Jump to content

Artillery Spotters and Night Vision


Recommended Posts

While playing WBW's Stolberg-Punching a Hole as the German's against the AI, I decided on the 2nd night to attack the US forces. I moved my 88mm Flak gun to the far rear for protection (in case any US units slipped through my attacking line) as I didn't think I would be using it for night action.

A few turns into the battle, shells start landing around the gun. I can't see any spotter or anyway the 88 could have been seen. So I try a little test and move the gun to the rear edge of the map and hide it between two tall buildings where the AI/US forces could not see it. 5 turns later. Shells start falling on the gun and knock it out. Alt U and can't find a US arty spotter anywhere on the map targeting the gun. What gives??? Invisible Arty Spotter with Night Vision??? No wonder the Germans lost.

PS. BTW WBW it is a great operation.

Link to comment
Share on other sites

Guest Captitalistdoginchina

Do you have FFoW turned on?

Otherwise i have no explanation!

CDIC

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

"Death solves all problems - no man no problem"

J.V.Stalin, 1918

Link to comment
Share on other sites

Guest Captitalistdoginchina

Hi,

Sorry i thought it was well known FFoW (Full Fog of War)

If you do not select FFoW before you start you and your opponent or the AI will know and see exactly where you are.

Default setting is On.

Hope this helps.

CDIC

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

"Death solves all problems - no man no problem"

J.V.Stalin, 1918

Link to comment
Share on other sites

Actually a handful of people have come on the board in the past describing exactly what Cos did, saying full FOW was on yet the AI still was able to target an out of the way gun. The last example I remember was from the Mortain op, no explanation was given as to why it happened. I guess the only recourse, if indeed FOW was on, would be to send the file(s) to BTS for them to have a look.

Link to comment
Share on other sites

It's not about FOV on or OFF. AI seems to allways know where your units are.

I was having Artillery bombardment at night in operation few hundred meters rear of my main defence. Targets were clearly 4 halftracks in middle of village. I hadn't yet seen any enemy units even get to my main line.

Link to comment
Share on other sites

Originally posted by illo:

It's not about FOV on or OFF. AI seems to allways know where your units are.

I was having Artillery bombardment at night in operation few hundred meters rear of my main defence. Targets were clearly 4 halftracks in middle of village. I hadn't yet seen any enemy units even get to my main line.

So the AI is cheating then. This is in essence what you're saying. Despite the fact that Charles has repeatedly explained that the AI doesn't cheat, and doesn't get to see

your units that it hasn't spotted.

I find it hillarious every time someone posts a story where they got whacked, and the only possible reason it could have happened was because the AI did something unfair.

Of course the AI couldn't possibly have targetted the village for bombardment for any other reason.....

Link to comment
Share on other sites

Originally posted by Captitalistdoginchina:

Hi,

Sorry i thought it was well known FFoW (Full Fog of War)

If you do not select FFoW before you start you and your opponent or the AI will know and see exactly where you are.

Default setting is On.

Hope this helps.

CDIC

FOW I would have got (I think) it was the 2 F's that got me confused. I didn't touch the FOW (as far as I know) when I started the game. I can't see the enemy units until they are visable under FFOW conditions when playing so I will assume it's still on.

Will look at things agian as I have kept the saved game to review once I got some ideas what might be going on from those of you more knowledge then me smile.gif

Originally posted by illo:

I was having Artillery bombardment at night in operation few hundred meters rear of my main defence. Targets were clearly 4 halftracks in middle of village. I hadn't yet seen any enemy units even get to my main line

I guess if your HT were moving around in the village there could have been sound contacts and the spotter could targeted in on that. If they weren't moving then nope, don't see how the AI should should know where they are.

I'll have to check on the 88 crew and see if they are having a really loud party of something like that or just taking a nap biggrin.gif

Link to comment
Share on other sites

Originally posted by Cos:

I'll have to check on the 88 crew and see if they are having a really loud party of something like that or just taking a nap biggrin.gif

Can you play it hot seat on the turn the fateful arty shows up? Maybe that will tell you who is doing the spotting if there is an area target line. Or it could tell you if the enemy can see your unit.

Link to comment
Share on other sites

Ok here's the update after checking it again. As far as I can tell FFOW is on. The AI spotter is in scattered trees and can not see the 88mm to target it (distance is 1137m at night and LOS is blocked by the buildings the 88 is hiding between).

Directly between the spotter and the town where the 88 is hiding is a hill with 3 Rifle sqds, HMG, and a Pz VIH shooting up a 60mm US motor and they have just eliminated a MG on the prior turn. The spotter is only 291m from the Pz VIH on the hill. So the AI doesn't go for this close threat, but chooses a gun it can't see. So I guess the AI can find guns at night anywhere on the map. Any idea/thoughts?

PS. The 88 crew were playing a game of cards and keeping the noise down as not to wake the people sleeping in town, so no sound contact possible.

Link to comment
Share on other sites

So, Phoenix here you see it. AI actually cheats...maybe only with arty, but it cheats.

Phoenix btw I didn't get whacked I got maybe few infantry casualties in that battle and thats all. I just wonder why AI decided to shoot halftracks in village which he didnt possibly see instead of my front troops.

I got all halftracks out when I saw first artillery rounds hit the ground.

It doesn´t bother me if AI cheats at least it gives more of an opponent that way.

Link to comment
Share on other sites

A couple questions: Were the shells falling on the 88 in the tight pattern you get with los, or the looser pattern you get without? Was the 88 in woods or in the open? Could any of his other units have possibly gotten los to the gun?

Had the ai seen the gun in the first battle of the operation? Was it shelled in the first battle?

I suspect the problem may be that the ai has trouble with forgetting it's seen things between battles in an operation...

Link to comment
Share on other sites

Originally posted by John Hough:

A couple questions: Were the shells falling on the 88 in the tight pattern you get with los, or the looser pattern you get without? Was the 88 in woods or in the open? Could any of his other units have possibly gotten los to the gun?

Had the ai seen the gun in the first battle of the operation? Was it shelled in the first battle?

I suspect the problem may be that the ai has trouble with forgetting it's seen things between battles in an operation...

Tight pattern like spotted. The 88 was between 2 tall buildings that blocked LOS to the AI lines (let alone all the trees and hills inbetween). None of the AI units at the beginning of this battle could see it.

The 88 was in action in the prior battle, I don't know if the AI arty had targeted on it. There are no shell craters in the area where the 88 was sitting in the prior battle.

I guess the AI could have targeted 88 in the prior battle and it carried over to the next one.

Link to comment
Share on other sites

Cos said:

While playing WBW's Stolberg-Punching a Hole as the German's against the AI, I decided on the 2nd night to attack the US forces. I moved my 88mm Flak gun to the far rear for protection (in case any US units slipped through my attacking line) as I didn't think I would be using it for night action.

I'm assuming you used this gun in the 1st battle, so the AI knew you had it and considered it an important threat. Thus, it would make an effort to destroy it.

Part of the AI functions on a higher level than the immediate tactical engagements. This makes the AI capable of using arty for other than direct tactical purposes, just as you and I can do. I think it's a reasonable assumption that enemy reserves often shelter in the rear, especially things that battlefield conditions make impotent at the moment, and I believe the AI can make a similar judgment. So just as I often shell rear areas on general suspicions, I think the AI does as well. At least it seems blindly to shell my rear areas sometimes.

Thus, even if you put the 88 back there in the set-up phase, there's no reason to assume the AI had to have known where it was. It could easily have just decided to shell there because it looked like a good hiding place for your reserves. Which IMHO is just realism I'm glad is in the game.

A few turns into the battle, shells start landing around the gun. I can't see any spotter or anyway the 88 could have been seen. So I try a little test and move the gun to the rear edge of the map and hide it between two tall buildings where the AI/US forces could not see it. 5 turns later. Shells start falling on the gun and knock it out.

Moving an 88 requires a vehicle, which usually creates a sound contact at night. Sound contacts don't show up exactly where the vehicle is, but the AI could easily figure out the true location because the shelling shows more or less where it started from. Hence, 5 turns later, more arty on the gun.

The 5 turns is critical proof here that the FO did NOT have an LOS. The "normal" time US arty FOs is 2 minutes, which is at least doubled when the FO lacks an LOS. Thus, this 5 minutes is exactly what you'd expect it to take to get the fire moved to your new location if the AI ordered that shift immediately upon hearing you move and you moved further than 100m from the initial impact area.

Tight pattern like spotted

Contrary to popular belief, LOS has NO EFFECT on the shape of the impact pattern. The only thing that changes pattern density is using the command Target Wide (about a 200m diameter circle) instead Target (about 100m x 50m ellipse oriented E-W). So if the AI's FO was using the Target command, its shells would have fallen in exactly the tight same pattern whether it had an LOS or not.

The only difference that no LOS causes is that every so often, 1 or more shells will fall WAY (up to several hundred meters) outside the normal pattern. This happens while the FFE is going on and can often pass unnoticed, especially if the stray rounds fall off the map due to shooting near an edge. But if you don't notice these strays, there's no way to tell whether the FO has an LOS or not just by looking at the impact pattern.

So anyway, I think your example here started out with bad luck. The AI decided to shell your rear area on general principles and happened to pick where you'd stashed your gun. When this fire flushed your gun out of hiding, the AI noticed and adjusted its fire. Simple as that. No "night vision", no cheating AI. Just realism not often found in an AI.

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

-Bullethead

In wine there is wisdom, in beer there is strength, in water there is bacteria.

[This message has been edited by Bullethead (edited 03-03-2001).]

Link to comment
Share on other sites

I played the Operation 21st Panzer Attacks a couple of days ago. FFOW was turned on.

Same Problem as Cos described. I placed my 88 in the rear and it was shelled with great accuracy (at night).

Of course it was knocked out, so i decided to load a saved game and place it in a more secure place. Guess what happened. Same story as before.

At this point, i decided to do a little test.

I placed the gun at the left and right mapedge, at the center in the rear and so on.

All tests showed the same result - the gun was knocked out, even when in terrain that cant be considered a good hiding place.

This leaves only one conclusion - the AI DOES know where your units are.

Link to comment
Share on other sites

So now I'm out of lurker mode here - if I'm following Mad Mike, he placed the gun in different places PRIOR (don't know how to underline here, sorry about the caps) to the first turn. So.. no vehicle noises in moving, etc. I'm curious what all of you who were so certain the Arty targeting was 'just luck' and so forth think now. I'm really a beginner here and know a lot of you have been on the board a long time, corresponded with BTW, etc. But - I'd sure like to hear a better explaination of this behaviour than what I read so far. Lots of unknowns still to consider I'm sure - but what???

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

For Great Justice

Link to comment
Share on other sites

Mad Mike said:

At this point, i decided to do a little test. I placed the gun at the left and right mapedge, at the center in the rear and so on. All tests showed the same result - the gun was knocked out, even when in terrain that cant be considered a good hiding place. This leaves only one conclusion - the AI DOES know where your units are.

Well, you got me there. This does seem a bit fishy. On the face of it, it's hard to reconcile these observations with BTS' repeated statements that the AI abides by the same FOW rules as we do. I don't see this happening in 1-off battles so I figure there is probably something different going on in an operation, perhaps unintentionally.

Let's see (just talking off the top of my head here).... In an operation, units that fought in a battle start the next battle in the same position they had at the end of the prior battle. But you can then reposition them inside the set-up area.

We know that more info about a unit than just its position also carries over into the next battle, such as casualties and ammo. And the AI's knowledge of the unit must also carry over to some extent to allow the AI to make plans with some sort of continuity.

Suppose that one piece of AI knowledge that carries over between battles is a unit's location if the AI had an LOS to that unit at the end of the prior battle. This knowledge retention seems almost to be required to prevent unrealistic, massive AI redeployments between battles. So assume that somewhere, the AI keeps a list of spotted units and their locations that carries over between battles.

Now assume at the start of a battle that the player redeploys a unit that's on the AI's list of known targets. This should take the unit's spotted status and its location off the AI's list, leaving only the mere fact that the unit exists. However, suppose there's a bug that doesn't remove this info from the AI's list, and in fact updates the unit's location with its new position after redeployment by the player.

In such a situation, the AI would know exactly where the target is even without an LOS from any of its individual units. And when the AI knows where something is, it tries to kill it, probably without asking itself how it knows the target's there. In these night cases, it can't use direct fire because that won't work without an LOS, but it can use arty. And the times reported for the arrival of the shelling indicate that the FOs don't have an LOS either.

So, here's a theory that fits all observed results: guns get shelled when they shouldn't and BTS hasn't deliberately made the AI cheat. What say ye?

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

-Bullethead

In wine there is wisdom, in beer there is strength, in water there is bacteria.

Link to comment
Share on other sites

I don't think BTS has deliberately designed the AI to cheat. I would say it is unintended and therefore a bug. I was convinced a long time ago that the AI knows where my valuable assets are. I did the same test that Mad Mike did back when this topic first came up. I was trying to do a 14" spotter test but my FOs were knocked out before their shells made it into the area. I made a map with a huge hill and heavy trees seperating the two forces, so there was absolutely no way the Germans saw my FOs. But no matter where I placed them they would get whacked within the first few turns. I know BTS has categorically denied that the AI cheats, but I am convined that it does in this situation.

Link to comment
Share on other sites

Originally posted by Bullethead:

Contrary to popular belief, LOS has NO EFFECT on the shape of the impact pattern. The only thing that changes pattern density is using the command Target Wide (about a 200m diameter circle) instead Target (about 100m x 50m ellipse oriented E-W).

Actually, you did experiments that verified that LOS does affect the size of the spread. This is consistent with the arty patterns that I see in current games, and why I prefer arty in LOS, rather than out of LOS. It's the gun size that has no effect on the pattern (except rockets, which just crash all over everything).

Here's a quote from your post that I cut out and put into my "useful articles from the forum" file. I don't have the original post pointer though.

Here are the average sizes of impact patterns under the possible states of the significant

variables:

Regular Impact Pattern Selected

LOS and TRP: 100m x 50m pattern oriented E-W

LOS, no TRP: ditto

Blind and TRP: ditto

Blind, no TRP: 200x x 100m pattern oriented E-W

Wide Impact Pattern Selected

LOS and TRP: 200m diam circle

LOS, no TRP: 250-300m diam circle

Blind and TRP: 200m diam circle

Blind, no TRP: 350-400m diam circle

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

"If you can taste the difference between caviar on a cracker and ketchup on a Kit-Kat while blindfolded, you have not had enough aquavit to be ready for lutefisk." (stolen from some web page about lutefisk)

[This message has been edited by chrisl (edited 03-04-2001).]

Link to comment
Share on other sites

Originally posted by Pvt. Ryan:

I made a map with a huge hill and heavy trees seperating the two forces, so there was absolutely no way the Germans saw my FOs. But no matter where I placed them they would get whacked within the first few turns. I know BTS has categorically denied that the AI cheats, but I am convined that it does in this situation.

I've done similar tests, but with about a dozen spotters hidden in buildings and woods against a dozen or more tanks controlled by the AI. I never lost a spotter to a tank that was more than 20 m away, in several runs of the test--virtually all of them got all their rounds off before getting popped, and very few of them got popped at all-- usually I ended up with a bunch of tanks milling around near the flags looking for things to shoot at. The arty also killed very few tanks, though it did generally button them up, and occasionally immobilized them or damaged the guns.

I've never seen any indication that the AI cheats-- the AI will punish you if you do something stupid, and will occasionally pull off something clever, like dropping arty where it thinks you would have forces (I do that too, but I prefer to drop a tight pattern on an identified target), but for the most part it's pretty predictable and easy to beat. I think you just had some bad luck.

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

"If you can taste the difference between caviar on a cracker and ketchup on a Kit-Kat while blindfolded, you have not had enough aquavit to be ready for lutefisk." (stolen from some web page about lutefisk)

Link to comment
Share on other sites

chrisl, I thought it was bad luck too. That's why I tried placing my FOs in different places - and in places you wouldn't normally hide them - so that the AI arty would not find them in the traditional hiding places. No matter where I placed my FOs the AI found them right away. Incidentally, my test was with FOs, but the same thing happens with AT guns and other soft valuable assets. I don't see the AI cheating in other situations - only when calling arty down on me.

Link to comment
Share on other sites

Guest Watson & Crick

I think I know the answer. In the previous battle your gun was spotted by the AI. This carries over to the next battle. So regardless of where you place the gun, the Ai will know where it is.

The converse also has happened to me. As an attacker, you spot a gun in a battle and that battle ends before you get a chance to eliminate the gun. If you pay close attention and look around the map in the next battle, you should be able to find the "pre-spotted" gun. I think that this only relates to guns and not tanks etc. IIRC there has been at least one thread regarding this problem. IIRC someone from BTS says that there was no fix.

Link to comment
Share on other sites

Originally posted by Bullethead:

Mad Mike said:

Let's see (just talking off the top of my head here).... In an operation, units that fought in a battle start the next battle in the same position they had at the end of the prior battle. But you can then reposition them inside the set-up area.

We know that more info about a unit than just its position also carries over into the next battle, such as casualties and ammo. And the AI's knowledge of the unit must also carry over to some extent to allow the AI to make plans with some sort of continuity.

Suppose that one piece of AI knowledge that carries over between battles is a unit's location if the AI had an LOS to that unit at the end of the prior battle. This knowledge retention seems almost to be required to prevent unrealistic, massive AI redeployments between battles. So assume that somewhere, the AI keeps a list of spotted units and their locations that carries over between battles.

Now assume at the start of a battle that the player redeploys a unit that's on the AI's list of known targets. This should take the unit's spotted status and its location off the AI's list, leaving only the mere fact that the unit exists. However, suppose there's a bug that doesn't remove this info from the AI's list, and in fact updates the unit's location with its new position after redeployment by the player.

In such a situation, the AI would know exactly where the target is even without an LOS from any of its individual units. And when the AI knows where something is, it tries to kill it, probably without asking itself how it knows the target's there. In these night cases, it can't use direct fire because that won't work without an LOS, but it can use arty. And the times reported for the arrival of the shelling indicate that the FOs don't have an LOS either.

So, here's a theory that fits all observed results: guns get shelled when they shouldn't and BTS hasn't deliberately made the AI cheat. What say ye?

I agree with that. My 88 was moved during set-up phase to the rear. It had been firing on infantry in the prior battle and stopped the AI advance and was high on the threat list from that battle and easly could have been targeted by the AI spotter, but since it was in the last 3 turns of that battle, shells hadn't started to fall yet.

Link to comment
Share on other sites

OK, some good thinking going on here. I'm impressed with Bulletheads analysis, esp. since later posts seem to confirm he is on to something. Now that we have a working hypotheses, (the AI can 'follow' previously IDd resources, even if moved between battles), how do we get BTS to think about that possibility, or do a check? Or should we do some more searching to see if they've answered this already? I've searched some, but didn't find - but people can get really crabby around here bout that sort of thing smile.gif

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

For Great Justice

Link to comment
Share on other sites

×
×
  • Create New...