Spotting in game version 1.06

[SlowMotion]

I made small test to see how units spot enemy units. Does it work like a player might think: for example if there are some units straight ahead, they are all spotted quickly and at the same time?

I moved a tank to a position from which it should "see" several destroyed enemy tanks, if a tank crew member just looked at this direction. So I was expecting that it would take a while for someone in the tank to scan the area around it and once looking to right direction then all the four tanks would be spotted.

What actually happened was that after a moment tanks 1 and 4 were spotted almost simultaneously. Then after maybe 10 seconds #2 and then soon after this #3.

[Redwolf]

10 seconds to see a tank-sized object in that distance, when you are a soldier who's life depends on spotting the enemy first?

What's the distance there?

[SlowMotion]

I tried this again and tanks were about 200-300 m away. Now it took more than 15 seconds to see all 4 tanks.

I set enemy tanks as destroyed because otherwise some of them would have destroyed my tank before it had even seen all those 4. I think this is quite surprising. Distance is quite short and there are no trees etc blocking the view.

[SlowMotion]

I put those tanks closer and tank crew veteran. Here the tanks are 250-270m away. I ran this test several times and saw that the command you use when moving has a big effect on how fast spotting works. When using Hunt, from spotting of first tank to last it took 20 seconds.

[Kwazydog]

Slowmotion, out of interest could you retry this test with the commander unbuttoned, those old Soviet tanks are notoriously blind when the commander is buttoned up.

Its not impossible some more tweaking does need to be done here, but overall Ive seen pretty good and realistive results in the couple of PBEMs Ive had time to play recently.

[Guest Guest]

deleted per user request

[SlowMotion]

Tested with unbuttoned commander and those 4 tanks within 20 meters. Took this same about 20 seconds again. Maybe a bit more. Then compared this old tank to T-72 (2001). Unbuttoned time was about the same, but *buttoned* time was actually 5 seconds shorter.

Also tried it with UAZ vehicle with veteran FO team onboard. So three guys watching, all "unbuttoned". From first to last it took about 11 seconds. If this same team had dismounted first and walked (Move command) to position where they could see the tanks it took the same 11 seconds.

And same time with bailed out 4 man tank crew.

I don't know how spotting works now, but as player I assumed that things that are the same size, equally visible and close to each other would be spotted almost simultaneously. That's how I see things if I rotate the map.

The fact that things don't work this way surely changes the way I play. It seems one needs to stop frequently and let this spotting process to happen before deciding what to do next.

I like 1.06, you just need to play it differently than old CMx1 games. How spotting works is clearly one of those things that change the game play. C2 is another.

[ February 17, 2008, 02:43 AM: Message edited by: SlowMotion ]

[Battlefront.com]

There are a bunch of factors that go into spotting, so it is far more than distance and obstructions (or lack of them). For one, spotting from vehicles is heavily dependent on the vehicle being used. By and large buttoned Soviet equipment has rather poor situational awareness. When on the move spotting becomes even more difficult.

On top of this, it is presumed that when something is spotted the unit temporarily (for a few seconds at least) gets tunnecould be looking at something inside the vehicle, checking footing while walking, etc. A few seconds to notice something isn't unreasonable in real life.

Oh, and number of eyeballs on the target is critically important. This is another reason why vehicles tend to have more problems spotting things because generally there are only 2-3 sets compared to a Squad's 8-9.

Steve

[Redwolf]

The gunner's sight is facing the enemy tanks, directly. How can he possibly not see them within a few seconds?

[Battlefront.com]

See previous comments. Unless a gunner has his eyes screwed into the sight 100% of the time there is a chance he won't see something instantly because he isn't looking.

It also has to be taken into consideration that this is a completely artificial situation that SlowMotion has created. In real life tanks don't magically appear 200-300m out of thin air directly in front of you. So it really doesn't matter what happens in such situations because it won't happen in a real game and obviously not in real life either.

I'll remind people that we've been through tests like this a million times before with CMx1 games. There are all sorts of artificial things people can do in the Editor that yield strange results, but in the game itself don't appear. Since the game is what matters, we don't worry too much about such tests

Steve

[Roter Stern]

Originally posted by Battlefront.com:

It also has to be taken into consideration that this is a completely artificial situation that SlowMotion has created. In real life tanks don't magically appear 200-300m out of thin air directly in front of you. So it really doesn't matter what happens in such situations because it won't happen in a real game and obviously not in real life either.

I'm sorry, but that's a ridiculous thing to say In "real life" people don't go around calculating the square root of 225 on a daily basis either, but when you punch it into a calculator for a test it better give you an answer of 15 or you might as well get a new calculator...

Besides, how is "appearing out of thin air" different than "appearing from behind a wall", as far as the simulation is concerned?

[ February 17, 2008, 06:54 PM: Message edited by: The Louch ]

[Battlefront.com]

A simulation is nothing more than a bunch of math that is tossed together in the hopes that the results generally come up right rather than wrong. It is imperfect at best, so it isn't at all surprising to us to find that some situations crop up where things don't work as expected. The CMx1 Forums are chock full of things that didn't work correctly in "artificial" circumstances, yet nobody appeared to see them in the game itself. Since the game itself is really the only thing that matters we have no desire to chase things that don't appear to be a problem in the game.

In theory, the "appearing from behind a wall" could be the same as "appearing out of thin air", but it is not likely to be so. That's the point I'm trying to make. That wall example you came up with would be in the context of an actual battle where tons of other variables are at work. Having an absolutely flat map, with one unit on one side and four on the other, with a tiny amount of distance between them, with the one unit perfectly facing the four... completely artificial. A unit popping out from behind a wall is not in the same ballpark

Steve

[Roter Stern]

Well, of course I couldn't just leave it there, so here's my "Real Life Appearing Out of the Thin Air Test"...

Bunch of vehicles get positioned behind a tall wall with a 16m wide opening in it. On the other side a bunch of AI's tanks and things are ordered to drive back and forth at full speed in front of the opening. Range between vehicles is about 180m, with the wall about half way.

Scene 1 - a split second after the first tank peeked though the opening - four out of six vehicles have the tank spotted.

Scene 2 - a few seconds after the first - can't really see the spotting info on this one, although it's pretty definitive that not only was the T-72 able to spot the two tanks, but it actually had enough time to engage both as they moved in front of the opening. Also notice that the BRDM2 has already scurried out of the harms way.

What does that prove? I have no idea. edit: actually I do know... it proves that Steve was right /gasp

edit 2: after running it though a bunch of times had a BMP-2 take out a T-55 with a single burst of the 30mm - must've hit the ammo compartment, since the whole thing went up in the biggest fireball I've seen in a while!

[ February 17, 2008, 07:47 PM: Message edited by: The Louch ]

[SlowMotion]

Yes it was artificial situation and on purpose. I tried to see how spotting works in CMx2 and now I understand it a bit better. I was wondering if spotting would work so that units would scan the map around them in chunks, let's say 6 x 60 degrees. All units inside one such field of view would be spotted at the same time. Then unit would look into other direction and spot whatever is there.

After these tests I saw it doesn't work this way. I got some idea of spotting times and the important new thing to me that it may take a while even if the units are in same direction.

My test was just a test, but a similar game/RL situation could be when you first have a unit behind a hill and you want to see what's on the other side. How long should my unit stay in a spotting place in order to see the possible enemy units? I have this kind of situation in almost all games I play.

Another situation could be Shoot & Scoot. In CMx1 a unit could target enemy units it cannot see, so individual unit's spotting time didn't matter much. Now you have to move your unit to Shoot position, wait for it to spot and shoot, then reverse.

The Louch: I wonder how would it work if those red units behind the wall were static and you moved your T-72 forward so that it could see the red units from the opening? I would think that moving tanks are spotted much faster.

And Steve: of course that the game works is what really counts. But I think that now when people are starting to really play CMSF (my guess) at least some of them want to get a better idea of some things like this that could be important, especially when playing real time mode games. In CMx1 people tried to estimate things like turret rotation speed or how quickly a gun can fire. Now some other things may be important.

In my tests some tanks weren't spotting any faster when the commander was unbuttoned. In CMx1 such tanks spotted clearly faster (during real game situations, not tests). Now it looks like it doesn't always work this way.

[ February 17, 2008, 10:40 PM: Message edited by: SlowMotion ]

[Redwolf]

Originally posted by Battlefront.com:

See previous comments. Unless a gunner has his eyes screwed into the sight 100% of the time there is a chance he won't see something instantly because he isn't looking.

I don't think that a tank commander moving into combat will allow his gunner to fool around with his gameboy all day. The gunner isn't looking as their tank is coming over a ridge line, that will take a while to convince me of.

Also, since two tanks are spotted and two are not, but all four are in the same place, there's no direct explanation along those lines.

I know it's abstracted, but it seems heavily pessimised.

[Taki]

Another situation could be Shoot & Scoot. In CMx1 a unit could target enemy units it cannot see, so individual unit's spotting time didn't matter much. Now you have to move your unit to Shoot position, wait for it to spot and shoot, then reverse.

You cAnt Shoot and Scoot in CMx2 anymore when you have the underdog Tank. Its as simple as it is. Spotting System exspecially with Tanks is weird sometimes and not only in Testing Situations. I talk of In Game behaviour in MP TPC/IP (and must be the same in Single Player)

Reason to me is: there isnt so much simulated. As Redwolf mentioned there should be better Spotting Times with Firearcs and units that has been spotted by other Friendly Units that can inform you where it is so you have a Clue where you have to search.

CMSF in its actual state is a Dice Roll with the bitter end for the weaker Tank almost all of the Time. You cant Shoot and Scoot, you cant flank. Its just to unpredictable and things that Logically would work in RL (like a Commander Unbuttoned comming over a Ridge knowing the Exact pos of an enemy M1 and just dont find it, or find it Plings with MG Fire and the M1 Turns and Kills you) just dont work.

[Battlefront.com]

SlowMotion,

Yes it was artificial situation and on purpose. I tried to see how spotting works in CMx2 and now I understand it a bit better. I was wondering if spotting would work so that units would scan the map around them in chunks, let's say 6 x 60 degrees. All units inside one such field of view would be spotted at the same time. Then unit would look into other direction and spot whatever is there.

The thing is that just about every unit has more than one set of eyes and in real life scanning doesn't work that way. People don't tend to "scan" very much because it's energy intensive and extremely distracting too.

My test was just a test, but a similar game/RL situation could be when you first have a unit behind a hill and you want to see what's on the other side. How long should my unit stay in a spotting place in order to see the possible enemy units? I have this kind of situation in almost all games I play.

Right, but remember that in the artificial scenario you set up there is no variability. The two bunches of units always start out in the same spot in full view with the same variables being used and not being used. When in the game spotting happens far more fluidly and dynamically because things are more fluid and dynamic You can spot something within 1 second of moving into view of it. You are correct, though, that the artificial test you came up with does give an idea of a sort of worst case scenario for that particular tank.

Another situation could be Shoot & Scoot. In CMx1 a unit could target enemy units it cannot see, so individual unit's spotting time didn't matter much. Now you have to move your unit to Shoot position, wait for it to spot and shoot, then reverse.

Er... got me confused here. In CMx2 you can not only tell a unit to target something it can't see, but it can track that target (depending on tactical circumstances) until it is seen and then fire at it immediately. In CMx1 the TacAI had no "memory" so when something went out of LOS it was "out of sight, out of mind". In CMx2 the TacAI has a memory, however if another target comes into LOS it may switch from the unseen target to the seen one, depending on circumstances.

And Steve: of course that the game works is what really counts. But I think that now when people are starting to really play CMSF (my guess) at least some of them want to get a better idea of some things like this that could be important, especially when playing real time mode games. In CMx1 people tried to estimate things like turret rotation speed or how quickly a gun can fire. Now some other things may be important.

Sure, perfectly understandable. Personally, I play more intuitively and don't bother trying to figure out stuff like this because in most game situations it won't be of much use due to the vast number of possibilities that can be present in any given second. For example, I wouldn't try to sneak a tank around the corner of a wall to hopefully get first shot off at a gaggle of enemy vehicles. In real life that would be a bad idea So instead I'd wait for them to come to me or I'd try to flush them out somehow. Or, if possible, I'd try to work around them and come in from a different direction. That sort of thing. Trying to guess how many seconds this or that thing will allow for is not value IMHO. But each person plays the way they want to so I'm not sitting in judgement here.

In my tests some tanks weren't spotting any faster when the commander was unbuttoned. In CMx1 such tanks spotted clearly faster (during real game situations, not tests). Now it looks like it doesn't always work this way.

In the game it probably matters more than it does in artificial situations. At least it should. I recall fairly in-depth testing of this and (depending on vehicle, of course) it did make quite a bit of difference. For something like an Abrams M1A2 SEP the difference wouldn't be much, if any, for example.

Redwolf,

I don't think that a tank commander moving into combat will allow his gunner to fool around with his gameboy all day. The gunner isn't looking as their tank is coming over a ridge line, that will take a while to convince me of.

Ask anybody that is in the military, especially one that is in actual combat, and the first thing they'll tell you is whatever SHOULD happen doesn't. It is impossible, and I mean physically and psychologically impossible, to be 100% vigilant 100% of the time. Reality is that it's far less than that, especially considering that most soldiers are tired most of the time, usually exhausted. Plus, all soldiers or crew members have more than one job to do. The tank gunner, for example, doesn't have the luxury of doing nothing but sit there with his eyes screwed into the sights. In fact, it might be physically difficult to do at times with a vehicle that doesn't have an independently stabilized gunnery system. So on and so forth.

Now, it is true that if the tank commander said "we're about to go over this ridge here, all eyes forward and fire at the first thing you see" that the gunner's eyes would be screwed into the sights. However, CM has no concept of such things, not now and not ever. The reason why is because this is impossible for the TacAI to determine on its own and allowing the player to micromanage this would be cumbersome and likely unrealistic overall (due to God View intel gathering).

Also, since two tanks are spotted and two are not, but all four are in the same place, there's no direct explanation along those lines.

The explanation is simple... CM doesn't have the ability to group spot. Therefore, every unit must be spotted individually. Usually, in a real game situation, this isn't a problem since the sort of clustering in SlowMotion's artificial example doesn't happen much, if at all. And if it does it usually is resolved in such a way that it doesn't matter. It's not like the spotting tank is physically capable of engaging all four targets simultaneously, therefore by the time it is physically capable of engaging targets 3 and 4 it's already spotted them. Or, as is more likely the case in a real game, someone's popped smoke, the spotting tank has been KO'd, someone's reversed out of LOS, etc.

Again, the artificial tests have some useful purposes, but as we've said for 10 years now... they have their limitations. Getting too focused on what an artificial test yields can be counter productive.

Taki,

As Redwolf mentioned there should be better Spotting Times with Firearcs and units that has been spotted by other Friendly Units that can inform you where it is so you have a Clue where you have to search.

Cover Arcs should give an advantage to spotting and with Relative Spotting you absolutely can know where the enemy is before physically seeing it. I'm not sure why you think this isn't the case. Certainly SlowMotion's test doesn't even attempt to demonstrate either of these things.

CMSF in its actual state is a Dice Roll with the bitter end for the weaker Tank almost all of the Time. You cant Shoot and Scoot, you cant flank. Its just to unpredictable and things that Logically would work in RL (like a Commander Unbuttoned comming over a Ridge knowing the Exact pos of an enemy M1 and just dont find it, or find it Plings with MG Fire and the M1 Turns and Kills you) just dont work.

Well, I think you're exaggerating just a little (OK, a lot ). You can do these things you say are impossible, though I do not denny that sometimes they don't work or they aren't 100% predictable 100% of the time. That's done on purpose. Real life isn't a bunch of static, narrowly focused, perfect cased events that are all nice and logical. Things in real life are prone to not working they way they should work on paper. If we optimized the game for the best case stuff the entire game would feel artificial and wrong. Robotic behavior is simply not realistic, neither is optimal behavior. Yes, this can be frustrating for some types of gamers, but that's a different story

Steve

[SlowMotion]

Originally posted by Battlefront.com:

The thing is that just about every unit has more than one set of eyes and in real life scanning doesn't work that way. People don't tend to "scan" very much because it's energy intensive and extremely distracting too.

AFAIK WW2 fighter pilots used this scanning technique and it saved many lives.

Er... got me confused here. In CMx2 you can not only tell a unit to target something it can't see, but it can track that target (depending on tactical circumstances) until it is seen and then fire at it immediately. In CMx1 the TacAI had no "memory" so when something went out of LOS it was "out of sight, out of mind". In CMx2 the TacAI has a memory, however if another target comes into LOS it may switch from the unseen target to the seen one, depending on circumstances.

I know about the memory (unit icon changes into '?' icon. But targeting units that my unit hasn't seen yet? If I try targeting those ?-icons that a friendly unit has spotted, I see text like "no line of sight" or something. I don't understand how targeting those would be possible, like it was in Borg spotting.

Cover Arcs should give an advantage to spotting and with Relative Spotting you absolutely can know where the enemy is before physically seeing it. I'm not sure why you think this isn't the case. Certainly SlowMotion's test doesn't even attempt to demonstrate either of these things.

Like you wrote earlier, I thought there would be many things involved in spotting. I didn't try to guess all of them, just did the very basic tests. But now that this was mentioned, I did test and it did shorten times. Another thing I noticed making these times worse was the smoke cloud produced by the tank. Once a moving tank stops it takes a while for this smoke cloud to clear.

[flamingknives]

If WW2 fighter pilot "scanning" was noteworthy, doesn't that indicate that people, ordinarily, do not "scan" very much? Possibly because it is energy intensive and distracting?

[Drusus]

As I see it, there is a specific situation where the game doesn't work as I would expect it to do. That is when the unit in question has (or should have) information about enemy units because of relayed information.

There are two things I don't see happening as should. First, the relaying of information isn't what I would expect it to be. Or maybe the information passing does happen, but the game doesn't show it as well as it should.

The second problem is that spotting in situations where information is (or should be) relayed isn't as fast as I would expect it to be. This might be because the information isn't relayed, or because the spotting is just too slow in these cases.

The typical situation is having some US infantry units spot enemy vehicles. If you want to shoot & scoot those spotted vehicles with your own unit, there is a big risk of bad results.

My example is from the Jisrah Shugur scenario. Here the enemy's vehicle is usually BMP-2 and your vehicle is Stryker MGS. Naturally, in this situation spotting first is critical.

Now, the problem is that even if my infantry units have known for a long time (5 minutes or more) where the enemy BMP is (which is usually on road or in completely open ground), the information doesn't seem to help in the spotting of the BMP when trying to kill it with the MGS. Giving covered arcs doesn't help either. The range is usually below 500m. And the time it takes to spot the enemy vehicle is often more than 20 seconds.

As I see it, there should be little chance of such long spotting time given that the MGS knows pretty well where the vehicle is and thus should be actively spotting the right area.

The solution is simple. Spotting of units whose information has been relayed to the spotting unit should be much faster. Also, the player should have some more knowledge of the information relayed to the unit.

Here I am assuming that such information is relayed when in battle. This is where the US excel, so my assumption shouldn't be wrong.

I have said this earlier, but there is some gamey ways to solve this situation. The first is that you as the player can see when the BMP starts rotating it's turret and issue a reverse command immediately. Works even better when trying to flank MGS with BMP. The other method I have used when I get bored is to use area fire against the BMP. For both problems a simple solution is to have a small pause before the execution of the command.

[SlowMotion]

Originally posted by flamingknives:

If WW2 fighter pilot "scanning" was noteworthy, doesn't that indicate that people, ordinarily, do not "scan" very much? Possibly because it is energy intensive and distracting?

I didn't mean to imply that it wasn't energy intensive. I got my idea of how it might be done in tanks/armored vehicles from WW2 air battles. If you know how this spotting is really done *when encountering enemy is expected*, please let me know.

[Redwolf]

I don't see why people drag that to an inter-vehicle communication issue. Two tanks a couple hundred meters away directly ahead that the gunner doesn't see with turret straight ahead? Not possible.

[thewood]

The one thing I still see few people bringing up is that when spotted, the enemy unit becomes ID'ed in detail immediately. I still don't see how you can tell at 4-500m whether its a regular infantry unit, an HQ, an ATGM unit, etc. That's not in the open, but loaded in an APC. Same with tanks. From 1000m you right away know exactly what model T55 is out there.

[flamingknives]

FM 3-22.1, Chapter 6, sections 1 to 3.

[Lethaface]

Originally posted by Redwolf:

I don't see why people drag that to an inter-vehicle communication issue. Two tanks a couple hundred meters away directly ahead that the gunner doesn't see with turret straight ahead? Not possible.

That is a Real Life (how unkown it might be to some of us ) situation.

What Steve argues is that the visibility system isn't build upon those situations. It isnt build for group spotting. To tune that down a bit; Mr. Redwolf walks in a forest. Right in front of him, @25,4 meters, is a tree. 0,7 meters to the right is another tree. Luckily for Mr. Redwolf his eyes allow him to spot both trees at the same time.

As I understand CMSF spotting processes make it so that any given unit has to spot any other given unit one for one. That might even produce results that unit X can spot unit Y, while unit Z is partially blocking view to unit Y. Still, unit X only spots unit Y but can't spot unit Z. In Real Life offcourse this wouldn't be possible. But there are a range of factors why it is possible ingame. Perhaps it SHOULDN'T be possible. But hey, there are more things in that category. AI 'eyes' dont operate like our ones. Hopefully in the future group spotting can be implemented in the game; or at least all spotting routines will be refined over time so they will produce more life like results. I'm afraid that we will never have Real Life alike spotting in any game (that is excluding CM SupaComputah which Charles and Steve are working on for me every hour they're not trying to convince you of their lies).

Since 1.06 im quite happy with spotting. I did have some strange things, like in my example above, one of my RPG squads couldnt see tank 1 that was actually blocking LOF to tank 2. They did see tank 2 and fired on it, but instead kept hitting tank 1. Actually, only tank 1 was at a disadvantage there, since it got blown to pieces before tank 3 shot up my RPG team. I'll never find out if my team would have spotted tank 1 or tank 3 first. But; who cares

PS Redwolf, no personal offense