Naive Spotting Question

[Philippe]

Does the CM spotting routine automatically spot something that is within its line of sight?

Just because I can see something doesn't mean that I will notice it.

I may have a perfectly clear line of sight to that artillery spotter in the open on that ridgetop two miles away to the north, but I may be paying closer attention to the North African swallow 200 yards away to the east.

I just looked at the two pages mentioned in the index of the CMBO manual that talk about spotting, and neither one seems to deal with this aspect.

In case I wasn't clear, the probability that I will notice the North African swallow is 100% because it is very close and carrying a coconut. The probability that I will notice the artillery spotter, even though he is in the clear, is more like 50%, unless I happen to be looking for him. Even if he's in the open, unless that ridgetop is completely bare he won't catch my attention. He could get temporarily ignored because, even though he's against the skyline, he's squatting in a twenty yard wide clearing between clumps of trees and maybe the odd church steeple.

It probably doesn't happen so much in CMBB and CMAK because of the covered arcs, but I'm constantly amazed when one of my gunners decides to start shooting at something on the other side of the map that I hadn't even realized was there.

Then again, I'm a rather daft puppet-master.

[And yes, the real title to this thread should have been Prolegomena for a Final Solution to Borg-spottting].

[Splinty]

But North African Swallows unlike arty spotters aren't migratory.

[YankeeDog]

I'm not sure if I exactly understand what you're going for, but here's how the CM spotting rules as I understand them:

(1) For any one unit, having LOS to a specific location does NOT necessarily mean that you will spot an enemy at that location. There's lots of factors that go into determining whether or not any one of your units actually spots and enemy at a given location, including distance, light conditions, available cover, profile of the enemy unit (i.e., sharpshooters are a lot harder to see than tanks), experience level of both the spotter and the (potentially) spotted, morale/suppression, whether or not the unit is moving and/or firing, etc.

With all these different factors, there are no universally applicable guidlines to know for sure whether or not you will spot enemies at a given range. You can get a general idea of how likely you are to spot, though, by looking at the color of the LOS line -- if it's light blue, your LOS is very good and your unit will have a good chance of spotting. If it's dark blue, your LOS is poor and you might not see enemies in the area, especially low profile enemy units like small infantry teams.

And yes, at long distance, even if you have a good, light blue LOS line, very low profile infantry teams like sharpshooters and tank hunters can somtimes remain unspotted even in open ground. They usually get spotted as soon as they try to move, though.

However, once an enemy unit is spotted by any ONE of your units, ALL of your units with LOS to that unit can see and are aware of that enemy unit, not matter how marginal their LOS is (i.e., how dark blue the LOS line is). All units with LOS will therefore react accordingly, targeting said enemy if the TacAI routines deem it appropriate. This is the infamous "Borg Spotting" issue.

There are a few exceptions to the above. For examply AFVs, and especially buttoned AFVs, appear to be slower than infantry units to react to new-appearing enemy units. I speculate, therefore, that there is some sort of fudge in the code here. When an AFV "misses" an enemy in it's LOS, you can still manually order the AFV to target the enemy, but it often won't on it's own for a turn or more, even if the enemy unit is in good LOS and at a range where the AFV can really hurt it.

Green and Conscript units also sometimes let fairly decent targets go by unscathed for quite a while, especially at medium to long range. Just like with buttoned AFVs, they will still obey manual targeting orders. I suspect this is intended to model the slow reaction time of such poorly trained troops.

Cheers,

YD

[dieseltaylor]

If you go to the MZO site a chap called POS did a fine series of jpgs. showing spotting ranges etc.

A fine site - there is a link on the top of the CMBB forum messages

[Philippe]

Thanks for the cogent and detailed reply, YankeeDog.

It occurs to me that the program could keep a yes/no awareness checklist for every sentient object listing all the things that it might be able to see (bad news for huge scenarios). For example, my Fallschirmjaeger squad would have a potential awareness list that includes the entire oob for both sides as well as all the buildings, bunkers, bridges, and anything that starts with the letter b.

If my Fallschirmjaeger can see something but isn't aware of it, the AI could be blocked from reacting to it. Or at the very least prevented from shooting at it. Good news for bumblebees on bridges.

The awareness list has another use. If my Fallschirmjaeger can't see something, but my sniper can, my Fallschirmjaeger can only react to it if it passes the awareness check. This could probably be done by some kind of arbitrary rule stating that awareness communicates up and down the chain of command only if the next step in the link is x [= a very small number] meters away and both are in command relative to each other [computer does another numbers crunch...].

So in the case of the sniper, he may see the bumblebee on the bridge and be aware of it, but he is unlikely to communicate that to the next unit up the chain of command unless they both use walkiie-talkies. Since his CO will probably fail the bumblebee awareness test (unless he happens to be standing very close to both the Fallschirmjaeger and the sniper), the Fallschirmjaeger will fail it as well. Depending on how intrusive the programming would have to be the Fallschirmjaeger would then be blocked from shooting at the bumblebee (except through area fire) or having its AI react to it (harder).

To make this work, you would also have to hijack an optional feature from Jagged Alliance whereby you only see what the unit you have selected sees (or is aware of). This would have the accidental effect of making officers more important by adding an extra layer to the FOW, because if you click on nothing you will see nothing, and to get any sense of what was going on on the battlefield you would have to start clicking on the different unit commanders. And yes, you will probably start losing your units in a big battle when they get separated from their HQ's.

What I'm describing will trigger a lot of behind the scenes numbers crunching, but I think modern computers could probably handle it, and it has the side benefit of making it much more important to keep everyone in command control.

[Moon]

What I'm describing will trigger a lot of behind the scenes numbers crunching, but I think modern computers could probably handle it, and it has the side benefit of making it much more important to keep everyone in command control.

That's a bold statement While developing the series, Steve and Charles learned early on to be very careful with this kind of assumptions. Fact is that back when the engine was first written, computers were definitely not up for this kind of number crunching. Which is the reason for the borg-spotting implemented in CM - a unit is automatically aware of an enemy that is within LOS and spotted by another unit. Individual awareness checks like you describe were impossible to implement without bringing the hardware to its knees (and lower... )

Since CMBO, the faster number crunching allowed people to play even bigger scenarios, but even today's hardware has its limitations.

But well... who knows what Charles will manage for the new CMX2 engine

Martin

[Redwolf]

The CM spotting model is simpler than many people think and I don't recommend researching it as it can spoil the game (I stopped before it did).

What is mainly missing, and for reasons Moon mentioned (CPU usage) is a random factor. In the real world, you can rush from one house to another and if only one or two people scan the area you might be lucky. Or not. It's essentially random. In CM it is not, if you rush and the sighting possibility is given you will be spotted 100% of the times. Even if you had the CPU power to throw a dice each time that would still be fairly useless as long as there is as much absolute spotting as there is now.

That's for moving in the open.

Other than the missing random factor, the spotting model is by concealment, whether the spotter is hiding (huge loss), whether the target is under control of a stealth bonus HQ, whether the spotter has binoculars (huge win). There is not much difference by target unit type, among infantry only the sharpshooter is special and vehicles suck.

There are various oddities about spotting from vehicles I rather not tell to not spoil the game (I don't exploit that knowledge when playing).

When a target fires most of these conditions fall, all spotters and spotting conditions become mostly the same.

[Other Means]

i really don't know how BFC are going to get around allowing area fire.

after all, the player is Jambooba in these circumstances.

i suspect it'll only be allowed if a unit has been seen by a unit that shares a radio net with the firing unit. and then after a delay to consider command delay.

but then what happens to recon by fire?

good luck lads.

[Redwolf]

Err, you are talking about fixes to absolute spotting now?

[Seanachai]

Originally posted by Other Means:

after all, the player is Jambooba in these circumstances.

Even Cesspoolers know it's 'Jamoomba', you tit.

[Other Means]

Originally posted by Seanachai:

</font><blockquote>quote:</font><hr />Originally posted by Other Means:

after all, the player is Jambooba in these circumstances.

Even Cesspoolers know it's 'Jamoomba', you tit. </font>

[YankeeDog]

Not a complete fix, but a couple of things that would actually work in the current engine to reduce the "Jamoomba" effect in re Area Fire:

1) Make Area Fire subject to command delay just like other orders. This would further slow the time it to activate an Area Fire command, and also realistically make Area Fire orders execute faster for units in command and/or with radios than those without.

2) Make Area Fire even less accurate. Right now, for example, if you order a tank to area fire HE onto the position of a known (but presently unseen) AT gun in a tree line, the fire falls in a pretty concentrated cluster around the aim point given by the player. It's a little bit less concentrated than direct, targeted fire, but not by all that much. I think it would be more realistic to spread out Area Fire even more to simulate the firer searching for a target. I would add one exception to this: Buildings should be accurately targetable objects. Buildings provide an easily communicatable reference point for communicating targeting orders, yet another reason why it's not such a good idea to hole up in them when the enemy has big HE around. It's a lot easier to specifically define a target such as, "Hit the farmhouse with the red roof just North of the bend in the river," than it is "Hit the tree line about 100m West of the large pine tree on the crest."

Inevitably, you're never going to be able to completely eliminate some of the unrealistic aspects of player control without making CM a very unfun game, but I do believe there are possible improvements that could dramatically reduce the player's ability to take advantage of his god-like perspective, at least in immediate tactical action.

Cheers,

YD

[tar]

I agree with YD. In fact, I think all player-given orders should be subject to command delay: Area Fire, Specific unit targetting, rotation.

[Redwolf]

Ideally, the area fire command should not be to a point, but to a circle (TacOps has something similar for fire triggers).

Then, you limit the minimum radius of the circle depending on spotting situation and you limit expoiting absolute spotting.

Another huge advantage of the circle area is that it finally allows sweeping MG fire on a whole area.

But I think the worst thing about absolute spotting right now, and it would be fixed by giving each units is own idea of which enemies are spotted, if the overwhelming return fire in anti-armor ambushes. If you button the enemy tanks up you should be able to pick them up more or less at the same rate that you can spot them. But in CM, but to too limited zeroing in and due to the fact that everybody shoots back at you almost instantly the ambusher almost always dies while firing at the first target.

[birdstrike]

I fear there might be no real CPU-friendly way to get around "borg-spotting". With dozens or well over a hundred units on the battlefield, it seems unlikely that individual spotting can be implemented without system requirements skyrocketing.

Really not much I can think of to solve this.

AKAIK in RTS games this is compensated by a limited firing range of all the units - but that's something which would do more harm than good in a game like CM.

But just out of interest: can anyone remember how this was handled in the Close Combat series?

I seem to recall there also was some kind of Borg spotting wasn't there?

[YankeeDog]

Don't remember with CC. That game generally had far less units per side than CM, so it had an easier job of it.

It also was 2D, which vastly simplifies the LOS and spotting calculations.

As to what's really possible with CMX2, CPUs and "fixing" Borg spotting, only time will tell. Computers are a factor of magnitude faster now than they were when CM first came out, but as you note the requirements for individual, or "relative" spotting as it's sometimes called, go up exponentially with the number of units involved.

There might be ways of abstracting it and reducing the processor load to a degree. For example, at least for infantry units it might make sense to simply assume that all units "In Command" of a specfic HQ see what all other units under command of that HQ see -- IOW keep "Borg Spotting," but just within platoons. This would reduce the number of units that would need to be tracked for spotting purposes by a factor of about 4.

There are probably other abstractions that could be made, too. We'll just have wait and see what Charles cooks up. . .

Cheers,

YD

[Redwolf]

Originally posted by birdstrike:

I fear there might be no real CPU-friendly way to get around "borg-spotting". With dozens or well over a hundred units on the battlefield, it seems unlikely that individual spotting can be implemented without system requirements skyrocketing.

It's not free but not prohibitive.

Right now both sides can have around 256 units.

The native approach would be to have one linked list for each unit, but since you know the maximum number of units in the scenario you can allocate an array, saving the pointers. So in the extreme you have 256 friendlies * 256 enemies * 2 side boolean variables, where the boolean variable would typically be a bit, a byte or a word. If it's a byte you use 128 KB. That's nothing.

That's for memory. Far worse is runtime overhead but much of it you already have.

When unit move you have to figure out who will not spot whom. However, while it is expensive, you already do that right now. The current scheme is a form of optimization in that you only have to do the check for units not already spotted. With individual spotting you can't do that unless every enemy spotted this unit which usually won't happen and isn't worth special-casing for.

I could go on speculating but I think it is pretty clear that BFC is going to implement this.

[birdstrike]

I hope you're right guys, would be glad to see that my old comp (which isn't that old actually) is still capable of running the next CM.

[YankeeDog]

Well, on the plus side, assuming CMX2 sticks with a "WEGO" format similar to the current engine (which is all but certain), I'm pretty sure the aforementioned relative spotting requirements would affect turn calculation issues only, and would not affect graphics rendering requirements.

So, to a point presumably older computers would just take longer to calculate turns, and so long as they had a half decent graphics card (something that's also relatively inexpensive to upgrade) would still be able to display the movie phase just fine.

I'm just making semi-educated guesses, tho. It is worth noting that in the past BFC has generally been pretty good about keeping the game at least runnable on less than state of the art systems. To put things in perspective, I first started playing CMBO on a 300Mhz PII with 128MB ram and a 16MB graphics card. It ran great (even w/some hi-res mods), and turn calculations were not overly onerous.

So BFC can increase the minimum hardware requirements of CMX2 a fair amount over the old engine, and still be nowhere near what's currently 'cutting edge'.

Cheers,

YD

[DavidFields]

IIRC, CC also had Borg spotting.

[arax3]

In a recent game, I was playing the Allies. The company CO had binocs and a radio. He advanced to the edge of some woods and got a clear LOS on on a Tiger which was doing a good job of imdeding the Allied advance. I tried to call in a 155 barrage on the Tiger from a remote spotter. The shells fell everywhere but on the Tiger, like the spotter was firing area fire. My attacking force was cut to ribbons!! How? They both had radios. This was CMAK game. A3

[Snarker]

Two things to consider in your example arax3.

1) The company CO is not an FO, probably doesn't have the proper charts etc.

2) Some one will stagger along and confirm or renounce this - I wouldn't assume the CO could contact a FO on his radio in WWII. From what I've seen and read, radio setup and performance varied widely and wildly!

[Redwolf]

Well, a company CO could of course call in artillery support if he could convince the artillerymen that he knew what he was doing, was capable of reading the map correctly, was informed of other friendly troops in the area and has a target worth working at.

[arax3]

My dad was an artillery forward oberver in the late stages of WWII in Germany. He had a driver, a jeep and the appropriate maps. His job was to call in artillery strikes on enemy positions which he spotted. A3

[Michael Dorosh]

Originally posted by Snarker:

Two things to consider in your example arax3.

1) The company CO is not an FO, probably doesn't have the proper charts etc.

2) Some one will stagger along and confirm or renounce this - I wouldn't assume the CO could contact a FO on his radio in WWII. From what I've seen and read, radio setup and performance varied widely and wildly!

In the CW, the "arty rep" or "FOO" travelled with the OC (company commander) of an infantry company, and used his own equipment to contact the artillery chain of command. The infantry radios were netted for communication within the battalion. The two radios did not talk to each other.

Generally, the FOO was located with the company (or battalion) headquarters group for that reason.