It's got to go.

[Baneman]

... Why that notional initial aim point can't be the x,y,z coordinates of the terminus of the target line when the mouse is clicked, regardless of the LOS status of the centre of the AS is unknown to us mere mortals...

[wild speculation]

Possibly something to do with the whole LOS-map that is created - if the engine pre-calculates LoS from AS to AS, it would be difficult to differentiate between, for eg. Hills blocking LoS ( and LoF ) and the wheat/wall/trees that could possibly be fired through or at.

ie. Your suggestion could end up with the AI shooting ( or thinking it can ) through hills.

[/wild speculation]

I do of course, agree wholeheartedly with the hope that the OP problem can be addressed, but it seems clear that it's rife with pitfalls.

[Ultradave]

I always just assumed you could not area target or call in arty through trees on the side of a slope, hill, mountain, what have you, because there is no LOS to the ground. But I could understand why someone would come away incredulous that this is so. Its seems so simple to be able to do that in real life. I chalk it up to engine limitations. But, yes, whatever the case may be, not being able to get LOS through trees to call in artillery or CAS to the side of a hill definitely has GOT TO GO.

Yes, being an ex-artilleryman, it does frustrate me that I cannot call artillery on a map coordinate that I don't have a LOS to with an observer (if I could I'd be happy with possibly some degradation of accuracy because this is WW2 not present day, so no adjustment is possible). If I have a map - I can hit it with artillery, without having a TRP. The TRP just makes it quicker because the battery will have already done the calculations.

[Lt Bull]

+1 to this thread.

I have just played through an otherwise awesome scenario/map (Counterattack at Son) where the 'untargetable' issues discussed here really end up being unavoidable and to a degree immersion breaking. Much of the terrain is flat with long grass/wheatfiled and infantry prone in that terrain are virtually invulnerable to area fire.

[womble]

Yes, being an ex-artilleryman, it does frustrate me that I cannot call artillery on a map coordinate that I don't have a LOS to with an observer (if I could I'd be happy with possibly some degradation of accuracy because this is WW2 not present day, so no adjustment is possible). If I have a map - I can hit it with artillery, without having a TRP. The TRP just makes it quicker because the battery will have already done the calculations.

AIUI, you're being a little bit optimistic about how much help a map would be, in WW2. They didn't have very good maps, by all accounts, and even having good sightings on fair landmarks, the battery's estimate of their own position (and therefore the first, uncorrected fall of shot) could be out by half the width of a CM battlefield, easy, and the spotter's less-well-surveyed (especially if they're a Green platoon CO, sheltering from MG fire) could be even less accurate. Walking the shells onto target was a necessity of the era.

[PhilM]

[wild speculation]

Possibly something to do with the whole LOS-map that is created - if the engine pre-calculates LoS from AS to AS, it would be difficult to differentiate between, for eg. Hills blocking LoS ( and LoF ) and the wheat/wall/trees that could possibly be fired through or at.

ie. Your suggestion could end up with the AI shooting ( or thinking it can ) through hills.

[/wild speculation]

I do of course, agree wholeheartedly with the hope that the OP problem can be addressed, but it seems clear that it's rife with pitfalls.

Most logical explanation so far - go to the top of the class! (For now ...)

I'm relieved that a few other people seem to be as sad as I am, and also find this interesting (as well as frustrating ...)

[Ultradave]

AIUI, you're being a little bit optimistic about how much help a map would be, in WW2. They didn't have very good maps, by all accounts, and even having good sightings on fair landmarks, the battery's estimate of their own position (and therefore the first, uncorrected fall of shot) could be out by half the width of a CM battlefield, easy, and the spotter's less-well-surveyed (especially if they're a Green platoon CO, sheltering from MG fire) could be even less accurate. Walking the shells onto target was a necessity of the era.

OK, perhaps I oversimplify. But my artillery experience dates to the days of charts and darts and firing sticks however, well before GPS, computer fire controls and all those cool things that are taken for granted now :-). I'm really not THAT far removed in my experience and techniques from those in WW2.

But even without good maps (which was a fairly common experience for me too), presuming I can figure out where I am to some reasonable level of accuracy (enough that I know I'm not calling fire down on my head), if I can see the edge of the woods and presumably somehow call fire on it, I can call fire 400m into the woods just as well. Granted, I can't see to adjust and walk it on to the target, but maybe all I'm trying to do is keep everyone's head down so I can move (I suspect they are in there somewhere and want to suppress them so I can move)

[LukeFF]

Just wanted to chime in here and agree with the OP. The amount of times this issue has popped up in my current KG Engel campaign mission is really getting to be frustrating.

[womble]

OK, perhaps I oversimplify. But my artillery experience dates to the days of charts and darts and firing sticks however, well before GPS, computer fire controls and all those cool things that are taken for granted now :-). I'm really not THAT far removed in my experience and techniques from those in WW2.

Understood, but it's not the techniques that are the root of the imprecision I'm referring to, but, AIUI, the maps themselves.

But even without good maps (which was a fairly common experience for me too), presuming I can figure out where I am to some reasonable level of accuracy (enough that I know I'm not calling fire down on my head), if I can see the edge of the woods and presumably somehow call fire on it, I can call fire 400m into the woods just as well. Granted, I can't see to adjust and walk it on to the target, but maybe all I'm trying to do is keep everyone's head down so I can move (I suspect they are in there somewhere and want to suppress them so I can move)

If you've already dialed your arty support in to a point you can see (and thus removed all the problems of potentially mistaken appreciation of the spatial relationship between spotter and battery (and target area, therefore)), sure you can call it 400m into the woods. Though I'd have to ask what you think might be hassling you from there... Unfortunately, CM doesn't remember previous missions... BFC seem quite determined that there will never be a "repeat mission" fire request, and have stated their reasons, which seem, on the face of it, to be reasonable, though arguable. I apologise for not being able to recall their resoning in detail. What convinced me, though, IIRC, was the "balance" argument. If you could call arty "blind" (even if you had to have zeroed one mission with the called-upon battery first), the God's Eye perspective of the player would make blind missions pretty much routine.

[kensal]

OK, perhaps I oversimplify. But my artillery experience dates to the days of charts and darts and firing sticks however, well before GPS, computer fire controls and all those cool things that are taken for granted now :-). I'm really not THAT far removed in my experience and techniques from those in WW2.

But even without good maps (which was a fairly common experience for me too), presuming I can figure out where I am to some reasonable level of accuracy (enough that I know I'm not calling fire down on my head), if I can see the edge of the woods and presumably somehow call fire on it, I can call fire 400m into the woods just as well. Granted, I can't see to adjust and walk it on to the target, but maybe all I'm trying to do is keep everyone's head down so I can move (I suspect they are in there somewhere and want to suppress them so I can move)

I think the reasoning behind not being able to lay artillery by map co-ordinates in the game is based on the principle that TRPs represent artillery targets that have been zeroed in by map co-ordinates, whereas if you don't have that already sorted before the beginning of a scenario you would not have enough time to do that within the timeframe.

[Ultradave]

But that's no different than firing at something you can clearly see - a normal on call fire mission. TRPs only save you the time to compute the data and send it to the guns. They aren't zeroed in and there's a good chance you might have to adjust from a TRP - in fact that's a lot more likely than calling a FFE on a TRP (the bad guys rarely oblige you by walking over your TRP location :-) )

(nothing here applies to firing at unspotted locations - just replying to the last comment)

[Battlefront.com]

I know this thread has been quiet for over a month, but today I stumbled upon it and figured I would address the issues here since, no doubt, they will come up again. In exchange I'm trusting you guys to help answer the next dozen threads that pop up

Three issues have been raised here. Each is somewhat different. However, it is important to understand this one critical component of CM:

LOS is, by far and away, the most taxing element on the game's use of RAM and CPU time. In order to allow CM to run and run well on an average computer there must be an underlying grid to avoid RAM and CPU limitations. The only way a grid works is to have ONE point represent each square (Action Spot) of that grid for purposes of LOS. Therefore, when a unit can not see the center of an Action Spot it is presumed LOS doesn't exist and the system does not attempt to proactively refine LOS. Even with that the system is stressed out pretty hard to deal with CM sized maps and average unit loads. Your computer quite simply can't handle the alternative.

Line of Fire (LOF) and Spotting are different matters, but let's put aside Spotting for now. When LOF (Line of Fire) is drawn it is between two specific points, not between two Action Spot centers. A unit either occupies a specific spot (soldier) or a range of spots (vehicles, which includes big guns). Each spot has a height associated with it depending on stance (soldiers) or the portion of the unit in question (vehicles). To keep the CPU and RAM from being overloaded there are a sent number of predefined heights, which we refer to as ELOS (Enhanced LOS).

When a shot is fired it travels a literal ballistics path between two points. If it collides with something the physics of that collision are dealt with and the effects applied. These on-the-fly calculations are "expensive" from a resource standpoint. However, CM can do this because there is only so much shooting that will go on at any one time. Well, most of the time for most people in most situations. Those of you with more marginal computers playing on huge open maps with tons of units WILL notice slowdowns as a result of LOF calculations. Which proves why Action Spots are necessary.

Spotting basically works the same way once a unit spots another unit. At that point tracking is point to point with the Action Spot system offering ways to shortcut otherwise massively "expensive" calculations.

OK, so now how does all this apply to the basic questions?

1. Why can't you Area Target a point on the map if you can't see the middle of it's Action Spot, yet if there is a unit there you can target just fine?

A unit can only trace LOS to centers of Action Spots or other units. That's because in each case CM has two distinct points to draw a line to. The first is specific friendly point to generic fixed point (Action Spot center), the second is specific friendly point to specific enemy point. What CM can't do is trace specific to random because that hits the computer too hard in context with all the other demands placed on it

For LOF to be traced the unit must first have LOS to the point in question. If there is no LOS then there can be no LOF. You can, however, have LOS without LOF, though generally the two go along together.

The problem arises when the player wishes a unit to fire upon a specific spot within an Action Spot it can not draw LOS to because it can't draw to it's center point. Unfortunately, as with any system that has a cutoff point, there are situations where the center is juuuuuuuuust out of sight making the entire Action Spot off limits. "Close enough" isn't something the computer can understand and therefore you either have it or you don't. Period.

Now, if a spotted enemy unit happens to be in the same exact spot, well... that's different! CM understands that the two units already know where each other is and therefore the Action Spots are, for the sake of argument, ignored. Now it is point to point. As stated above, the average computer can handle a reasonable amount of these situations at once so it's allowed.

2. Why can't you Area Target a particular point of an Action Spot at various heights, not just the default 1m height (which IIRC is the default)?

It is a similar answer to the above. You can target a unit, with it's specific height considered, in a way that makes sense because it is a point to point LOF calculation. But to Area Target a specific point of an Action Spot isn't straight forward. Sometimes you want it to be the ground, other times you want it to be 1m in the air, other times you might want it to be even higher. How is CM supposed to understand what it is you want to do? It can't. And having some sort of clunky UI to manually change the aim height is something we won't do.

BTW, it is wrong to say there is no grazing fire. Dead wrong. It's in the game and it works perfectly well. What you can't do, at least easily, is ensure that a specific spot of ground is covered by a specific height of fire.

3. Why can't you use artillery to target areas you can't see?

There is no game mechanics reason this can't happen. It's all specific to gameplay. You, as player, are blessed with waaaaaaaay too much information and therefore waaaaaaaaaay too many ways to abuse artillery. You could, for example, strike the enemy's likely assembly areas without anything other than a time penalty. Realistic? NO. In real life you would need intel and opportunity to use artillery blindly like that, but in a CM game you go into the game with both in your head. You know, for example, that there's an enemy out there. You know if the enemy is in the process of attacking you or is going to be setup for defense. Because of the limited size of the maps you know pretty much where he'll be. You also know pretty much what he has for forces too. Etc., etc., etc.

As realistic as grid firing is as a capability in real warfare, ironically in the game it would trash the overall realism to the point of nobody wanting to play any more. So we don't allow it. It's as simple as that.

Hope that helps!

Steve

[womble]

I know this thread has been quiet for over a month, but today I stumbled upon it and figured I would address the issues here since, no doubt, they will come up again. In exchange I'm trusting you guys to help answer the next dozen threads that pop up

Thanks for popping in, and I'm sure we'll do our best. I'd better bookmark this post for reference!

LOS is, by far and away, the most taxing element on the game's use of RAM and CPU time. In order to allow CM to run and run well on an average computer there must be an underlying grid to avoid RAM and CPU limitations.

Understood.

The only way a grid works is to have ONE point represent each square (Action Spot) of that grid for purposes of LOS.

That's certainly true while the grid is the only thing you can use. But the grid isn't the only thing you use now: units are spotted without the assistance of AS grid centres (at the final assessment).

Even with that the system is stressed out pretty hard to deal with CM sized maps and average unit loads. Your computer quite simply can't handle the alternative.

Ah. While forming my response, I've stumbled across a realisation that this is always going to be true in flowing RT play. When the game is in pause or during the order phase, the considerations of map size and unit count can potentially be given significantly lower priority while the targeting tool is being used, since pretty much the sole focus of the player is the use of that tool. Background functions (unit animations, sound, ambient terrain animations and the like) could be halted entirely, if necessary to free up the Hz for the LOS calculation. However, since the tool has to work in the dynamic environment of the RT game unpaused, such parking isn't practical and the "suspended" modes of play have to work within that limitation.

Line of Fire (LOF) and Spotting are different matters...

A distinction that sometimes makes itself painfully (for the poor bloody pTruppen) known at times... :-/

1. Why can't you Area Target a point on the map if you can't see the middle of it's Action Spot, yet if there is a unit there you can target just fine?

This is the issue the thread was started for. Thanks again for taking the time. In my OP, I raised 5 initial circimstances where this was a significant problem, and I can certainly see that the issues involved with targeting "any old AS" with deadly or smoke ammo are intractable in the current architecture, which leaves the three inabilities to shoot at terrain objects which are eyeballable, but on hidden ground.

A unit can only trace LOS to centers of Action Spots or other units. That's because in each case CM has two distinct points to draw a line to. The first is specific friendly point to generic fixed point (Action Spot center), the second is specific friendly point to specific enemy [or terrain - Ed] point. What CM can't do is trace specific to random because that hits the computer too hard in context with all the other demands placed on it.

To move a previous statement here so it can be considered in a new context:

A unit either occupies a specific spot (soldier) or a range of spots (vehicles, which includes big guns).

And now back to sequence, with an elision...

Now, if a spotted enemy unit happens to be in the same exact spot, well... that's different! CM understands that the two units already know where each other is and therefore the Action Spots are, for the sake of argument, ignored.

So, given the constraints above, how would it be if buildings and other objects with a vertical component had a targetable spot in the middle of each side of non-negligible dimension, per "building elevation level"? The targeting tool already calculates the intersection with the plane of the face, in order to determine whether it blocks LOS, and knows to ignore it for the next AS past the wall on the line of sight (so we can target facades as per currently). These spots could exhibit a hybrid of AS and unit behaviours. They wouldn't need to be considered for the "general map" (as unit target spots don't), since that's already calculated using the LOS-blocking properties of z-axis objects which is far more complete, nor for spotting (like all terrain, there's no FoW). They would only need to be considered one at a time, since there's only one targeting tool active at a time, and only when the targeting tool has already detected an intersection with a LOS-blocking plane that is not the terrain grid, which could maybe set aside the consideration of AS-visibility in the same fashion as when a unit is targeted.

I don't think there's a general need or desire to be able to target any square decimetre of a building's facade, but it would be a huge step forward to be able to shoot at any floor you can see at least the middle of. Yeah, it's another advantage for the player over the AI, but it would greatly reduce one of the big frustrations.

BTW, it is wrong to say there is no grazing fire. Dead wrong. It's in the game and it works perfectly well. What you can't do, at least easily, is ensure that a specific spot of ground is covered by a specific height of fire.

Yeah, I use grazing fire a fair bit. Decide what you want to suppress, then drag your LOS cursor out past it until it intersects with something you can target. Bullets whizzing past overhead do work to keep heads down.

3. Why can't you use artillery to target areas you can't see?

[snip]

'Nother reason to bookmark your post to be able to refer people to a "horse's mouth" explanation...

Hope that helps!

It certainly does. I hope in return that you guys can at some point find a way to do something to mitigate one of the remaining frustrations of the engine. I certainly feel it's worth some diversion of processor and development cycles to get some sort of resolution, and I don't recall any posts disagreeing that it'd be a welcome improvement if it could be done to a useful degree. I don't even dare to imagine any such large amendment could be done before v3, and it's obviously a big deal so wouldn't be the kind of thing a patch could address... v5 maybe? Our computers will be better then...

[A Canadian Cat]

Thanks for popping in, and I'm sure we'll do our best. I'd better bookmark this post for reference!

Yes, thanks - I am pretty sure I have read that kind of information before. Given that this (the inability to hit buildings that you cannot see the centre of) is an issue that I run into often, especially in an urban environment, I would like to see if there is something that can be tweaked.

So, given the constraints above, how would it be if buildings and other objects with a vertical component had a targetable spot in the middle of each side of non-negligible dimension, per "building elevation level"?

Interesting idea if it is workable it would certainly help.

I had an alternate suggestion. Add 8 additional spots in each AS (so each would have 9 with the current one in the centre). Those extra 8 would be nothing more then target choices. They would *not* used for any of your current LOS or spotting calculations. So no additional load as the game just runs. But when the orders are being calculated for targeting using the area target tool then those other locations would be eligible to be the "other end" of the target line.

Clearly a non trivial task but being able to aim at the corner of a building would be extremely helpful and probably put an end to this kind of frustrating problem during urban combat.

[Battlefront.com]

Thanks for popping in, and I'm sure we'll do our best. I'd better bookmark this post for reference!

Certainly a good idea This is one of the rare instances of CM, as a game, having to be a slave to limited computer resources. And that requires an unusual amount of understanding of the underlying game engine's design.

That's certainly true while the grid is the only thing you can use. But the grid isn't the only thing you use now: units are spotted without the assistance of AS grid centres (at the final assessment).

Not true. Spotting is solely dependent on the Action Spot grid as the first step. Everything initially uses the Action Spot grid. If two units are in two Action Spots which can not draw center to center they will *never* spot each other.

However, once a unit is already spotted it is tracked point by point so that LOF, which is point to point specific, determines if something can/can't happen. As I said, this causes a big hit to computing resources but it is usually acceptable because the number of units actively needing such calculations is usually manageable.

Back in CMx1 days we had similar problems, but with a much cruder game system. You could have a game with a battalion of tanks on either end. This worked fine provided no more than a couple of platoons were in action at any given time. But if you put them all within LOS of each other at one time, the system ground to a halt. Rune made an infamous "Battle of the Bulge" scenario that took something like 30 minutes to crunch a turn and it ran at about 1 or 2 fps on my system which, under normal circumstances, didn't have a problem.

Similar issues with CMx2, but obviously it's capable of handling a LOT more detail. Partly because of better hardware, but mostly because of the Action Spot system.

Ah. While forming my response, I've stumbled across a realisation that this is always going to be true in flowing RT play. When the game is in pause or during the order phase, the considerations of map size and unit count can potentially be given significantly lower priority while the targeting tool is being used, since pretty much the sole focus of the player is the use of that tool. Background functions (unit animations, sound, ambient terrain animations and the like) could be halted entirely, if necessary to free up the Hz for the LOS calculation. However, since the tool has to work in the dynamic environment of the RT game unpaused, such parking isn't practical and the "suspended" modes of play have to work within that limitation.

Theoretically true, but I'm not sure it would work like that in WeGo.

There is something called an "LOS Map" which is precalculated when the map is made and it is loaded into RAM for rapid access during the game. This basically tells CM, ahead of time, which Action Spots can see which Action Spots. There are *no* on-the-fly calculations going on. And that's why your Target tool works snappy under all circumstances.

When CM draws LOF it *only* draws between two known points, which it can handle fine in most situations on most computers.

What you're theorizing is that the Target tool will continue to work smoothly if the LOS Map is bypassed and instead the tool continuously draws point to point. For short distances, even in complex terrain, I have no doubt it can handle it. But stretch that line 1000m in complex, broken terrain... I'm not so sure it can. In fact, I'm wiling to bet that it could not.

Which is to say that even if we were willing to consider an entirely different method of play for WeGo vs. RealTime I don't think it would work for most people enough of the time to be viable.

So, given the constraints above, how would it be if buildings and other objects with a vertical component had a targetable spot in the middle of each side of non-negligible dimension, per "building elevation level"?

I'll answer this specifically in my next post.

The targeting tool already calculates the intersection with the plane of the face, in order to determine whether it blocks LOS,

Actually, it doesn't. At least not technically. All the Target Command does is regurgitate the precalculated LOS Map information. There's nothing on-the-fly and the Target Command line itself does absolutely nothing.

I don't think there's a general need or desire to be able to target any square decimetre of a building's facade, but it would be a huge step forward to be able to shoot at any floor you can see at least the middle of. Yeah, it's another advantage for the player over the AI, but it would greatly reduce one of the big frustrations.

I agree that what we call the "oblique angle" problem sucks when you encounter it. But it's not just buildings or vertical surfaces, it's all Action Spots where the center is "juuuuuuuust barely" out of view. Believe me, we spent a ton of time with this back in 2006-2008 when the issues were far worse than they are now.

The reason buildings seem to be more affected by this than other terrain is because in more open environments usually landing Area Fire on the adjacent Action Spot produces a sufficient result. This is not true for buildings or vertical surfaces which fall into the "oblique angle" problem.

Yeah, I use grazing fire a fair bit. Decide what you want to suppress, then drag your LOS cursor out past it until it intersects with something you can target. Bullets whizzing past overhead do work to keep heads down.

Works great provided you can draw LOF to something beyond the point in question. Fortunately that's usually possible.

It certainly does. I hope in return that you guys can at some point find a way to do something to mitigate one of the remaining frustrations of the engine.

It is possible that we could do something about it now or fairly soon that would work on pretty much current higher end computers. What we have to consider is that the bulk of people do not have such systems. Since this feature is one that must be shared by all people all the time, I don't think we'll get this sufficiently fixed for a while yet.

Steve

[Battlefront.com]

OK, here's an answer to something Womble had addressed. I wanted to separate it to make it easier to find.

I had an alternate suggestion. Add 8 additional spots in each AS (so each would have 9 with the current one in the centre). Those extra 8 would be nothing more then target choices. They would *not* used for any of your current LOS or spotting calculations. So no additional load as the game just runs. But when the orders are being calculated for targeting using the area target tool then those other locations would be eligible to be the "other end" of the target line.

Unfortunately that's not how it works.

As I mentioned above, the LOS Map is the heart and soul to the whole Action Spot system. It's what gives CM it's speed, at the cost of extra load time and a lot of RAM that is proportional to map size and map complexity. The bigger and more complex the map, the bigger and more necessary the LOS Map becomes, the more time it takes to load, and the more RAM it takes once loaded.

What you're suggesting means that the number of possible point to point connections that have to be precalculated goes up proportional to the number of buildings involved and (worse) the amount of LOS possibilities that exist (i.e. higher buildings in clearer terrain VASTLY increases the number of LOS possibilities vs. a 1 story house in the middle of a forest). This likely means an exponential increase in the complexity and size of the LOS Map. Or at least we'd have to plan on that happening since we have no restrictions on what people do with the Editor. And that gets me to this point...

Some reading this might think "hey, if ARMA 3 can handle huge maps with point to point LOF, why can't you?". The simple answer is that ARMA, and games like it, impose game restrictions that CM players would never accept.

Ask yourself... what is the target maximum number of point to point LOF calculations that happen in ARMA 3? About 64 (above that practical server and bandwidth issues nix larger amounts). In real world military terms that is roughly 2 infantry platoons. How many games do you think you guys would like to play if we tailored the game to work with 2 infantry platoons? Exactly

Another one is that in ARMA the LOF function only happens when the shooter is actually shooting. There's no pre-tracing a possible aim point, as happens when you move the Target Command around. This means the CPU and RAM need only be able to handle roughly 64 LOF calculations at any specific time, worst case.

Even better, as far as I know there are no spotting calculations at all since that's left to the player's own eyeballs. If you can see it with your eyes you have spotted it. If you don't then you haven't. CM has to handle hundreds, potentially thousands, of entities checking constantly.

The list of restrictions and limitations of ARMA 3 designed to make the game practical on even high end machines is as long as my arm. CM, unfortunately, can't do almost any of those things because people would no longer be interested in playing CM.

Steve

P.S. When I initially proposed the Action Spot system to Charles back in 2004 I had each floor consist of "Mini Spots". There was one for the center, one for each wall, and one for each corner. Why did I do that? Because I anticipated these sorts of problems before coding had begun on the game portion of the engine (Charles was working on the graphics elements already). So trust me when I say we understand the issues

[Skwabie]

OK, here's an answer to something Womble had addressed. I wanted to separate it to make it easier to find.

Unfortunately that's not how it works.

As I mentioned above, the LOS Map is the heart and soul to the whole Action Spot system. It's what gives CM it's speed, at the cost of extra load time and a lot of RAM that is proportional to map size and map complexity. The bigger and more complex the map, the bigger and more necessary the LOS Map becomes, the more time it takes to load, and the more RAM it takes once loaded.

What you're suggesting means that the number of possible point to point connections that have to be precalculated goes up proportional to the number of buildings involved and (worse) the amount of LOS possibilities that exist (i.e. higher buildings in clearer terrain VASTLY increases the number of LOS possibilities vs. a 1 story house in the middle of a forest). This likely means an exponential increase in the complexity and size of the LOS Map. Or at least we'd have to plan on that happening since we have no restrictions on what people do with the Editor. And that gets me to this point...

Some reading this might think "hey, if ARMA 3 can handle huge maps with point to point LOF, why can't you?". The simple answer is that ARMA, and games like it, impose game restrictions that CM players would never accept.

Ask yourself... what is the target maximum number of point to point LOF calculations that happen in ARMA 3? About 64 (above that practical server and bandwidth issues nix larger amounts). In real world military terms that is roughly 2 infantry platoons. How many games do you think you guys would like to play if we tailored the game to work with 2 infantry platoons? Exactly

Another one is that in ARMA the LOF function only happens when the shooter is actually shooting. There's no pre-tracing a possible aim point, as happens when you move the Target Command around. This means the CPU and RAM need only be able to handle roughly 64 LOF calculations at any specific time, worst case.

Even better, as far as I know there are no spotting calculations at all since that's left to the player's own eyeballs. If you can see it with your eyes you have spotted it. If you don't then you haven't. CM has to handle hundreds, potentially thousands, of entities checking constantly.

The list of restrictions and limitations of ARMA 3 designed to make the game practical on even high end machines is as long as my arm. CM, unfortunately, can't do almost any of those things because people would no longer be interested in playing CM.

Steve

P.S. When I initially proposed the Action Spot system to Charles back in 2004 I had each floor consist of "Mini Spots". There was one for the center, one for each wall, and one for each corner. Why did I do that? Because I anticipated these sorts of problems before coding had begun on the game portion of the engine (Charles was working on the graphics elements already). So trust me when I say we understand the issues

I thought of somewhat the same idea as ian.leslie but now, perhaps not. Having experienced frustrations as well in the LOS/LOF department, the conclusions I reached is a return to the desert maps with more open terrain...

Edit: But since the LOS map is already pre-loaded, maybe an "LOS disc" option mentioned somewhere before wouldn't be such a bad idea. It sure would greatly reduce the tediousness of positioning the units, and foretell the player where he would likely run into the LOS engine problem so as to expect it or re-position his unit.

[womble]

I said:

...the grid isn't the only thing you use now: units are spotted without the assistance of AS grid centres (at the final assessment).

I worded that badly, may even have mistyped. I don't think I meant "without the AS grid"... I was thinking, I think, of units that are in places where "the ground can't see the ground".

Not true. Spotting is solely dependent on the Action Spot grid as the first step. Everything initially uses the Action Spot grid. If two units are in two Action Spots which can not draw center to center they will *never* spot each other.

So how can a unit spot a unit in an AS which it can't fire at unless there's a unit there? [puzzled] This happens all the time (mostly to me, as I move a unit through a place I didn't think the enemy could see... ) and is the root of the question we're asking. I get that there are AS-AS pairs that really never can have mutual LOS (there's a hill in the way, or half a dozen buildings and 400 trees). It's the "second step" after the LOS map has decided that "if a unit is present in the AS, it might be visible from there" where the issue lies. What I'm suggesting is that the spot on the wall be a sort of "latent" unit-type thing, which only the targeting tool is concerned with, and only once the LOS map has said it's worth checking a unit. So the same process that says "there's an infantry unit on that floor you can shoot at" also recognises that "there's a target spot on that wall you can shoot at".

However, once a unit is already spotted...

Maybe it would help if you explained a bit how that happens, particularly in the case when the unit is in an AS that, absent the unit, the observer wouldn't have LOS to? Also, does the LOS map account ahead of time for all the possible observation sources in the AS, from a prone pTrooper to the commander of a Sherman? Those make a difference, as does the precise location in the observing AS. The most common example of this I find is the "cowering gunner" who can't see over the bocage berm to area target the next berm, one minute, but once he's mustered the courage to stick his precious bonce in the way of potential bullets, he can, without having moved position at all. That sort of assessment certainly involves some elements of on the fly consideration. And on the target end, how does a prone, unspottable infantry team get converted in the game's processes into a spottable, kneeling infantry team?

There is something called an "LOS Map" which is precalculated when the map is made and it is loaded into RAM for rapid access during the game. This basically tells CM, ahead of time, which Action Spots can see which Action Spots. There are *no* on-the-fly calculations going on.

Except, surely, for the units (that aren't, AIUI, in the LOS map)?

The reason buildings seem to be more affected by this than other terrain is because in more open environments usually landing Area Fire on the adjacent Action Spot produces a sufficient result. This is not true for buildings or vertical surfaces which fall into the "oblique angle" problem.

I don't think it's to do with the lack of an alternative means of effect, so much as it is to do with the sheer perceptual disconnect. "I can see the window they were shooting out of; why can't I target it." Shooting into a vast field of waving corn isn't anywhere near so distinct.

[A Canadian Cat]

sheer perceptual disconnect. "I can see the window they were shooting out of; why can't I target it." Shooting into a vast field of waving corn isn't anywhere near so distinct.

That sums up the problem for sure.

[A Canadian Cat]

Unfortunately that's not how it works.

As I mentioned above, the LOS Map is the heart and soul to the whole Action Spot system.

Bummer, thanks for the detailed explanation. Adding all that to the look up table would clearly not be work able.

I am left with the same question though : if a unit in a building can see my men and then ducks down I can no longer target the front of the building they were just seen in. But it spotting only works between AS es that have LOS then how did those unit see and fire at each other in the first place?

[PhilM]

Bummer, thanks for the detailed explanation. Adding all that to the look up table would clearly not be work able.

I am left with the same question though : if a unit in a building can see my men and then ducks down I can no longer target the front of the building they were just seen in. But it spotting only works between AS es that have LOS then how did those unit see and fire at each other in the first place?

I read the explanations - thanks Steve - twice (!), still not sure I understand them ...

But based on a quote from Steve that

"CM understands that the two units already know where each other is and therefore the Action Spots are, for the sake of argument, ignored."

... is this (my bold bit above) because, when the unit in the building is, say, kneeling up at a window and firing out at your guys, it is a unit to unit spot and LOF ... but when they move away from the windows and, say, lie down in the room centre there is no longer a unit-unit LOS, and it becomes - or rather, fails to become - a unit-AS centre LOS, which doesn't exist?

So, for as long as the guys in the building are face down away from a wall/window, they too don't have LOS and so cannot fire? Fire is only possible unit to unit, and not unit to AS?

Make any sense at all??

[Battlefront.com]

So how can a unit spot a unit in an AS which it can't fire at unless there's a unit there? [puzzled]

I've asked Charles for some clarification on this point as it's been a loooong time since I've got into this much of a nuts and bolts discussion. But as I remember it, this isn't possible.

What can happen is Friendly Unit has knowledge of Enemy Unit because either they were previously in two Action Spots that could see each other or the information was passed on by another unit. In either case there is now a "special relationship" between the two units which supersedes the Action Spot center requirement. As soon as the two units break contact that relationship disappears and now it's purely Action Spot related.

Again, think about LOF rules. These are from point to point. If you see an individual enemy soldier, and can draw LOF to it, then you can shoot it. The Action Spot's center is irrelevant. The enemy unit can not be unfairly penalized so it has a chance of spotting who is shooting at it and can theoretically shoot back. Again, the Action Spot center is not relevant because we have two known points. Take away one of those units and now it is a known point to an Action Spot center.

What I'm suggesting is that the spot on the wall be a sort of "latent" unit-type thing, which only the targeting tool is concerned with, and only once the LOS map has said it's worth checking a unit. So the same process that says "there's an infantry unit on that floor you can shoot at" also recognises that "there's a target spot on that wall you can shoot at".

This breaks Fog of War rules. Why should a player be able to move the LOS tool around and find "oh, I didn't see an enemy unit here, but it's allowing we to target off-center so there must be one there. Whoops! I can't target there any more, so it must have moved". That's a deal killer right there. The technical implications may be another reason for this not being possible, but I can't speculate on that.

Also, does the LOS map account ahead of time for all the possible observation sources in the AS, from a prone pTrooper to the commander of a Sherman?

It cross references all the heights.

Those make a difference, as does the precise location in the observing AS. The most common example of this I find is the "cowering gunner" who can't see over the bocage berm to area target the next berm, one minute, but once he's mustered the courage to stick his precious bonce in the way of potential bullets, he can, without having moved position at all. That sort of assessment certainly involves some elements of on the fly consideration.

No, it doesn't actually. When the soldier is prone (Height 0) the LOS Map is consulted and it's determined that it can't see in that particular direction, though it can in other directions. When the soldier pops his head up the LOS Map is consulted again, but using a different criteria (Height 1) for determining what can be seen. Going prone again means consulting the LOS Map again (Height 0). While certainly the results change, the calculations are not done on the fly, only the decision which set of data to use is. Huge difference in terms of processing time.

This also serves as a hint about how often LOS is checked. Yup, every time a Soldier changes position it changes his LOS point of view, which means LOS needs to be reconfigured. Doing that on the fly is simply not possible. Doing it to any point on the map isn't either.

And on the target end, how does a prone, unspottable infantry team get converted in the game's processes into a spottable, kneeling infantry team?

Spotting is a separate function. A Conscript Soldier standing in the open is more spottable than another member of his Team that is prone behind a wall. Which is why the system allows you to spot individuals within a Team and not have it be an all or nothing approach. Getting better intel on the unit is an intelligent aggregate of the individual information already obtained.

Except, surely, for the units (that aren't, AIUI, in the LOS map)?

Yes, as I said these are exceptions because there are two known points involved. The system CAN handle a certain number of these at one time, but I've also said that the game as it exists today can be whacked pretty hard if too many of these things are going on simultaneously and/or if your computer isn't all that great.

I don't think it's to do with the lack of an alternative means of effect, so much as it is to do with the sheer perceptual disconnect. "I can see the window they were shooting out of; why can't I target it." Shooting into a vast field of waving corn isn't anywhere near so distinct.

Not quite. If you can see a unit shooting out of a window then you can target that unit, provided LOF is possible. The issue only comes up when you can NOT see the unit shooting out of the window, but suspect it is still there, and you can not draw LOS to that Action Spot's center from where you are. At that point you want to do Area Fire and aren't allowed to. But as soon as that unit pops back up again, you can target it again.

Bummer, thanks for the detailed explanation. Adding all that to the look up table would clearly not be work able.

I am left with the same question though : if a unit in a building can see my men and then ducks down I can no longer target the front of the building they were just seen in. But it spotting only works between AS es that have LOS then how did those unit see and fire at each other in the first place?

Hopefully the above answers your question!

... is this (my bold bit above) because, when the unit in the building is, say, kneeling up at a window and firing out at your guys, it is a unit to unit spot and LOF ... but when they move away from the windows and, say, lie down in the room centre there is no longer a unit-unit LOS, and it becomes - or rather, fails to become - a unit-AS centre LOS, which doesn't exist?

You get a gold star on your forehead!

So, for as long as the guys in the building are face down away from a wall/window, they too don't have LOS and so cannot fire? Fire is only possible unit to unit, and not unit to AS?

Correct. That's because the unit in the building needs to have LOS between Action Spot centers just as the unit outside the building does. Which is why the game is so relentlessly fair There are no situations in CM where LOS and LOF are anything other than fundamentally reciprocal. Spotting is also inherently reciprocal, but battlefield conditions can make the actual results for a specific situation one sided.

Steve

[A Canadian Cat]

What I'm suggesting is that the spot on the wall be a sort of "latent" unit-type thing, which only the targeting tool is concerned with, and only once the LOS map has said it's worth checking a unit. So the same process that says "there's an infantry unit on that floor you can shoot at" also recognises that "there's a target spot on that wall you can shoot at".

This breaks Fog of War rules. Why should a player be able to move the LOS tool around and find "oh, I didn't see an enemy unit here, but it's allowing we to target off-center so there must be one there.

That is not what he means. He is suggesting that buildings always have a "unit-type thing" that can be targeted. So when faced with an unobstructed view of a building you could target multiple points on each floor on the face of the building instead of targeting the centre of the building (which I guess you could so as well).

That way when you see an obstructed view of a building you can still target one of the "unit-type thing" on each floor.

[A Canadian Cat]

Bummer, thanks for the detailed explanation. Adding all that to the look up table would clearly not be work able.

I am left with the same question though : if a unit in a building can see my men and then ducks down I can no longer target the front of the building they were just seen in. But it spotting only works between AS es that have LOS then how did those unit see and fire at each other in the first place?

<your actual explanation snipped>

Hopefully the above answers your question!

Yes, for sure.

Reordered:

Not quite. If you can see a unit shooting out of a window then you can target that unit, provided LOF is possible. The issue only comes up when you can NOT see the unit shooting out of the window, but suspect it is still there, and you can not draw LOS to that Action Spot's center from where you are. At that point you want to do Area Fire and aren't allowed to. But as soon as that unit pops back up again, you can target it again.

Which is the problem. When the enemy team ducks down I cannot give target orders to supporting units and keep their heads down. I have to wait until they popup again so my units are exposed and I cannot lay down the suppressive fire that I really should be able to do.

You have expressed the issues with computational power etc that mean your options are limited to improve things. I get that. We are just shooting ideas around in the hopes that it inspires a way to solve the problem. I know, wishful thinking, since we don't have the detailed knowledge of how the system works.

This has bothered me since I ran into it in my first major scenario and I have personally had a problem with this issue in pretty much ever game that involves more then farm or two that I have played. I continue to hope that some kind of change can be made that will make this less of an issue.

Thanks for spending some time thinking about this and sharing with us some of the issues.

[Freyberg]

These days I often just give a 'face' command and let the units do their own thing.

[eniced73]

http://www.battlefront.com/community/showthread.php?t=111250&highlight=attached+targeting+order&page=6

Ian found this out above which worked. BF deemed it gamey and eliminated it in a patch. I tried to search around but could not find any evidence of people complaining about this being a "gamey" feature. If anything I think they created a problem which you guys are discussing here. If they would have left well alone everything would not be perfect but satisfactory in my opinion. This little hint also allowed you to fire through that little invisible forcefield called "smoke". So if this is deemed gamey then using smoke should be eliminated from the game as being just as "gamey". Who was the idiot the complained on this being gamey in the first place? just curious as to why BF changed it as I did not see many people complaining.