Casualties Wounded to dead

[womble]

It's a simulation, so it is a model that approximates to reality. I would suggest that factors associated...having no secure line of retreat...

And there you have touched on something that the current AI has absolutely no idea how to even begin thinking about considering how to determine. Not one. "Retreat" according to the AI (i.e. once it gets off the human-generated AI plan) is "towards a map edge marked as friendly" by the most direct route, so even if there's a covered route, it'd better be straight. And given the state of the surrender modelling at the moment (with headless chicken squads running back and forth in the open under fire from tanks and distant HMGs but still not surrendering) it's debatable as to whether adding any more factors is a priority.

[SmittyJ]

I was also struck by the skewed KIA/WIA ratio- until I realized that I wasn't doing anything special to render aid to my wounded. Has anybody made it SOP to use small groups as medics? Has anybody seen it make the difference between victory and defeat in a tight game? I think I'll try using my HQ support guys as full time buddy aiders and see if the results are worth the effort.

[Apocal]

Yup, we don't want to change the control and flexibility scenario designers haven and instead impose upon them draconian, hyper realistic conditions which would remove diversity and overall realism. It is our call and I'm confident that the vast majority of CM customers, hardcore grogs and gamers alike, are supportive in the way we have approached things.

There is nothing realistic about being able to push dudes consistently past 50% casualties and still have them actually obey orders. In-game right now, there is nothing I can do to force a realistic morale model on the player, even at -2 motivation, they just rally too quick and in far worse situations. Meanwhile, even neutral morale (to say nothing of +3 or +4) gives me the ability to bang infantry together even in broken state.

Would it be reasonable if I asked for something like -3 or -4 morale states?

[JonS]

Has anybody made it SOP to use small groups as medics?

No, but I do go out of my way to buddy aid as much as possible. I don't use dedicated 'clean up crews' because generally I want any ammo and wpns that are recovered to remain with the team or section that took the loss.

Has anybody seen it make the difference between victory and defeat in a tight game?

Nope, I haven't seen that. But that's not why I do it anyway.

[Erwin]

"Has anybody made it SOP to use small groups as medics?" Many of us use the plethora of HQ units as medics: HQ Support units; XOs; 2ICs etc.

Just curious, when a unit does take any casualties, why not make em unavailable for orders for some random amount of time (depending on their morale and experience level I suppose) before they can be used again, per what Steve was saying - since as he pointed out we tend to keep em going while in RL they would be out of action for quite a while?

Or is that a game/playability issue?

Re the idea of shooting and hiding, the idea would be good esp in WEGO as currently we can only have one or the other for a whole minute. It would be great to be able to order a FIRE BRIEFLY accompanied with a HIDE.

The issue re ambushing is also problematic in WEGO. If you HIDE behind a wall (even with a covered arc), the enemy will usually walk up to the wall and shoot you. If you don't hide and are behind a wall, the enemy usually can see you before any ambush can be sprung. And when the enemy fires at you behind your wall, all your men can be hit - even though one would expect that only one guy would be exposed while the others are covered by the wall.

It would be great to be able to have an AMBUSH command which is effectively a HIDE in which one guy would poke his head up a few times per minute to check around and if suitable spring the ambush.

[JonS]

Just curious, when a unit does take any casualties, why not make em unavailable for orders for some random amount of time (depending on their morale and experience level I suppose) before they can be used again, per what Steve was saying - since as he pointed out we tend to keep em going while in RL they would be out of action for quite a while?

Or is that a game/playability issue?

That's already in, but it depends on the circumstances.

It would be great to be able to have an AMBUSH command which is effectively a HIDE in which one guy would poke his head up a few times per minute to check around and if suitable spring the ambush.

Isn't that HIDE + COVERED ARC ?

[Childress]

Nope, I haven't seen that. But that's not why I do it anyway.

Ahh, that's touching, Jon.

[Battlefront.com]

There is nothing realistic about being able to push dudes consistently past 50% casualties and still have them actually obey orders.

True, but how is the game supposed to know if you're moving them with combat in mind or retreat in mind? It can't differentiate from that. Which leaves the game with two horribly difficult things to determine:

1. When to wrest control from the player and submit the units to AI control.

2. How to program an AI that is reasonably capable of handling your units in a way that doesn't make you scream "NO YOU MORONS!!! THAT'S WHERE THE ENEMY IS!!!" (even if in real life they wouldn't know).

In-game right now, there is nothing I can do to force a realistic morale model on the player, even at -2 motivation, they just rally too quick and in far worse situations. Meanwhile, even neutral morale (to say nothing of +3 or +4) gives me the ability to bang infantry together even in broken state.

Would it be reasonable if I asked for something like -3 or -4 morale states?

This is definitely something we need to look at again as we head back to the Eastern Front. It could be that some reduction in recovery or other tweaks might make some positive differences to how the game plays.

Here's a way to think of it. The reason Combat Mission is so enjoyable, so realistic feeling, and so diverse compared to other wargames is because of it's TacAI. The reason Combat Mission can at times be frustrating, unrealistic, and seemingly narrow minded is also because of it's TacAI.

We're fortunate to have a game engine that largely does the former and more rarely does the latter. Most games have it the other way around.

Not to say we won't strive for improvements, because we will. I just feel (as always) that perspective should be kept reasonable and relevant to reality.

Steve

[Erwin]

"Isn't that HIDE + COVERED ARC?"

Jon: I think we've determined that HIDE + covered arc is a very dangerous/unsuccessful method of attempting an ambush as most times, your troops will be "lying with their faces in the dirt" while the enemy literally walks up and shoots em in the back.

It's my understanding that when ordered to HIDE, no one is keeping any sort of a "look-out".

Conversely, as I mentioned, if your troops are behind a wall (for example) and NOT hiding, the enemy can often see em first before an ambush can be sprung. And when the enemy fires, all one's troops are vulnerable as if they are all looking over the wall, as opposed to only one guy.

So, what people are asking for is an intermediate command where your troops will be HIDING waiting in ambush, but one guy is looking around every (say) ten seconds or so, in order to spring the ambush (and ensure that the enemy doesn't just walk over and shoot everyone in the back while they are HIDING).

[womble]

It would be great to be able to have an AMBUSH command which is effectively a HIDE in which one guy would poke his head up a few times per minute to check around and if suitable spring the ambush.

Watch your HIDEing units next time. They do periodically have someone SPOTTING rather than HIDING. It's just a lot less than all the eyes all the time, so HIDEing units get surprised/contribute less to the overall intel picture than non-HIDEing troops.

The difference would be the last bit. I think one of the problems with HIDEing troops is that it takes a more clear and present threat to make them ignore their HIDE orders and come up shooting.

[Battlefront.com]

And there you have touched on something that the current AI has absolutely no idea how to even begin thinking about considering how to determine.

Yup. This is the sort of programming that takes a long time to implement, and much longer to sort out. Extremely "twitchy" behavior is guarranteed. And that is the thing that sucks so much about AI programming. An AI function in CM that gets a 90% success rate might still be viewed as a failure. Problem for us is that 10% could take 3-4 times longer to dent than it took to make the 90%.

And given the state of the surrender modelling at the moment (with headless chicken squads running back and forth in the open under fire from tanks and distant HMGs but still not surrendering) it's debatable as to whether adding any more factors is a priority.

I think people who are fondly remembering CMx1's surrender and routing model might want to see how the above statement applies

"Has anybody made it SOP to use small groups as medics?" Many of us use the plethora of HQ units as medics: HQ Support units; XOs; 2ICs etc.

I've seen loads of people use HQs, bailed out vehicle crews, ammo teams without ammo, totally spent Rifle Squads, etc. for just this sort of thing. I know I have.

Just curious, when a unit does take any casualties, why not make em unavailable for orders for some random amount of time (depending on their morale and experience level I suppose) before they can be used again,...

This does happen, but not strictly because of casualties.

The problem here involves both the points I raised in the previous post (when to cut the cord with the user, then how to handle the units under the AI) plus a new one. And that new one is...

Or is that a game/playability issue?

The more we wrest control from the player, the less he's playing and the more he's watching. Anybody that thinks this doesn't exist, or that we need not be concerned about it, is almost certainly not employed in a position of creative responsibility within a games company. If by some strange chance someone is, I will append my previous statement with "not for long" at the end

Re the idea of shooting and hiding, the idea would be good esp in WEGO as currently we can only have one or the other for a whole minute. It would be great to be able to order a FIRE BRIEFLY accompanied with a HIDE.

It would certainly be useful, but here's the problem. How does CM know you want to HIDE *after* instead of *before*? And I for one would not like "Target Briefly" to assume I want my guys to do both 100% of the time (which is the only other option).

This is why I think we need a dedicated Ambush Command. Hide would be assumed as the "before" and therefore an explicit Hide could be interpreted as an option "after" request.

Steve

[JonS]

Jon: I think we've determined that HIDE + covered arc is a very dangerous/unsuccessful method of attempting an ambush

Ah. Did we? Ok. I'll have to apologise to all my opponents, then, for all those successful ambushes. Must have been a bug or sumfink.

It's my understanding that when ordered to HIDE, no one is keeping any sort of a "look-out".

Your understanding is wrong.

[akd]

Depends on location. If behind a wall/hedge/bocage, no one is looking to the far side if HIDING.

[Apocal]

True, but how is the game supposed to know if you're moving them with combat in mind or retreat in mind? It can't differentiate from that. Which leaves the game with two horribly difficult things to determine:

1. When to wrest control from the player and submit the units to AI control.

2. How to program an AI that is reasonably capable of handling your units in a way that doesn't make you scream "NO YOU MORONS!!! THAT'S WHERE THE ENEMY IS!!!" (even if in real life they wouldn't know).

I'm obviously not familiar with your code, but you already have something that approximates No. 2 reasonably well in the form of "friendly directions." Obviously there are times when it can be "gamed" or otherwise produce unrealistic results.

Not to say we won't strive for improvements, because we will. I just feel (as always) that perspective should be kept reasonable and relevant to reality.

Just so I'm clear, I actually do play CMx2, enjoy it and feel it offers something different than competitors. But I also want it to improve.

[Michael Emrys]

In CM how many players cancel their orders for a Squad that's taken 3 WIA casualties?

Not precisely that, but in one game I had a platoon get hit by a couple of mortar rounds as they were moving out of the setup zone. Two of the squads lost a two or three men each and the morale of all of them plummeted. From then on that platoon was not used for any kind of heavy action. They trailed well behind the rest of the company and were used solely to occupy objectives that the rest of the force had captured.

Michael

[JonS]

Depends on location. If behind a wall/hedge/bocage, no one is looking to the far side if HIDING.

Yup, and if they're in a windowless room they won't see the enemy placing a dems charge up against that interconnecting wall either.

Context. It matters

Still even behind a wall, that blindness isn't all bad.

[Erwin]

"This is why I think we need a dedicated Ambush Command. Hide would be assumed as the "before" and therefore an explicit Hide could be interpreted as an option "after" request."

Many players will be very happy with this statement from Steve. Hopefully, it's now merely a question of ETA.

[Battlefront.com]

I'm obviously not familiar with your code, but you already have something that approximates No. 2 reasonably well in the form of "friendly directions." Obviously there are times when it can be "gamed" or otherwise produce unrealistic results.

Actually, we don't. What we do have is a very short range, short term concept of immediate threat. There's no "intelligent" direction, just a bee-line to the map edge when retreating. In other situations it's a bee-line for the nearest cover, with some sense of direction. But once the unit arrives it's done. That's all it can process.

Just so I'm clear, I actually do play CMx2, enjoy it and feel it offers something different than competitors. But I also want it to improve.

Thanks and to be clear we want to improve the game as well. It's just when those improvements require a lot of AI work... we shudder

Not precisely that, but in one game I had a platoon get hit by a couple of mortar rounds as they were moving out of the setup zone. Two of the squads lost a two or three men each and the morale of all of them plummeted. From then on that platoon was not used for any kind of heavy action. They trailed well behind the rest of the company and were used solely to occupy objectives that the rest of the force had captured.

That is not only sensible, but it is probably realistic. Captain sez... "Smitty, you guys got the rough end of it. Hang your guys back a bit and pull security. We can't as more of you right now, bit try and pull your stuff together in case we really need you."

Big difference compared to what most players do. "Smitty, you got... what... 3 guys in 1st Squad effective? How about 2nd Squad? 4? Hmmm... 3rd Squad is kinda messed up morale wise, yes? But they have 5 guys and a BAR? Excellent. You're going to assault that MG nest over there while the rest of us tackle the hard stuff. You don't think that's a good idea? Oh grow up... we're here to win, not to survive!".

Steve

[Steiner14]

They do periodically have someone SPOTTING rather than HIDING. It's just a lot less than all the eyes all the time, so HIDEing units get surprised/contribute less to the overall intel picture than non-HIDEing troops.

My impression is, that there could be room for a conceptional improvement of the spotting capability. Right now it seems, that the spotting capability of hidden units is proportional to the number of eyes spotting. This leads to the strange effect, that a single pair of eyes can have a much worse spotting capability than in RL (hidden units don't see the tank in front of them).

A tank driving in 100 m distance: Twenty eyes should not see more than two eyes because two eyes should already recognize it with 100% probability.

Infantry on open ground coming close to 50m: two eyes should see them and ten times more eyes should not deliver better spotting results.

But:

A tank 1000 m away, or an enemy gun hidden in a wood, or enemy infantry hidden in wood:

Here the number of eyes should make a big difference.

Thesis:

The smaller the visual impact (VI) as a function of size (camouflage reduces the visible size), distance, color, shape and light, the more important the number of eyes become, while the bigger the VI, the less important the number of eyes are.

If the VI reaches a certain value, then one spotter already will be enough for reliable spotting.

The hide command could be connected with the spotting task, which is defined by the cover arc:

The size of the cover arc, determines the maximum distance. A armour cover arc would mean, that for the maximum distance of this arc the VI of the smallest tank is calculated and this gives the number of eyes, that need to spot.

A shorter CA would result in less spotters raising their heads, while a long range CA would need more spotters.

The normal CA for all targets would mean, that infantry, which has a smaller VI, must be spotted at the en of the CA, which would result in more spotters raising their heads.

Additionally each unit has only a limited number of binocs. At a certain small VI, only the number of spotting binocs should be able to gain additional spotting info.

I think there would be plenty of room for enigneered optimizations of the spotting system that could make the need of new commands (ambush) obsolete.

[Hister]

Thesis:

The smaller the visual impact (VI) as a function of size (camouflage reduces the visible size), distance, color, shape and light, the more important the number of eyes become, while the bigger the VI, the less important the number of eyes are.

If the VI reaches a certain value, then one spotter already will be enough for reliable spotting.

The hide command could be connected with the spotting task, which is defined by the cover arc:

The size of the cover arc, determines the maximum distance. A armour cover arc would mean, that for the maximum distance of this arc the VI of the smallest tank is calculated and this gives the number of eyes, that need to spot.

A shorter CA would result in less spotters raising their heads, while a long range CA would need more spotters.

The normal CA for all targets would mean, that infantry, which has a smaller VI, must be spotted at the en of the CA, which would result in more spotters raising their heads.

Additionally each unit has only a limited number of binocs. At a certain small VI, only the number of spotting binocs should be able to gain additional spotting info.

I think there would be plenty of room for enigneered optimizations of the spotting system that could make the need of new commands (ambush) obsolete.

If that ain't a good preposition then I don't know which one is! It can be translated into the coding language elegantly right!?

I don't see the ambush command option should be left out if it gets implemented though.

[poesel]

...

The smaller the visual impact (VI) as a function of size (camouflage reduces the visible size), distance, color, shape and light, ...

Easily written, difficult to execute. This function would have to simulate the pattern recognition scheme of a human being. Would sure be nice to have.

[Steiner14]

Sure and not every person has the same visual capabilities and therefore it needs to be modeled for each soldier, too...

[Battlefront.com]

One pair of eyeballs in CM is theoretically just as capable as another. Experience, Morale, Bonuses, Binoculars, current action, and probably some other things modify a particular pair of eyeballs to be more or less capable as another. Further, any specific pair of eyeballs has to trace a specific line of sight (LOS) between it and whatever it can potentially spot. The more eyeballs that can theoretically spot a particular enemy unit, the greater the chance that one of those pairs of eyeballs can confirm a sighting.

This is as it should be. I don't think anybody will argue against the way it works in the game, today, compared to how it works in the real world.

The problem is that in the real world there is a great amount of imperfection in spotting things. Personal distraction, inexperience, bad luck, etc. all come into play. The problem we have is coming up with a reasonable catch-all "oops, you missed it" system instead of trying (in vain) to directly simulate things such as lighting effects, shadows, blood pulsing in the eyes, short range vision focus (a HUGE problem for spotting), etc. Any time we have to roughly approximate something there's going to be arguably "unrealistic" outcomes. There's no way to prevent that from happening.

Here's an example from real life that many of you might have experienced. Recently I was about 3 hours into a 4 hour night time drive. I am an excellent driver by any definition. I constantly check my mirrors to track what's going on behind me. At this point in my drive there are only tiny pockets of artificial light because, basically, I'm driving through a forest. So it should be very easy to track the progress of cars behind me because the headlights make it quite easy to see them. Easier than in the daylight, in fact.

Yet I suddenly found a car on my immediate left that I had no idea was even close to approaching me. What's more, it wasn't going that much faster than me, so it's not like a German Autobahn example where in a blink of an eye a BMW can be passing you when just a second before there wasn't a car in sight (try driving on the Autobahn with a 4 speed POS rental car... you'll get that a LOT).

So there I am... a Crack Driver, with +2 eyeballs (I have 20/20 vision), easily spotted "enemy", good Morale and Suppression ratings, and what not. But I missed spotting the primary thing I was looking for despite no visual obstructions, good weather, and highly contrasting light/dark benefits. If this CM were a driving simulation you guys would tell me this would be IMPOSSIBLE to have happen. But real life just doesn't work that way.

And what this has to do with KIA:WIA casualties... mystery to me

Steve

[Mord]

And here we'd thought you'd finally run out of automobile analogies/anecdotes.

Mord.

[JonS]

The more eyeballs that can theoretically spot a particular enemy unit, the greater the chance that one of those pairs of eyeballs can confirm a sighting. This is as it should be.

Except ... doesn't that assume that everyone is looking with the same intensity all the time? Which is certainly not how it works in the real world. In a platoon the radio op probably isn't looking at all - he's too busy trying to get the frigging radio to work, or writing down some lame ass message about how vital it is that the CQ gets a signature on the Form 125A. The platoon commander is either looking at his map, compass, or the objective on some hill 2km away. The platoon sgt is more concerned with making sure Pte Numbnutz isn't goofing off again. Besides, he's a sgt, and sgts don't watch arcs. The section commanders are busy trying to find more ammo. The MG #2s are busy making a brew. Most of the riflemen are having a durry, and the rest of them are stacking z's. About the only guys actually watching at all are the three MG gunners, and they figure that they don't have to watch too hard, because, hey - there's a whole platoon keeping watch. Right?

Meanwhile that small team over there KNOW there's only four of them, so while three might busy with other stuff, the fourth knows he's it, he's the man, so he does his job.

There are good reasons why recon patrols are composed of only a few men, rather than massive herds blundering about the place. Overall, small things are harder to see than big things, and that applies just as much to small things with only a few eyes finding it easier to see big things with lots of eyes. As an extreme example, imagine two hills, about 800m apart. On Hill A is a pair of riflemen. On Hill B is an infantry battalion. Both are endeavouring to remain concealed.

1) Which has more eyes?

2) Who will see whom?

If the answer to 1) and 2) are the same, ur doin it rong.

I am an excellent driver by any definition.

We all are