CMx2's new movement system explained

[aka_tom_w]

I think Steve has sort of defined a "command game" as some form of text only game paradigm where you simulate the role of only one level of command and send text orders one command level down and wait for a text message back from that level of command without ever seeing the battlefield or the other actual units.

Within that definition Steve was talking about CM:SF is not technically a "command game" (one level) because the player can play the role of all levels of command.

AS you said:

This game is exactly a platoon/company/troop/battalion commander simulator

This implies a different "kind" of game more like a (and I dont' think this is the name Steve called it) combat simulator rather then a "command level" game.

Sorry I did not mean to split hairs

somewhere there is a very lengthy post by Steve about what a "command game" is and how they are not interested in developing command level games. (or something)

-tom w

[M1A1TC]

I know everyone sees CM games as whatever they want. Who cares as long as you have fun

[aka_tom_w]

Some of this might be interesting take another look at

look at the OLD date, this is from back in Feb 2005 almost 10 months ago

he he

Battlefront.com

Administrator

Member # 42

posted February 02, 2005 01:12 AM

Interesting discussion. The most interesting part about it (no offense) is that this is the same discussion we always have. Well, at least when you boil down the ideas. This is not a slam, rather a painful recognition about how few options there are open to us. To summarize these options, here are some definitions we can put to the different concepts for limiting the God problem (i.e. seeing everything there is to see and acting on that info):

Player as Commander - in this the player can only see and interact with the information his character (Commander) has become aware of through realistic C&C systems.

Positive - Seeks to stop the problems before they start in the most straight forward, realistic, and simplistic way possible. If you can't see it (or talk to someone who can), you can't act on it (or instruct someone else who can).

Negative - very restrictive in terms of traditional gameplay. The player's ability to play the game (traditionally speaking) is now at the mercy of circumstances that may not be within the player's ability to control.

Command Delays - player can see whatever he sees, but can't necessarily act on that information effectively.

Positive - Has the effect of slowing down the player's ability to act on "ill gotten gains" while not interfering with the player's control of units per se.

Negative - it is very difficult to make the desired outcome realistic because the problem (excessive intel) is already present. Since modern (i.e. WWII - present) communications are fairly quick when things go right, limiting reaction times runs the risk of making actions unrealistically slow and/or inflexible.

Fog of Game - game system purposefully hides information from the player that a unit under his control can actually "see" from a simulation standpoint.

Positive - player retains control of his units but does not get information he "shouldn't" have.

Negative - player's orders for the unit are lacking information that unit SHOULD have. Difficult for the player to understand what is known but hidden from the unknown and not yet discovered.

MIA Units - when a unit goes out of C&C it not only stops reporting what it sees, but it is now hidden from the player entirely. It's as if the unit went "poof" and disappeared.

Positive - units no longer can act as eyes and ears without a mouth and listening ears to hear reports. Also eliminates uncertainty because if you can't see your unit you can't see the enemy's units either.

Negative - oh boy... lots of them!! First of all there is the whole issue of units leaving player control at the last second of a turn, being unavailable during the Orders Phase, then 1 second into the next turn suddenly becoming aware. Now all the player can do is watch and hope the unit doesn't get eliminated or lose C&C by the next orders phase. And speaking of that... how is the unit going to behave without the player to guide it? SOPs might help, but that can't substitute for good AI (tactical and slightly higher) which is also hard to do.

Relative Spotting - units can only see what they themselves have spotted. A unit can't shoot or do anything else (directly) to unit/s it isn't directly aware of.

Positive - eliminates many fundamental problems with the current game system, including a host of ones not directly related to this topic. One of which is to make enemy units harder to spot in the first place since less eyeballs in one spot means less chance of spotting.

Negative - when done right the feature has no negatives, though it does have some shortcomings. Players can still get around some of the restrictions (area fire for example) some of the time and Relative Spotting does nothing in terms of limiting the God influence on the enemy intel that is gathered legitimately.

Dedicated Scouts - only certain units have scout like capabilities.

Positives - limits the range of units that can be used for "gamey recon" to those which the designers have chosen.

Negatives - the last bit says it all This does nothing to address the problem, but instead simply narrows down which units can be abused. Players will almost certainly figure out how to use other ones in some ways in some circumstances. On top of that, it most certainly involves unrealistically blinding the rank and file units, which is indefensible from a realism standpoint.

Those are the major groupings as I see them. There are variations on these, but I'd say this is the bulk of them.

So, which work to solve the problems of gamey recon? NONE. Each one of these does not fix the problem because the inherent problem is the Human player can see too much and act in too coordinated a fashion. The Player as Commander concept is the only one that even attempts to simulate the real world restrictions of commanders in the field, but even that doesn't go far enough. The others tend to penalize realistic behavior in the pursuit of curbing unrealistic results. Results that are inherent to the Human player, not the system. Tough challenge.

Ok, so how are we going to fix gamey recon (which is just a part of the God problem)? We aren't. There is no way we can. However, by using a combo of concepts, carefully measured, we do expect to come up with a system that greatly reduces the effectiveness of acting on information that shouldn't be there in the first place. Relative Spotting is the #1 part of that plan, but of course there is a lot more to it than that.

Just my thoughts

Steve

AND

Battlefront.com

Administrator

Member # 42

posted February 02, 2005 08:18 PM

As for bailed out crews... they already are less likely to spot, more likely to route, and are easier to kill than any other unit. But that doesn't make much of a difference sometimes

One way to fix this is to apply a MIA UNITS feature to Panicked/Routed units. This is, IMHO, a legitimate possibility. Close Combat had something similar, though you never lost sight of the guys. It is something that is being considered, though there are large implications that need to be considered once some other elements of the game have been finalized.

Yes, CoPlay (co-op play) eliminates all sorts of Intel problems through very hard nosed restriction of intel sharing. Unfortunately, that creates another set of challenges for us since players will need some way of sharing intel. And that is why it isn't happening for CMx2's first release

Importance of HQs in CMx2 will be much more realistic because we will have a full bore simulation of Command and Control. That in turn is made possible by Relative Spotting, since without it there is little we can do to make HQs realistic. Can't say more than that at this point.

I will remind you all again... we are very much against implementing any system that can not be defended from a realism standpoint. Command delays that are not somehow related to realistic command delays are not up for consideation. Delaying spotting information even one second from the unit that spotted it is also completely out of the question. Both of these things have very large potential for gameplay problems that are worse than the problem they seek to fix (cure worse than disease sort of thing).

One has to remember that the overall realism of the game is important. Hobble one aspect of realism and you almost certainly cause something else to be unrealistic. Usually quite a lot of things, sometimes quite small and only noticable in certain circumstances. Then we spend tons of time trying to fix the side effects, which likely cause other side effects, and that requires more time for fixing, etc., etc. Therefore, any game design element which we know to be inherently unrealistic is not up for consideration.

Steve

[Abbott]

Originally posted by Battlefront.com:

Commands are divided up into 4 logical groupings:

Movement Commands - basic Commands to get a unit from A to B.

Combat Commands - basic Commands to instruct a unit how to use its weapon/s and/or weapon/s systems.

Special Commands - Commands for speciality functions specific to that type of unit and/or

I am guessing that spraying the Silly String will be under the "special" command?

[lucero1148]

This is going to be a great game.

[Battlefront.com]

Quick answers...

CM isn't a command level game. When CoPlay is introduced it will be more like that, but only when playing with other people. The reason? Because each person will be assigned a specific and limited role, which is absolutely not the case when playing single player and anything over about Platoon size.

We can not code highly detailed, condition specific AI into each type of unit. Tankers spend years getting it right... it seems rather unreasonble to think that we can code all situations up in a few days. Give us a few months and we could do something pretty damned nice, though it would likely bog down the hardware and still leave us with many more months of work to do for other types of units.

The TacAI can only handle the basic stuff. Knowing when it should rotate its turret in anticipation of a threat that may or may not exist is simply not going to happen. So... we either allow the player no ability to do subtle commands for their units, like CMx1, or we do. We felt allowing this was a much better option.

Yes, scenarios will have the ability to contain "event" scripts of various sorts. The Commands are not designed for that, though we do intend on allowing units to start out with Commands already issued.

Yes, Silly String is a Special Command for the M92A3 non-lethal projector Actually, that is another reason we don't want to do a stability ops game. Let's keep everything nice and lethal and make our lives easier!

Steve

[Abbott]

Originally posted by Battlefront.com:

(snipped)

Yes, scenarios will have the ability to contain "event" scripts of various sorts. The Commands are not designed for that, though we do intend on allowing units to start out with Commands already issued.

Steve

Are we talking about a few tools scenario designers may be able to use in programming simple unit actions in addition to the AI’s responses?

[M1A1TC]

Ok, say a platoon of tanks in a wedge formation is moving across the desert. Each one is assigned a fire sector, which is the same regardless. The gunner of the lead tank, for example, scans from about 10 o'clock position relative to the hull to 2 o'clock position. The rest are scanning in their own sectors and avoid firing if friendly unit is infront. Why this can not be modeled?

Say you have a staggered column, then the last tank scans to the rear for targets automatically. The second tank scans left, third scans to the right. If the enemy is detected to the right, for example, each tank turns to the right, gets on-line, and gunners scan 10-2 o'clock.

Will the units in the game be able to tell what formation the platoon is in?

If you need help, I have Tank Platoon SOP, Tank Gunnery TMs, Sight Picture Training M1 Tank, etc

[ November 05, 2005, 04:28 PM: Message edited by: M1A1TankCommander ]

[aka_tom_w]

Have you done away with the cover arc??

or is it one of the things that worked well in CMBB an CMAK that will be kept in CMx2??

(or is there something like it but not actually a "cover arc" command?)

just curious

thanks

-tom w

[ November 05, 2005, 09:57 PM: Message edited by: aka_tom_w ]

[M1A1TC]

Here is a scan I just did to show what I am talking about:

Gunner's scanning for targets is realistic and I was sad not to see it in CM games

[Michael Emrys]

Originally posted by M1A1TankCommander:

Ok, say a platoon of tanks in a wedge formation is moving across the desert. Each one is assigned a fire sector, which is the same regardless. The gunner of the lead tank, for example, scans from about 10 o'clock position relative to the hull to 2 o'clock position. The rest are scanning in their own sectors and avoid firing if friendly unit is infront. Why this can not be modeled?

What makes you assume it won't be? I've done this very thing in CMx1 many times using covered arcs. From the sound of what Steve has posted, either those will be retained or something effectively similar will be included in x2, the only difference being that you can order a reorientation of the covered arc at each waypoint rather than being stuck with the same one throughout a turn.

Michael

[Michael Emrys]

Originally posted by M1A1TankCommander:

How many squad leaders do you know that fly around the battlefield giving orders to the whole battalion, besides their own squad?

Er, you seem to be laboring under a misunderstanding. It's not that one squad leader is flying around the battlefield giving orders to all the other units. It's the player who becomes the leader/commander of each unit as he clicks on it. Not totally realistic either, but I guess the best compromise that could be come up with at the time.

Michael

[Cameroon]

Originally posted by sgtgoody (esq):

I want voice recognition software so I can hook up my old CVC to my computer and relive my glory days!

Seriously thought it sounds interesting, can't wait to try it out.

Take a look at a program called "Shoot" -- it uses the Voice Recognition tools that Microsoft provides for Windows to "press keys" and such. I use it for the game X2: The Threat and others; works brilliantly.

I imagine it could be real handy for CMx2, too

[juan_gigante]

Yeah, M1A1, my understanding from Steve's description is that the scenario you described would not only be possible, but something one would do with some regularity.

And I have always thought of CMx1 as a "hive mind" game. In game, I represent the combined consiousness of all the unit commanders on the field (on my side, that is). As I think, they think. But that's kind of cop-out way of describing it.

[Bluefish]

Originally posted by Battlefront.com:

...especially because it now has a memory for targets.

In CMBO terms, that panzerschrek 50m off in the woods has just dropped out of sight. My Shermans turet will still point in his last known direction instead of rotating to the front?

This would be an absolute breakthrough improvement of the TacAI if so.

[kipanderson]

Hi,

Without sounding too much like a mindless fan… which I suppose I am , it does all sounds great.

I was not sure what Steve had in mind when saying that the urban, closing fighting will be greatly enhanced, but now I do understand.

For those who are truly WWII nuts this must also be good news. Imagine how these changes will impact on the modelling of fighting in the hedgerows of Normandy or streets of some Soviet city or village. Stunning !

Greatly looking forward to CMSF.

All the best,

Kip.

PS. I can also see why Steve said that most of us will wish to keep the games fairly small with new system. (But come CoPlay and huge games will be much in demand from me… there is always one .)

[aka_tom_w]

any more news on if the old cover arc is in or out?

-tom w

[Andrew H.]

Originally posted by M1A1TankCommander:

Ok, say a platoon of tanks in a wedge formation is moving across the desert. Each one is assigned a fire sector, which is the same regardless. The gunner of the lead tank, for example, scans from about 10 o'clock position relative to the hull to 2 o'clock position. The rest are scanning in their own sectors and avoid firing if friendly unit is infront. Why this can not be modeled?

Say you have a staggered column, then the last tank scans to the rear for targets automatically. The second tank scans left, third scans to the right. If the enemy is detected to the right, for example, each tank turns to the right, gets on-line, and gunners scan 10-2 o'clock.

Will the units in the game be able to tell what formation the platoon is in?

If you need help, I have Tank Platoon SOP, Tank Gunnery TMs, Sight Picture Training M1 Tank, etc

and

The TAC AI should detect the street on the right, and tell the gunner to scan down the street. I dont want to do it myself- as a commander- for each vehicle. Thats gunner's job.

You supposed to be playing a bigger role

I suspect that this sort of SOP and this sort of tac ai behavior - both of which are already in CM1 - will remain. Specifically, if you drive down a street, the tank will spot any threats down a side street and turn its turret to engage the threat. Similarly, you can (and should) use the CA command to give your tanks specific spotting areas.

When steve wrote:

Knowing when it should rotate its turret in anticipation of a threat that may or may not exist is simply not going to happen. So... we either allow the player no ability to do subtle commands for their units, like CMx1, or we do.

I believe that he is talking about a situation where, say, the primary threat is from the front, but other units have reported a possible RPG team down a side street. In this situation, you want the tank to mostly face forward, because that's where the greater threat is, *except* when it crosses a side street.

You can't do this in CM1, nor is it covered by normal facing SOPs. You can have a tank *always* face in the direction of the side street - but then it is vunerable to shots from ahead, *and it will keep facing sideways both before and after it crosses the dangerous sidestreet*. Or you can have a the tank always face forward, but then it is vulnerable as it crosses the sidestreet.

I don't see any way for the tac ai to deal with this realistically, and players would soon be annoyed if their tank turned its turret to look down a side street every time it crossed one, since this would make them much more vulnerable from the front.

[Battlefront.com]

M1A1TankCommander

Ok, say a platoon of tanks in a wedge formation is moving across the desert. Each one is assigned a fire sector, which is the same regardless. The gunner of the lead tank, for example, scans from about 10 o'clock position relative to the hull to 2 o'clock position. The rest are scanning in their own sectors and avoid firing if friendly unit is infront. Why this can not be modeled?

Because it is a ton of work, that's why Yes, in theory we can do formations and the AI that goes with it, but man oh man oh man is it a lot of work. If we had to only do ONE formation for ONE type of vehicle in ONE type of situation... no problemo. But that's not the way things are.

Will the units in the game be able to tell what formation the platoon is in?

No, because that is one of the fundamental problems with doing this sort of AI work. First, we have to define what a formation is and what SOPs go along for each vehicle within the formation. Nex we have to define positions within a specific type of formation, what is expected of each position, what it should do in the event it can't keep in formation, how it should react to different types of threats in order to remain consistent with the formation, and probably other things like this. Then w Then we have to create a decision matrix for a simulated Platoon Leader so that he chooses the correct formation for the setting. And to do that the simulated Platoon Leader needs to know what all the possible settings are and how good, or bad, a particular formation is for each one. That, of course, means programming all possible formations and their parameters. Then we have to move along and do the same thing for Company level since the parameters are not inherently the same. Nor are the formations. Then we have to do this for every other type of unit, then do entirely different matrix so that the chosen formations and parameters can be influenced by what formations and parameters units that it is working with are using.

It's a big, stinking mess of AI work. Yes, the FMs and TMs make a lot of this work much, much easier... but it still means a lot of coding work. All the FMs and TMs do is make it more likely that the end result of all that work has a decent chance of generally working as intended.

We'd like to do formations, and I think we will someday, but it certainly will not be in CM:SF. I doubt it will even be in CM's second release.

Steve

[Battlefront.com]

Abott,

Are we talking about a few tools scenario designers may be able to use in programming simple unit actions in addition to the AI’s responses?

Yes. The exactly nature of these tools is not something I can discuss at this moment, but the AI can be given a lot of help so that it doesn't have to figure out everything on its own. In relation to Commands, I would like there to be a way that even friendly units MUST carry out an initial batch of Commands. For example, let us say there is a scenario where a convoy is ambushed. It seems silly to start out the scenario with all the vehicles starting out at a dead stop and the player in the position of altering things in a "gamey way" (i.e. knowing this contact will happen and therefore taking preemptive evasive action). So instead perhaps all vehicles MUST drive at x speed down y road until z spot. Then you are assured of an ambush happening, and therefore assured that the player will have to react to it.

Michael.

What makes you assume it won't be? I've done this very thing in CMx1 many times using covered arcs.

Correct, but I think M1A1's beef is that the AI doesn't do this for you automatically, and instead forces the player to do this (or not) on his own. Because, as you say, it is very easy to get the sort of behavior out of a tank platoon in CMx1 as M1A1 is asking for. But not automatically. It is one way to distinguish a good player from a bad one since the TacAI doesn't automatically do this for everybody all the time every time according to doctrine. A good player who issues these sorts of Commands will likely do better than a player that doesn't.

From the sound of what Steve has posted, either those will be retained or something effectively similar will be included in x2, the only difference being that you can order a reorientation of the covered arc at each waypoint rather than being stuck with the same one throughout a turn.

Correct.

Er, you seem to be laboring under a misunderstanding. It's not that one squad leader is flying around the battlefield giving orders to all the other units. It's the player who becomes the leader/commander of each unit as he clicks on it. Not totally realistic either, but I guess the best compromise that could be come up with at the time.

Correct. As we have discussed hundreds of times before, unless we have brilliant TacAI (like M1A1 is asking for) or CoPlay (with a player directly commanding nothing more than a platoon), the CMx1 system of one player wearing many hats simultaneously is the only acceptable game design to go with.

Bluefish,

In CMBO terms, that panzerschrek 50m off in the woods has just dropped out of sight. My Shermans turet will still point in his last known direction instead of rotating to the front?

Correct. There will come a time when the Sherman will decide to move on (either proactively or reactively) to other targets, but it will do so only when it makes sense to instead of automatically when the target moves out of LOS. The latter is how CMx1, obviously works.

Cover Arcs work the same in CMx2 as they do in CMx1. It's just that you don't have to be stuck with the same Cover Arc for the entire duration of the unit's movement.

Steve

[Bonxa]

Will we be able to re-map the keys? I'm not sure I like to have the commands all the way over by the keypad.

[aka_tom_w]

Cover Arcs work the same in CMx2 as they do in CMx1. It's just that you don't have to be stuck with the same Cover Arc for the entire duration of the unit's movement.

Steve

That sounds great

thanks for the update

-tom w

[RSColonel_131st]

Originally posted by Bonxa:

Will we be able to re-map the keys? I'm not sure I like to have the commands all the way over by the keypad.

Seconded. For me in CMx1, associating commands with keypresses works well because of the labels "F=Fast, M=Move, A=ADvance" etc.

Associating commands with the geometrical position on the keyboard is a bit less "direct".

[Battlefront.com]

Yes, we plan on allowing custom key mapping. However, they won't be dynamically assigned based on what group of Commands are present. In other words, you can't have F = Fast for Movement Commands and F = Fire for Combat Commands.

Steve

[dalem]

Steve, since you're giving stuff out here, can I have a pony?

-dale