Jump to content

Why infantry combat in CMx2 is so different


Recommended Posts

In JonS' combat DAR (During Action Report) he described a few soldiers, in two units, withdrawing under close observation of the enemy and yet being able to pull back out of the line of fire. You can see his post here at the very bottom. When I read it I was suddenly hit with an "ah-ha" moment that I felt translated well into a "teachable moment".

To do this right I have to first explain things in CMx1 terms since many of you are used to thinking in these terms. For those of you who have played CM:SF/CM:A for a while, and have commented about how infantry feels a lot more realistic, this might help you put your finger on why. But it's complicated and requires a lot of detailed info. Therefore, what follows is pretty dense reading. You have been warned :D

-----

I am sure I'm not the only CMx1 player who quite frequently found myself looking at a unit and thinking "they are dead if they move, so they might as well fight it out where they are". Sometimes, of course, in real life the move would have been suicide, but far too often withdrawing should have been a more viable option than it was. This helped contribute to the unrealistically high body counts at the end of games (though there are many reasons, many found in nearly all wargames).

Assuming you guys know what I'm talking about, why is this not so much of a problem in CMx2 as it was in CMx1? Because CMx2 was written, from the ground up, to better simulate these sorts of things. And to do that, we added a ton of features not seen in CMx1. The big ones that are relevant here are:

1. 1:1 Soldiers

2. 1:1 Spotting

3. 1:1 LOS/LOF

4. Relative Spotting

5. Smaller (8x8m) Terrain

Most of you have seen us talk about these individually, and sometimes that means people don't get why they are so important. Understandable, because in a simulation the strength of the whole is the sum of its parts. Likewise, any one part on its own doesn't produce a rich and complex outcome without the help of others.

Therefore, I am going to first talk about the CMx1 system's parts and the outcomes they produced. Then I'm going to compare that to how CMx2 works and how it's outcomes come about. And don't get me wrong about CMx1... we're very proud of it as I am sure Steve Jobs is still proud of the Apple IIe. It's just that we can be both proud and unsentimental at the same time.

-----

A simulation rule of thumb is that the closer the fidelity of the code matches the fidelity of the simulated environment, the more likely the results will be acceptable (or in our case, realistic). The other rule of thumb is that the more consistent the fidelity is within the simulation, the more consistent the results will be. If you do both then your results should be consistently acceptable (in our case consistently realistic). This is what we attempted to do with CMx1, but there were serious technical limitations to how fine the fidelity could be. Which meant we had adequate and consistent results, but not necessarily optimal for CM's particular simulated level of combat.

The map format and feature set literally sets the stage for the rest of the game. Unfortunately, the technical limitations of the day forced us to adopt a format which wasn't as finely resolved as tactical combat really calls for. It would have been much better for a system where Platoons were the smallest unit, not Teams. As a result the rest of the code had to work with terrain "tiles" that were heavily abstracted and not all that varied in terms of composition.

To keep things internally balanced the other systems within the game were designed to work within the limitations of the map structure. Systems such as LOS/LOF, for example, had to make allowances for things like folds in the ground, uneven vegetation, trees, etc. because they weren't simulated in a direct way. There wasn't much point having individual soldiers simulated (even non-graphically) since each terrain tile was internally identical. Which in turn meant ballistics couldn't be simulated directly since there weren't 13 points of fire for a 13 man unit, but rather 1 point of fire. Since there is only one point of fire there's no reason to simulate individual weapons firing on their own. Instead a volley of fire, made by "adding up" the individual weapons, was the only way to go. With no terrain variations or soldiers to use them, why bother having direct simulation of cover and concealment? So on and so forth.

The resulting simulation at the heart of CMx1 was, therefore, a very well balanced set of abstractions which could only simulate things to the extent their abstractions would allow. It worked beautifully as a game and, to a large extent, a simulation of tactical warfare. However, the fidelity was simply not fine enough for the type of tactical combat Combat Mission set out to simulate. We also had some systems that were "outcome based" in order to make up for the lack of fidelity.

"Outcome based" systems are those which data and/or equations are manipulated to produce specific results, as opposed to a "process based" system which gets its results from a more-or-less scientific approach. A typical wargame has an outcome based armor/ballistics system, which CMx1 largely did not. While it is true outcome based systems can produce some great results with fairly modest programming effort, they tend to produce a lot of bad results. They also tend to break if they are asked to cover too many possibilities or complex relationships. Here's a typical wargame problem you should all be able to recognize:

"Hmmm... that Panther was killed at 3000m by a Stuart. That's not right. So we'll up the Panther's armor a bit. But now a Sherman can't kill a Panther even at 1000m. Hmm... well, we'll put the Panther's armor back to where it was and decrease the Stuart's penetration value. Ooops... now the Stuart can't take out a PzIV from the rear, which it should be able to do. Hmmm... well, we can..." and so on and so forth. If the environment is simple, like RTS games are, then this sort of tail chasing usually isn't too bad. But for something like tactical combat, where realistic results are expected, it's usually futile. Which is why MOST of CMx1's game systems were not "outcome based". But some definitely were.

Absolute Spotting was one of the worst "outcome based" systems in CMx1. Trying to balance things so an AT rocket team wasn't spotted 1000m away, yet wouldn't remain near invisible at 50m was EXTREMELY difficult to do. And once the unit was spotted we had to make the TacAI not go about targeting it with everything it could theoretically use. In the early days of CMBO we called this the Yellow Lines Of Death. As good as we got the system to be, relative to where we started, it was still far away from where we all (including you guys) wanted it to be. But there wasn't really anything more we could do with it.

So why not just get rid of Absolute Spotting and go with Relative Spotting? To do that we would need to have a new Command and Control (C2) system to know which information a unit should and shouldn't know. This, in turn, required the core of the unit data to be ripped out and replaced. Doing that would require re-hooking up various things which were expecting the old data format. That would probably, in turn, require even more sections of code to be significantly altered. Cause, effect, cause, effect... it's a nightmare sometimes.

Doing all of this for a game system that we knew didn't have the right terrain fidelity and effects was never considered viable. It's why in 2001, half way through programming CMBB, we decided the CMx1 game engine had to be abandoned and a new engine created if we wanted push the envelope again. Especially if we wanted infantry to be the focus of Combat Mission, and we very much did. Enter CMx2...

----

During the CMBB development we were disappointed with our inability to significantly improve infantry combat modeling. It was a that point Charles and I decided the prime focus of CMx2 was to improve both the graphical experience and also the infantry combat. The two go hand in hand because one without the other isn't really possible. At the time we didn't know what the specs would be for the new engine, we just knew that as soon as CMBB was done we would start on it. And that we did, while at the same time making CMAK.

Keeping in mind the two rules of thumb noted above, the CMx2 engine was based around the 8x8m Action Spot system instead of the old 20x20m Tile system. This allowed us to increase the fidelity of all the sub systems proportionally. It took a TON of work to get these systems in place and working well with each other, but the results show that the effort was well worth it. As JonS' little retrograde move (remember that? ;)) shows quite clearly.

In Jon's DAR he had two small units leapfrog back to protective cover. They did this despite already being in contact with a much larger enemy force. So how was this more likely to happen in CMx2 than in CMx1? Because the many sub systems in the game allowed it to happen "naturally". In no particular order, since these all happened near simultaneously, here's what happened under the hood...

Elvis' soldiers were only allowed to shoot at Jon's guys if they had actually spotted them. This removed a significant number of potential shooters from the equation that would have been available in CMx1's Absolute Spotting environment. Of the soldiers who were aware of Jon's guys on the move, not all were physically in a position to take aim at where Jon's soldiers actually were. Perhaps a slight fold in the ground blocked their LOF, perhaps a tree got in the way of tracking the target, etc. Reducing the pool of potential shooters even further, even if Elvis' soldiers had LOF on Jon's guys, they might have been reloading, looking elsewhere, perhaps firing elsewhere, taking cover from something, took too long to aim, etc.

On Jon's side of things the terrain really helped out here. As his soldiers withdrew they interacted with the surrounding terrain. Tall grass, trees, etc. offer cover and/or concealment to some extent even when on the move, though more so when stationary (obviously). This screwed around with the enemy's ability to track and fire effectively. And when Jon's guys went to ground they might have even become no longer viable targets due to height differentials between Elvis' and Jon's forces. And since Jon only had to withdraw a fairly limited distance to near fully concealed terrain, Elvis' soldiers had a limited window of opportunity to strike them down. If this had been CMx1 the terrain wouldn't have had any affect unless Jon's guys moved from one 20x20m tile to another which offered situationally better cover/concealment, or perhaps a significant elevation fold in the middle of the tile. Meaning, as Jon's soldiers started moving they would usually have been just as vulnerable with each meter moved.

So what about random variables? Surely CMx1's system had variables involved which would have allowed Jon's men to slip away in one play through and then eliminated in another, right? Well, yes and no. Variability was certainly built into CMx1's systems, however they had to be outcome based. Set the variables so a 2 man unit at x distance in y terrain can withstand a couple of platoons firing at it and suddenly you find other situations where entire enemy platoons slip away because the internal weights are now favoring the defending unit too much. Push the variables around so that situation produces better results and now those 2 guys are again toast the second they stand up. And so goes the tail chasing.

The more extreme the range of unrealistic possibilities, the less likely you're going to wind up with a good range of results. In one situation Elvis could have had several hundred guys, several tanks, and some mortars able to shoot at Jon's 2 hapless soldiers. In another situation Elvis might only have had 2 guys capable of shooting at Jon's 2 soldiers. That's quite a spread to ask any system to be able to handle equally well in any and all situations.

In CMx2, however, such spreads don't happen because the systems "naturally" chip away at the unrealistic possibilities and leave only the realistic ones to be considered. The volume of Elvis' firepower was "naturally" restricted to only those who really should have had a chance to fire in the first place. And when they did fire they had to deal with individual accuracy, individual quality of LOF, individuals being ready to shoot at all, etc. This means we only have to balance the individual's results (1 on 1) and "naturally" 10 soldiers firing will be realistically treated because it's 10 individuals and not a volley of 10 grouped shots. The fidelity of CMx2 makes this possible in a way that CMx1 simply couldn't do at all. Period.

Simply put, without the 1:1 simulations, finer terrain treatment, and Relative Spotting Jon's little pixeltruppen would have probably been wiped out nearly instantly. This is not to say CMx2 is always going to produce a realistic result and CMx1 never did. That would be wrong. I'm just saying that given the same set of circumstances in both games, CMx2 is more likely to produce a realistic outcome more frequently than CMx1 would. It is even possible that CMx1 would never produce a realistic result for this specific situation, while I would not say that possibility exists for CMx2.

Hope that long read was interesting :D

Steve

Link to post
Share on other sites
  • Replies 264
  • Created
  • Last Reply

Top Posters In This Topic

Hope that long read was interesting :D

Steve

Very...and it was one of the things I wanted sooo much after playing CMX1 games...better infantry modeling....CMX1 always felt like the tanks and other armor where handle really well but the troops where more tacked on because, well you needed infantry. Now it feels like they are given the same lovin'.

Something that I think is very important (and you kinda touched on it) and worth noting is that in CMX1 if your squad was hit from two sides (or more) you only had two choices...die or flee...and if you fled (on purpose or by breaking) you died some more...In CMX2 I have seen my guys many times, engaging three different enemies in three different directions...it's very cool not having your dudes based around a combat modifier but displaying realistic tactics...not all bunching their fire on one squad...but taking shots when opportunities present themselves (one dude can end up making a real difference). A good example of multi-directional fighting is in this pic...

cmnormandy.jpg

It's gonna be a whole new world having 1:1 in WWII...I've been dreaming of it for so long It's hard to believe it's almost here.

Mord.

P.S. What Elmar said.

Link to post
Share on other sites

Cracking read, thanks for taking the time to explain. Suspect i will never go back to CMx1 after reading that simply because i would feel like its a waste of time when an more realistic tactical simulation/engine in available. I think naively i just thought CMx2 had better graphics.

Link to post
Share on other sites

There are many times while playing CMSF where my soldiers are on a small ridge line or crest of a small hill for enfilade cover and shooting at a distant enemy through trees or something equivalent. But I noticed that not all are shooting. Maybe half of the squad (or less) is shooting. The other half are in la-la land (at least that's my thinking). And I'm yelling "dont you see them!?!?! Shoot you dumb SOBs!!!

Link to post
Share on other sites
There are many times while playing CMSF where my soldiers are on a small ridge line or crest of a small hill for enfilade cover and shooting at a distant enemy through trees or something equivalent. But I noticed that not all are shooting. Maybe half of the squad (or less) is shooting. The other half are in la-la land (at least that's my thinking). And I'm yelling "dont you see them!?!?! Shoot you dumb SOBs!!!

This does happen occasionally, but I find that most of the time when the squad makes contact, the NCO yells at his men and they take better firing positions aligned at their target.

I witnessed the beauty of this when playing a small British module scenario yesterday where 2 Challanger 2s are stuck in the mud inside a city. You only have a handful of squads racing to the rescue and setting up the perimeter. I didn't have to assign covered arcs because the AI automatically adjusted to threats from different directions and kept 1 guy covering the other approaches. :cool:

In another vanilla CMSF scenario, assault on a factory, I only had to "Quick" an FO team, MG team and sniper team to a ridge and they automatically took positions on the ridgeline.

Link to post
Share on other sites

I just did a quick setup to show the 1:1 aspect for infantry. Below is a link to a video showing the "test range":

In it, a Dutch Engineer squad is observing an area which is completely divided by a high wall. Team A (4 men) is on the right and Team B (5 men) is on the left. Each team can only see its side of the wall and can therefore only engage enemies on their side (which can be easily seen in the video). The enemy force consists of 2 syrian recon squads and their platoon hq, which were ordered to "Quick" towards the dutch squads position.

So, more than one enemy squad can be engaged by different members of the squad. Also, the Teams cannot engage enemies on the other side, restricting the overall squad firepower affecting each enemy squad.

I think it is quite obvious that this is an improvement over CM1 squad to squad (or single point to single point) engagements.

Link to post
Share on other sites
Ok, I'm using the exact same mod but thought it sounded a bit different. Might be the compression.

It can sound pretty different depending on how you setup up the mod's modules and extras. Mad Mike, for example, appears to be using the A10 impacts extra.

Link to post
Share on other sites
But if a real life commander knew the men were there, he would order his men to area fire it too. :D

Sure, but it would first need to go up via C2 and then back down. In the meantime the withdrawing team would be gone :-)

Thus the desired end result would be the same, even if the player was aware of the situation.

Link to post
Share on other sites

If you want to learn how to withdraw while under fire in CMx2 just play as the insurgents in any scenario out there. Withdrawing under fire is basically a way of life for Red Forces in CMSF. Yes, it can be done while under area fire as well. In fact, if you are good enough and your opponent likes to do a lot of area firing, you can make your opponent waste a lot of ammunition firing at ghosts from old contacts. It just gets dicey if your pixeltruppen are subjected to so much firepower that they get pinned - then you are pretty much screwed unless the fire slacks off and you can slink away a little later. So, you have to time your withdrawal carefully. Wait too long and you get hosed. Do it too early and your position is surrendered before it needs to be.

Link to post
Share on other sites

Mord,

Good point about the multiple targets thing. Yet another layer to the onion that even I take for granted now!

In CMx1 we had to add special code to handle engagements by multiple units against a single unit. The code included taking into account the differences in engagement angle, so that a unit would behave differently if taking fire from two points in the same arc or (for example) directly in front AND directly from behind. The effect realistically increased the chances of casualties (can't hide equally well in all directions at once!) and Morale ("they're everywhere!!"). And from the attacker's standpoint this was really good.

From the defender's standpoint we had trouble balancing this out. Since the unit could only literally fire at one enemy unit at one time (that's the way the game core was coded), we couldn't have the unit split it's fire between two targets. Plus, even if we could this wasn't necessarily the best course of action nor necessarily the most realistic. Other factors, such as which enemy unit posed the biggest threat, which could be engaged more effectively, which direction was the "front", etc. would have to be included to have everything work correctly.

Some people tried to work around this by Splitting Squads. This allowed two points of fire and reduced the chances of all men being engaged by multiple enemy units simultaneously. But for other gameplay reasons we penalized units being split and therefore the separated parts were weaker than if they had been a whole unit. It also added about 30% more units to mess with if used extensively, which many people probably felt took some of the fun out of playing. Yet another example of tail chasing in terms of game cause/effect. Fix one problem and create another.

Another fix was put into CMBB, especially for MG units. And that was more rapid target switching. A unit could still only fire at one target at a time, but it could now switch between targets quicker if there were a lot of targets available. This made a huge improvement for both attacker and defender alike, but especially for the defender. It certainly lessened the problem situations both in frequency and bad outcomes, but it still left the basic problem unsolved.

In CMx2 things are very different. Individual soldiers fire at targets of opportunity. If they aren't able to shoot at Enemy Unit A, but can engage a soldier in Enemy Unit B... he has at it. So while the unit does try to concentrate its fire on a particular enemy unit, the individual soldiers within can make their own targeting choices. This means, in theory, a 9 man squad can engage 9 different units simultaneously. Obviously that probably has never happened because the conditions almost definitely never exist for this, but in theory the game mechanics could do that.

Another thing that really helps, and Mad Mike's video shows this, is that Squads now consist of either 2 or 3 Teams. From the game's perspective, the Teams are separate units. They have, inherently, their own targeting priorities. Think of it as a hierarchy of target selection priorities...

Squad

Team

Soldier

When you issue a Fire order to a Squad the simulated Squad Leader issues orders to its virtual Team Leaders to instruct its Soldiers to fire at at that target. If the Team can't engage, or feels it should really shoot at something else instead (mostly out of self defense), the simulated Team Leader directs its Soldiers to fire on a different target instead. If an individual Soldier can't get a shot off he will try, to some extent, to maneuver himself so that he can. But we only allow this to a certain extent in order to maintain spacing and proper use of cover. No sense having your Soldiers move to better firing spots only to be killed because there is no cover there. And all the while the TacAI is deciding if there are other things to worry about and if someone should shoot at it.

This system is far more "natural" than the CMx1 system. It means the system can be stressed out more and yet still function consistently well when compared to CMx1. It also means that we have far less tail chasing to do in CMx2 than we do in CMx1.

That being said, we're not saying that CMx2's behavior is 100% perfect. It isn't and will never be. What we're saying is that it is vastly more flexible and produces far more result that are consistently realistic in a wider range of situations.

Some of the problems that exist have nothing to do with the described firing/targeting behavior. The well known "ridge line" problem, for example, has to do with limitations of the Action Spot system. Fix that and the infantry behavior on those spots will be improved. It's something we can fix without tail chasing. In fact, it's on my own personal short list of things to try and fix before CM:BN ships. I don't know if we will get to it before then, but eventually we will.

Steve

Link to post
Share on other sites

I wish I knew what part of the game you guys are talking about. Steve post was written very cleverly to not reveal where the action took place and which units were involved. My guess it that it was during the bloodbath on Hill 154 when his platoon seems to just melt away into the woods.....

Link to post
Share on other sites

Does it model squads still firing even though they have no direct target? As Jon's Troops are being pulled back it makes sense that Elvis's Troops would still fire even though they can't see anybody. I know in CMSF that they will fire a few bursts at a building after loosing sight of the individual target. Maybe a happy medium can be developed where they will keep firing but with little accuracy and few casualties but increasing retreating troops "panic" level or level of suppression. Again....don't know if this is already included.

Link to post
Share on other sites

Great topic Steve, and thanks for the insights. At least in the early days CMSF infantry (especially poor hapless RED) suffered -- not unrealistically -- from overexposure in the relatively open desert environment (plus IMHO overuse of the default building layout with windows on all sides made units in them excessively/quickly vulnerable to suppressive fire from multiple angles). Not to mention the fact that a modern infantry squad has several times the FP of a WWII one.

So CMBN, with its better cover/concealment and lower FP will be a far better showcase for the ballistics model. (although it looks like CMA has already shown some movement in that direction).

P.S. I viewed the vid on my IPhone which made me all dreamy about the future prospects of CMing on a mobile device. Goodbye work productivity! :-)

Link to post
Share on other sites

Steve, interesting post and the new engine certaily seems as though it is now capable of going through several mid-service life refits, couple of questions though (if any have been previously answered I apologise)

1. In CM 1 when a 'volley' was fired did all weapons of the same group fire, ie Rifles, SMG's etc? It would explain why the Russian SMG troops were as effective as the Grim Reapers scythe!

2. Does this now mean the squad leaders proficiency affects the squads performance, ie targetting, morale etc? In CM 1 a weakness was the lack of any representation of this apart from the psychic influence of the platoons the damn it sarge! we was good at hiding but now the lines are black, not red I've forgotten how to do it" syndrome. I really hope the guy with the SMG is now better modeled than someone who's good at shooting close range.

3. Does the experience of troops affect the speed at which they aquire targets and how well they shoot, also, correspondingly, does it afford them more protection for any given terrain.

4. Will doctrine be represented, the US semi-auto rifle enabled a high rate of fire at area targets, the bolt action rifles represented a tradition of firing at identified targets (hence the post-war revisions to infantry docrine that followed the analysis combat in WWII). Obviously the Germans were ahead of the game with their MG series being specifically designed to engage rapidly moving and fleeting targets with maximun fire power.

5. Will any approximation of the reluctance to shoot be included? You said that there is a reason why wargames end in blood baths, one of the reasons is the vastly overscaling of firepower because of following theoretical maximums. Most fire is suppressive in nature with the killing work at short ranges, or somebody does something daft, like attack in a human wave.

6. Will we ever have the ability to organise out squads into formations (line arrow head, double interval etc)?

7. Will jamming or overheating of weapons be simulated?

Link to post
Share on other sites

Nearly forgot,

8. Will the ammo expenditure of units be affected by their proficiency level?

9. How does the psychological state of the men, taking fire, taking casualties affect their accuracy, ROF and aquisition time and will training mitigate these affects?

Link to post
Share on other sites
Guest
This topic is now closed to further replies.

×
×
  • Create New...