variable power through a wood (how many squares do I still get clear vision after the LOS cuts through 1 square of a density 3 wood). So indeed it must be based on a more detailed model, not on a broad-brush "per square" model. Also, there are differences between individual tree-types.

It can't be the treetrunks themselves though. The situation posted above makes this pretty clear. Also, I sometimes had a more powerfull LOS when the LOS was cutting right through a trunk! So my guess is that it might rather have something to do with the shape of the trees' foliage/soft blocker-hitboxes?

 

I also run an experiment in order to check whether units ignore the terrain in their own square for LOS-calculations. I put unit A on a heavy-concealment-square (heavy forest, 3 trees, hedge) and another unit B on an open square and checked how far they could see across the same kind of terrain. Both could see the same distance, which suggests that unit A's LOS was not degraded by the heavy cover in it's starting square. But I need to run the rest many more times to make any real conclusion. 

Edited May 30, 2017 by Kaunitz

[Bulletpoint]

2 hours ago, Kaunitz said:

It can't be the treetrunks themselves though. [...] You can place a single soldier directly with his face pressed against a tree - he can still see through (but he might not be able to fire).

 

 

I think you might be confused because trees in the same square as the spotter don't block LOS/LOF. But once you get a bit farther away, they do block LOS/LOF.

That's also the reason why you can hide your tank behind a tree and the tank can fire out, but incoming rounds will very often be blocked by the tree.

Edited May 30, 2017 by Bulletpoint

[Kaunitz]

I'm quite certain that tree trunks do not affect LOS (LOS =/= LOF). Distance doesn't matter. It must be the foliage (must be some kind of hitbox around a tree trunk) that degrades LOS (quite heavily, depending on tree type). 

A little video: 

The soldier is positioned on the grass-square. He has decided to be at the corner where the neightbouring square's trunk offers some protection. As you can see, seeing through tree-trunks is not a problem, even 1 or 2 squares away. After the second square though, LOS is blocked. Not by a tree-trunk, but what seems to be the border between the two actions squares (this is how I came up with the diagrams above). This is also why I thought trees might just act like ground-tiles, on a square-basis. I assumed that LOS had been degraded by cutting through two squares of density 1 trees so much that it could not enter the next square of trees. 

But upon closer inspection, and also with other types of trees, the results are very variable, so that the individual position of trees (with their foliage-hitboxes) must play a role here. Maybe the hitbox of this particular type of tree is so large that it produes more consistent LOS-results. Other trees' hitboxes might be narrower, so that there is a chance that a LOS comes at an angle that does not cut through the hitbox and is not degraded/blocked. if this was a tree of a different, taller type, foliage-hotboxes must look differently or have a smaller LOS-blocking effect as units can see much farther through them/their foliage-hitboxes).

Here is a second video that shows a different tree type (same density as above) and a different result. Unlike before, here the position of the tree(trunks) seems to matter. My suspicion is that - unlike the former tree- this tree's foliage-hitboxes are narrower, so that a LOS coming at a certain angle can evade them (so the LOS keeps its power and can go on). If you hit the foliage-hitboxes, your LOS is degraded. Note that when my LOS cuts through trunks/narrow foliage-hitboxes (in this case, I cut through two) I can still target the point immediately behind the trunk in the same square, but the LOS is degraded so that it cannot enter the next square.  

 

 

Edited May 30, 2017 by Kaunitz

[Bulletpoint]

2 hours ago, Kaunitz said:

As you can see, seeing through tree-trunks is not a problem, even 1 or 2 squares away. After the second square though, LOS is blocked. Not by a tree-trunk, but what seems to be the border between the two actions squares

Thanks for the videos. I think what is happening here might be that while your visual targeting line is drawn from the individual soldier, the game calculates LOS from centre to centre. If there's a tree intersecting the line drawn from the centres of two squares, you will not be able to gain LOS between those two squares, unless 1) you get into the square with the blocking tree and there's no other trees blocking the line, or 2) you have an enemy contact marker at the target square, in which case LOS will begin to be calculated directly between your troops and the enemies there, based on their actual locations within the square.

I am not 100 pct sure but I think this is how it works. And I think this is also the reason why LOS in orchards (where all trees are in the dead centre of their squares) seems so wonky.

Edited May 30, 2017 by Bulletpoint

[Vanir Ausf B]

LOS calculation is a two-step process. The first is a simple LOS check from center of the action spot to the center of the other action spot. If this LOS check fails no unit in either action spot may spot each other and no further action is taken. If the center to center check passes then each individual soldier's LOS in that action spot can be checked against any point in the other action spot. In the former instance individual positioning within the action spot makes no difference, in the latter case it does.

Edited May 30, 2017 by Vanir Ausf B

[Kaunitz]

@Vanir Ausf B Thank you for the info! It certainly makes sense. From what I can say by looking at the game, spotting seems to be highly sophisticated. So I wondered how the game could ever handle all the calcualtions behind it. 

--------------------

LOS conclusion/theory

I think it’s time to draw a line and summarize my risky theory. Once again, I have to point out that all of this the result of me trying to understand what’s happening in the game. None of it is official/insider-information or confirmed in any way. The following is based on my observations in the game and these official passages by battlefront.com:

The way it works is the LOS line is "degraded" as it is drawn from point to point.  The more cumulative crap in the way the less strong the line becomes.  The quality of the spotter, the less restrictions on view, etc. give the line a higher starting value than a LOS line drawn from a unit with negative factors.  Some of the factors are specific (restrictions on range of view or height for example), others are general (optics of X type vs. eyeballs is the best example).  Each piece of terrain has ratings which determine how much the line is degraded when it comes to that piece.  At some point the line is so degraded that it is considered "blocked". 

The 3D graphics of the foliage plays absolutely no role in LOS drawing since that is simply impossible to do because the computers are totally not up for it.  This is the primary reason for potential disconnect between the visual representations in the game and the ability to spot.

LOF, which is not the same as LOS, is a little different.  LOF does pay attention to the pixels on the screen at least to some extent.  Foliage is one of the things that is a "no" except for the trunks.  This is possible to do because shots are fired in low enough quantities that there is enough CPU power to allow for more calculations.  Plus, LOF is a binary decision about if something should keep going or stop, not the sort of conditional quality stuff going on with LOS. 

[http://community.battlefront.com/topic/121832-something-very-wrong-with-los-through-trees/?do=findComment&comment=1652113]

 

So, here comes my attempt at understanding LOS (the one that is shown to you when you issue a target-command). LOS seems to be drawn from point to point, not from action-spot to action-spot (8x8m). I don’t know the exact “resolution” of the grid used for this (1m intervals?). 

When using the target-command (during turn-resolution, the game uses a routine described by Vanir Auf B ), the starting point of a LOS is calculated from the positions of your unit’s soldiers within the square. If there is just one soldier in your unit, then the LOS starts at his head. If there are two soldiers in the square, then LOS starts in the center in between them, etc. I don’t know how the starting height of a LOS is determined (kneeling/standing/prone) if there are more than one soldiers in the unit.

For LOS-purposes, terrain features and objects/houses (except for flavor objects?) have a hitbox and a LOS-blocking value. Ordinary ground-types (like dirt, grass, long grass, etc.) are also treated as a terrain feature for this purpose and come with a hitbox and a blocking-value  respectively. Whereas objects such as trees and walls have comparatively small, individually shaped hitboxes, ground-hitboxes fill the whole action-spot (8x8m). Ground-hitboxes still come with varying height though - for example, the hitbox of tall grass seems to have a greater height than that of ordinary grass, and the forest hitbox seems to be even taller.

At any point, a LOS has a certain strength or quality. The starting strength of a LOS depends on the quality and equipment (binoculars) of the spotting unit and probably also on light and weather condition of the scenario.

Whenever the LOS of a unit (drawn from it’s starting point as described above) touches/cuts through the hitbox of a terrain-feature, the LOS’ current strength is compared to the terrain-feature’s blocking value. If the strength is high enough, the terrain-feature’s blocking value is substracted from the LOS’ strength and the LOS is allowed to pass through the terrain feature. If the LOS’ strength is not sufficient to penetrate the terrain feature, however, the terrain feature’s hitbox acts like a wall. Depending on the angle between the spotter and the hitbox, this might create a “reverse slope”-LOS. 

As ground-hitboxes (grass, dirt, mud, etc.) are omni-present on the battle-field, it is often a ground-hitbox that finally stops a LOS. This might be the reason why one gets the impression that LOS is working based on (8x8m) action-spots rather than points?

The stronger a LOS is on a certain point, the more likely the chance to spot enemies there (but see Vanir Ausf B's comment on the game's spotting routine). In the game, you can’t really see how strong your LOS is on a certain point, you only get to know at what point it reaches zero (no LOS).

Having a LOS on a point does not automatically mean you can effectively fire on it. LOS-calculations use LOS-hitboxes – for bullet-mechanics, the game uses a different system. E.g. tree trunks block bullets, but not neccessarily LOS - it depends on whether the (invisible) LOS-blocking ("foliage")-hitbox of the tree is exactly where the trunk is and how strong its blocking value is. Often, one tree is not enough to block a LOS completely - it takes 2 or more trees to reduce LOS to zero (this is why people get the impression they can see through nearby trees).

What I keep asking myself is if and how this system generates spotting/LOS-advantages. I.e. how come two equal units can have a different chances to spot each other. Does LOS-blocking value of terrain increase with distance? Is the terrain of the action-spot in which the spotter is positioned ignored for calculating his LOS? 

-------------------------------

 @Bulletpoint My attempt to explain the LOS-behaviour in the second video:

I do believe that the target-command-LOS is drawn/calculated from the point that the game actually shows you. I think that the video makes this quite clear because touching trees has an effect on my LOS, and the line for determining whether my LOS touches a tree or not is not drawn from the center of the action-spot, but in relation to the soldiers in the spot.

There are two odd things in the second video:

1) my unit can see through 1 tree-trunk – a LOS that touches one tree trunk is not stopped

2) a LOS that touches 2 tree-trunks (and two action-spots of dirt?) however, is stopped - but not immediately at the second trunk, but rather at the edge of the next action-spot (!). [this is a very consistent observation - when your LOS touches too many trees (depending on their LOS-blocking values), it will very often be stopped at the edge of the next action-spot, not directly at the tree itself]

This could be explained by the theory pretty well. The LOS has a starting strength of - let’s say - 10. The first hitbox it could touch is the ground-hitbox of the neighbouring square. However, since the ground-type is dirt, I don't think there actually is a hitbox here (or at least it's effect would be trifling). So, the LOS keeps its value of 10. Next, it touches the hitbox of a tree (this hitbox's position might or might not be identical with the tree-trunk that you see – different trees come with different hitboxes). This heavy foliage-obstacle reduces LOS by 5, so it is now 5. Next, the LOS crosses over to the next action-spot - again dirt, so no effect. Then comes the second tree which reduces LOS to 0. Therefore, LOS cannot penetrate the next obstacle on its way, which is a ground-hitbox (grass). Unlike dirt, grass has an actual hitbox and a blocking value. And since LOS power is already 0, LOS is blocked at the grass-square. What I do not understand, however, is why LOS is immediatly blocked. My theory would suggest a reverse-slope LOS (prone spotter versus grass-hitbox at 3 spots distance) from this point on. 

By comparison: if the LOS only touches one tree on its way, it has enough power to penetrate the grass-spot. 

Also, LOS was able to go through 2 trees and carry on over dirt-ground - i.e. when it was not stopped by a grass-hitbox.

-------------------------------

HE impacts & trenches

Now that I've tried to understand concealment, my other point of interest for creating my scenario (a mechanized platton attacking a dug-in squad) is artillery, or more generally: HE-rounds (including horizontal HE such as tank-shells or RPGs). The lack of any kind of overhead protection seems to make off-board artillery a no-go for CM:BS as airburst-shelling seems to obliterate infantry in foxholes and trenches a bit too quickly for my taste. And even ordinary ammo is quite effective against troops placed in trench-units. This made me wonder how HE rounds are handled by the game. I assume that they do indeed work on a physics-basis. So, does an explosion/shrapnel take into account the angle of the ground at the impact-spot? Are individual shrapnel pieces modeled? etc. 

If it was physics based, one could alter arty-effectiveness by modifying the ground-mesh (dig real trenches, so to speak, not just placing trench-units). I have tried out some designs and I do think my battle-positions do indeed have a positive effect on HE-protection. But testing is a very delicate matter here, and the offmap-arty-panic-issue makes it even harder. A small sketch about my thought-process for the trench-design:

The best battle-position would be as narrow as possible (to reduce target-area for direct vertical hits and airburst-shrapnel) with sharp edges (against vertical hits) and have a flat silhouette (to counter horizontal HE fire). 

 

Edited May 31, 2017 by Kaunitz

[Bulletpoint]

All your tests are done very close to the trees. Does anything change if you try to place the spotter about 300m away from the trunks?

 

[Bulletpoint]

55 minutes ago, Kaunitz said:

This made me wonder how HE rounds are handled by the game. I assume that they do indeed work on a physics-basis. So, does an explosion/shrapnel take into account the angle of the ground at the impact-spot? Are individual shrapnel pieces modeled?

I don't think so. I think HE is basically a case of the game drawing an invisible line from point of impact to each soldier within the potential kill range. If the line is blocked by a wall, ground, etc. then no further action is taken. If the line is clear, then there will be a "dice roll" made based on the size of the explosion and distance to target, plus cover type. In CMBS probably there's some modifier for body armour too - no such luck for my WW2 grunts.

But even though the game doesn't model individual fragments etc. the end result is pretty much like your drawings.

Edited May 31, 2017 by Bulletpoint

[Kaunitz]

1 hour ago, Bulletpoint said:

All your tests are done very close to the trees. Does anything change if you try to place the spotter about 300m away from the trunks?

 

I've tried it out with greater distances (but still in clear terrain - i.e. no terrain degrades my LOS before it hits the trees). It doesn't change a thing. Even if there is a greater distance between the spotter and the trees, the spotter can consistently see through 1 tree. But if his LOS cuts through two trees, it will be stopped at the next obstacle (which - for a prone spotter - is grass in most cases). 

However, there is a bit of inconsistency, regardless of distance. Sometimes, I should be cutting only through one tree (i.e. my LOS into the next action-spot should be clear) but the result suggests that I'm cutting through two trees (my LOS is blocked at the next obstacle). This seems to happen sometimes when there is another tree close by so that my LOS runs "in between" the two trees and touches only the trunk of one tree. It happens more often with some types of trees than with others. My suspicion is that it might be pointing to different hitboxes for trees and that their LOS-hitboxes might be randomly aligned around the visible trunk. But I admit that this is a very bold hypothesis:

 

1 hour ago, Bulletpoint said:

I don't think so. I think HE is basically a case of the game drawing an invisible line from point of impact to each soldier within the potential kill range. If the line is blocked by a wall, ground, etc. then no further action is taken. If the line is clear, then there will be a "dice roll" made based on the size of the explosion and distance to target. In CMBS probably there's some modifier for body armour too - no such luck for my WW2 grunts.

Well that's pretty much what I had in mind when I said "physics"-based.  So this would suggest that you can intentionally design positions to offer protection against HE rounds. For example, placing sandbag walls can be quite handy for blocking shrapnel from vertical HE/artillery, especially in broad trenches (elevation works by aaction spot 8x8 meters - which results in very broad trenches). PLacing sandbag walls would seperate the trech into several compartments/reduce the risk (at least for kneeling or prone soldiers).

Edited May 31, 2017 by Kaunitz

[Bulletpoint]

33 minutes ago, Kaunitz said:

However, there is a bit of inconsistency, regardless of distance. Sometimes, I should be cutting only through one tree (i.e. my LOS into the next action-spot should be clear) but the result suggests that I'm cutting through two trees (my LOS is blocked at the next obstacle). This seems to happen sometimes when there is another tree close by so that my LOS runs "in between" the two trees and touches only the trunk of one tree. It happens more often with some types of trees than with others. My suspicion is that it might be pointing to different hitboxes for trees and that their LOS-hitboxes might be randomly aligned around the visible trunk. But I admit that this is a very bold hypothesis:

This is starting to feel a bit like quantum physics  Maybe LOS in this game works a bit like a particle and a wave at the same time.

Seriously though, I suspect we need to see this as based on squares, not based on "tree hit boxes".

You can fire out of your own tree-square without penalty. You can fire through one (1) tree-square, and you can then fire into, but not through a third tree-square. So that's the maximum LOS penetration you can have in the game.

I'm just pretty sure I've seen units engage each other through a lot more tree squares than that though.

What's your take on this old thread?

 

[Kaunitz]

Bulletpoint, I'm only looking at LOS (the target-command-line) not at fire combat. So, it's just about spotting, not firing.

I'm quite sure that LOS is not based on 8x8m squares, but on a more detailed "resolution" (and some form of hitboxes). If it was all based on squares, then how could the game create those "reverse-slope"-effects (LOS-shadows/dead ground) behind walls? Also, the game needs a very sophisticated model in order to handle minute undulations in the ground-mesh and soldiers' stances as excellently as it is handling them right now.

Also, the videos and my observations clearly suggest that LOS through trees is not based on squares but that it does indeed take the position of individual trees into account. There are only that many variables once we know that the game considers individual trees: 1. the actual tree trunk (we know that it affects bullets, but what about LOS?), 2. an (invisible) LOS-hitbox (we know that the game represents foliage somehow; the exact shape is unknown, 3. the blocking value of the LOS-hitbox (we have official information that LOS is "degraded" when it cuts through terrain features).

The link you posted includes an official bit of information by battlefront.com (quoted above; page 2 in the thread). This information and my own observations were the basis for my "theory" ( ) on. 

Edited May 31, 2017 by Kaunitz

[Bulletpoint]

36 minutes ago, Kaunitz said:

Bulletpoint, I'm only looking at LOS (the target-command-line) not at fire combat. So, it's just about spotting, not firing.

But to fire, you need to spot. At least when talking about opportunity fire. In that thread, the guy says his tank was knocked out through a lot of forest. I've seen similar things happen. I always assumed the reason was that each tree halved the chance of a spot. So spotting through ten trees would be 1/2^10 chance (a bit less than one in a thousand chance). This is a low chance, but you can multiply by ten units on the other side of the forest, and multiply again by ten games played, and then there's a one in tenth chance you'd see it happen.

36 minutes ago, Kaunitz said:

If it was all based on squares, then how could the game create those "reverse-slope"-effects (LOS-shadows/dead ground) behind walls?

Because the game does those centre-to-centre checks, but not only ground-to-ground, also at various height levels. Probably it has 5 different height levels it checks for (prone, kneeling, standing, plus 2 vehicle heights). So the blue targeting line tells you there's LOS from ground-to-ground, and then reverse slope means that there's no ground-to-ground LOS but at least one of the other potential target heights is visible.

Edited May 31, 2017 by Bulletpoint

[Kaunitz]

This might help: 

I took a look at a group of trees (density 1) and noted down where exactly the lines of sight of my spotter got blocked. The red dots are tree-trunks, the grid equals action-spots, ground type was grass, spotter prone. 

Observations:

• LOS always gets stopped at the border between-action-spots. I'm not sure if that means that the whole LOS-system is based on spots. On one hand, I've never seen a LOS blocked directly at/by a tree-obstacle (sometimes, when the tree is placed very close to the border, it looks as if this is the case), but only at the next action-spot-border that the LOS tried to cross. On the other hand, obstacles such as buildings and walls can instantly block LOS, within an individual action-spot.
• Obviously the footprint of trees for LOS purposes  (= "foliage") is larger than their footprint for LOF purposes (= the trunk). You can clearly see that LOS lines are shorter even when they're close to a tree - they do not need to touch the trunk. Whereever a LOS is stopped early on, it should be safe to assume that it has touched the LOS-footprint of at least two trees. (Not depicted here, but when a LOS only touches a single, isolated tree, it is not stopped).
• The exception to this rule are the cases marked with blue dots in this image: 
• 

         Note that in all these three cases, one would expect the LOS to be blocked sooner because LOS certainly goes through two tree-LOS-hitboxes. The reason why LOS can reach out one spot further seems to be that they cross into the next spot almost exactly at the intersection of 4 spots (blue dot). 

• The position of individual trees does matter. If it didn't, we should not be able to get the clear LOS in the center.

-------

With short trees (treetop closer to ground) things seem to get more messy and random (orchard-phenomenon?).

Edited May 31, 2017 by Kaunitz

[Bulletpoint]

43 minutes ago, Kaunitz said:

I've never seen a LOS blocked directly at/by a tree-obstacle (sometimes, when the tree is placed very close to the border, it looks as if this is the case)

Yes I feel as if I've seen this many times too, but I guess I was tricked by trees close to the edge. As I'm not playing with a gridded terrain mod, I am not always sure where individual squares begin and end.

45 minutes ago, Kaunitz said:

Obviously the footprint of trees for LOS purposes  (= "foliage") is larger than their footprint for LOF purposes (= the trunk).

I'm pretty sure the foliage also blocks LOF. At least sometimes. Artillery and mortars will often burst against the leaves of a tree. At least it looks like that to me, but maybe it's actually the shell hitting the trunk's hitbox?

I think you're on to something, but not sure we're quite at the definitive answer yet...