Some things that need fixing in CMBB (not in 1.01)

[Dirtweasle]

Originally posted by dalem:

I am forced to vehemently disagree with this line of reasoning: it's a relatively simple consequence of game engine mechanics going on, that's it....

Who's forcing you Dale? In any event - I agree the game is only doing what it's programmed to do, nothing more or nothing less.

Ridiculous on the face of it, no matter how many stories we could make up to justify it.

Pardon me all to hell for thinking out loud with an example. Geez Dale you Ok this morning? I'm not justifiying any set behavior, though it sure seems like 4-6 guys could shoo off a couple others if push came to shove. Sounds completly plausible to me. I thought I was just mentioning this as an aside anyway.

My central idea on this subject is that I'd rather see them head for the hole, then get right on top of it an notice it's full already, and as you say the good unit would not leave, and the routed/broken one continues to the next best cover to the rear.

This brings me to a point about CM with which I've always disagreed: actions of Routed units, or at least the perceptions of those actions. I've always found it difficult to reconcile the fact that Routed and/or Panicked units will, at times, run around crazily and illogically on the battlefield, even in the face of withering fire, with the fact that they remain in play as units at all. The most extreme panicked battlefield reactions I've read about, i.e. the running around aimlessly, the catatonia, the complete breakdowns and dirt-eating, are more appropriate to individuals, not units. If a squad loses a few guys, we all know that the little crosses can mean anything from death to hiding to running away - and in my mind, these are the guys who are panicking severely enough to get themselves hurt at times.

But the rest of the unit? As long as it's a cohesive unit it should be fighting, hiding, or moving away to cover, not charging around in some weird spiral. And in the game, even a Panicked squad, since it can Rally, is a unit.

I'd agree with much of that, it soumds reasonable, and it would be an added touch if somehow a routed portion split off and went to the rear. Since CM uses units and not indivuals there seems to be little chance of single soldiers busting out, but it would certainly be more realistic for the majority of cases.

[dalem]

Originally posted by Dirtweasle:

I'd agree with much of that, it soumds reasonable, and it would be an added touch if somehow a routed portion split off and went to the rear. Since CM uses units and not indivuals there seems to be little chance of single soldiers busting out, but it would certainly be more realistic for the majority of cases.

Oh, I'm not asking for any "half squad routs" or anything like that (although maybe that's a way to handle it), I just want my "units", whether Good Order, Panicked, or Routing to act as units, and if they're not units anymore, then disappear.

-dale

[Halberdier]

Dirtweasle sayeth:

I'd agree with much of that, it soumds reasonable, and it would be an added touch if somehow a routed portion split off and went to the rear. Since CM uses units and not indivuals there seems to be little chance of single soldiers busting out, but it would certainly be more realistic for the majority of cases.

I have been thinking of that too. But I'm not sure if the AI is capable of recombining split squads that had been split. Has anyone seen it recombine?

Did entire squads route together? Probably sometimes, but I also think that it can vary from between one man and the entire platoon. Another wish for the engine rewrite :cool:

cheers,

-gabe-

[Eden Smallwood]

Originally posted by Halberdier:

But I'm not sure if the AI is capable of recombining split squads that had been split. Has anyone seen it recombine?

The Book Of All Wisdom clearly states that Split Squads will recombine if you get them to Snuggle Up next to each other at the end of a turn. ISTR that Snuggling occurs at less than ten meters.

Yes, it works too- I've made it happen. Useful.

Eden

[Tarquelne]

The Book Of All Wisdom clearly states that Split Squads will recombine if you get them to Snuggle Up next to each other at the end of a turn. ISTR that Snuggling occurs at less than ten meters.

Yes, it works too- I've made it happen. Useful.

Eden[/QB]

It doesn't work with SS squads, though - they'd be shot for snuggling.

[Halberdier]

Eden the smartass said:

The Book Of All Wisdom clearly states that Split Squads will recombine if you get them to Snuggle Up next to each other at the end of a turn. ISTR that Snuggling occurs at less than ten meters.

Yes, it works too- I've made it happen. Useful.

Eden

That's not what I meant, and thought that was clear, but I admit being a bit too brief by habit and am sometimes misunderstood. I meant the AI's units. The non-human. The 'puter controlled side (and a gamey one at that). The not-me side. The side that I don't have any control over. The side that doesn't cuss back at me. The other side. The side that I can't see until it's too late. The side that I'm trying to clobber. The side that has my flags. The side that bushwacks me...

To restate: I have been thinking of that too. But I'm not sure if the AI controlled side is capable of recombining its squads that had been split. Has anyone seen it recombine its own units?

cheers,

-gabe-

[Broken]

I agree with Redwolf on restricting AI "traffic jam" movement to on-board. I have lost infantry squads in the same way. In one tournament game, a platoon moving line abreast up a map edge lost a squad when they stopped because the AI adjusted the position of the edgemost squad so that it moved off-map.

Redwolf has some good points on panic movement. The AI could make more intelligent (realistic) decisions for panicked movement without using "memory". The AI seems to consider such things as location of threats and potential cover, but in a very "binary" and inconsistant way. I don't know if the "panic AI" has changed much since CMBO, but here are some examples of what I have seen:

A. Infantry panicking a few meters from good cover will often turn around and run 80 meters across clear terrain to other cover.

B. Infantry panicking a few meters from leaving LOS of the threat will often turn around and run to cover a few meters from the threat, covering 80 meters of clear to get there.

C.Infantry panicking in good cover will leave said cover and "panic dance" in the open.

D. Infantry panicking near a building that is under HE direct fire will often run into the building, typically moments before it collapses.

E. Panicking infantry does not consider LOS shadows to be cover, unlike vehicles. Stone walls and reverse slopes are not considered to be cover, it seems, for panicked troops.

F. Infantry panicking on the fringe of an arty pattern will run to the center of the pattern if the cover there is better.

Admittedly, D and F are fairly complex situations. What seems to be missing from the other situations is a better balance of threat vs distance and the use of LOS information. For example, a location out of LOS should be considered good cover.

[Battlefront.com]

Gentlemen... a quick reminder that "simple" behaviors does not mean simple coding. It *might*, but never, ever, EVER assume that because you find something simple that it is something that Charles could whip up in a couple of minutes, hours, or even days.

With 9 years of development experience with about a dozen programmers (some good, some horrible, and Charles ) I have long since lost any illusion that I am capable of knowing what is simple, complicated, or impossible. I am often correct, but I always go on the assumption that I don't know what the F I am talking about. For those of you who have never made a wargame, I assume you think along similar lines

I can tell you one thing for sure... the Broken/Panic/Routed behavior of flittering back and forth between two pieces of terrain is not a purposeful action. It is one that we tried to code out of existance as a matter of fact. But because units have no memory (see previous description of this on Page 1) we can not. Personally, this is probably the #1 thing I want to see fixed with the new engine. It will not be fixed with CMBB.

Also... remember that even truely simple changes, ones that take Charles only an hour or two to code, can sometimes have massive unanticipated consequences. Take one of the three recent suggestions from Redwolf for example. Charles might look at one and say "ah, piece of cake. I'll add it". But then something totally unexpected happens to some other behavior. Then he tweaks that and something happens to a different behavior. Then so on and so on. Suddenly that "quick fix" is a week's worth of coding and a month's worth of testing.

Bottom line... if it has to do with AI, any assumption OTHER that even a simple request might be practically impossible is just a bad way to think about what we can/can't do within the current constraints.

That having been said, Charles will look into this when he gets back.

Steve

[dalem]

Steve-

One thing I thought of tonight involves a piece of BB that was obviously deliberately added, and therefore has already passed muster with you and Charles, and the beta crew. Better MGs I like, more brittle infantry, in theory, I like. But the frequent automatic resetting of my Advance/Sneak/Move orders forward by the TAC AI to Sneak orders backward when under fire is, tactfully speaking, helping to ruin CM:BB for me (maybe not for anyone else). It's not just that I have to now check each face-down unit at the beginning of each turn, which is boring enough (for me); it's that, since I cannot simply drag the Sneak order point a full 180 degrees about in order to keep units under fire moving forward, I am forced to Halt my unit and give it fresh Sneak/Move/Advance orders in the direction I want it moving, with all the attendent command delays associated with a new maneuver. Likely as not, the unit will need the same treatment next turn, assuming it doesn't simply Break under fire because it's sitting stationary contemplating it's "new" orders.

Maybe it's impossible to simply allow Sneaking waypoints to be dragged fully 180 degrees, even with poor quality units - I don't know, not my code, not my expertise. But this effect, while no doubt realistic in its own way, is one of the things that is quite frankly killing the game for me personally, despite my attempts to learn to live with it. I hope this comes across as constructive, as I'm not trying to be a whiner, but if I want my troops at point X, then darn it, I don't care if they have to get there by crawling for five turns, I want them to keep trying as long as they are in Good Order, without being babysat.

Off the top of my head, I see two main elements "missing" from the attacker's arsenal which would be necessary to truly get the Attacker:Defender balance "to where it should be":

1) True Area Fire. I.e. designating an Area to fire into per weapon/unit, not simply a point to fire around. This would allow maximally efficient use of MGs and other suppressive fire on suspected and known enemy positions, and therefore allow more covered advances by the attacker.

2) More specific artillery requests. I.e. designating a smaller number of rounds for smoke screens and the like. With only 150 rds of 81mm mortar fire, I am loathe to use it for smoke screens, not because smoke is a bad idea tactically, but because I can "waste" a large number of rounds stoking a screen that is already "sufficient". Here I feel I am on much weaker ground because I have a relatively poor understanding of artillery doctrine - maybe they just asked for "smoke" and got 60 seconds of smoke rounds, I don't know. But it doesn't "feel" efficient.

These sorts of things are not "patchable", I know. But they are indicative, to me, of how the balance has swung away from the attacker, where it squarely rests in BO, over to the defender, where it now rests, maybe not as squarely however, with the defender.

-dale

[ October 27, 2002, 04:57 AM: Message edited by: dalem ]

[CMplayer]

Originally posted by dalem:

1) True Area Fire. I.e. designating an Area to fire into per weapon/unit, not simply a point to fire around.

This is a really good idea. How many times would you like a tank to put series of shells into the length of a treeline, for instance? Realistic, useful, and missing from the game.

[Cameroon]

Originally posted by CMplayer:

</font><blockquote>quote:</font><hr />Originally posted by dalem:

1) True Area Fire. I.e. designating an Area to fire into per weapon/unit, not simply a point to fire around.

This is a really good idea. How many times would you like a tank to put series of shells into the length of a treeline, for instance? Realistic, useful, and missing from the game.</font>

[Renaud]

It seems to me that this whole thread highlights the conflict between two major philosophies of design. I'm going to TRY to be objective in my description of the two camps even though I am certainly on one side. Note that splitting people into two camps is an oversimplification and not literally true, yet I think it is a useful 'thought-experiment'.

One philosophy stresses a high level of control over events, expressed in the various desires to have your men do more specifically what you want them to: lob an exact number of shells, cover an exact patch of ground, go where you tell them until they are all dead. This group could be called the 'chess-game' or 'micro' camp. Events such as squads panicking and bouncing around out of control under MG fire like pinballs in a pinball machine annoys these folks!

The other philosophy advocates an abstracted level of control on the assumption that commanders of CM-size forces in actual engagements have at best limited control over events. CM of course allows us far more than that on the assumption that squad leaders have initiative and our commands as players model this initiative as well as directives from HQ units. This camp could be called the 'realism' or 'macro' camp. This group tends to be more forgiving of quirkly tacAI behavior on the assumption that you can reasonbly expect a lot of apparently crazy behavior from men under fire.

What camp Steve and Charles fall into is not known to me. Due to apparent limitations of current code base, the game is currently more amenable to the 2nd, or 'macro' control, camp. That's why we mainly see folks from the 'micro' group posting concerns and lobbying for greater level of control and more predictable behavior. Note that here I am considering the thoughtful 'micro' people, not the larger crowd of 'shoot-from-the-hip' folks who immediately raise an outcry when they see something they MUST HAVE 'fixed' based on one odd event or lost game (ie 'control-freaks').

Anyway, food for thought I hope.

Ren

[Redwolf]

Ren, I don't think your points apply here, although they are generally useful. But what we have here is that the player is forced to have the detailed exposure of his men making their own movement, down to a detail level which is pretty tiny. At this detail level the game is broken for a number of situations (IMHO).

So, the "high-level" gamers cannot avoid being exposed to these details, even if they wanted to, and the "control freaks" cannot help the control being taken away from them.

That made me complain, I honestly believe the current system doesn't please anybody. As I said, it hurts defense in depth most, which is a tactics it should rather reward.

[Michael Emrys]

Originally posted by dalem:

True Area Fire. I.e. designating an Area to fire into per weapon/unit, not simply a point to fire around.

You mean like combining area fire with the lasso feature? Other than possible difficulties in programming, I do not immediately spot any objections to it.

Michael

[Eden Smallwood]

Originally posted by Michael emrys:

</font><blockquote>quote:</font><hr />Originally posted by dalem:

[qb]True Area Fire. I.e. designating an Area to fire into per weapon/unit, not simply a point to fire around.

You mean like combining area fire with the lasso feature?</font>

[dalem]

Originally posted by Michael emrys:

You mean like combining area fire with the lasso feature? Other than possible difficulties in programming, I do not immediately spot any objections to it.

Michael

That is absolutely what I mean. I know it can't be done with the current engine (I asked) but it is something I sincerely hope can be done with CMII. I think it is a critical feature.

-dale

[JonS]

Originally posted by dalem:

[...]

2) More specific artillery requests. ... Here I feel I am on much weaker ground because I have a relatively poor understanding of artillery doctrine - maybe they just asked for "smoke" and got 60 seconds of smoke rounds, I don't know. But it doesn't "feel" efficient ...

AFAIK, no army has ever used a method of fire control based on multiples 60 seconds worth of fire Fire control orders are always given in terms of the number of rounds to be fired. The main alternatives I am aware of are "x rounds FFE" and "y rounds at z rounds per minute"

Examples:

"12 rounds fire for effect" means each gun under control (could be a battery, a battalion, or a corps, doesn't matter) fires 12 rounds as fast as they can. For lighter calibres, this could be in less than a minute. For larger calibres it will be longer, but the point is that the FO wants the rounds as fast as they can be delivered. This is useful for quickly neutralising targets that are currently affecting your operations.

"12 rounds fire for effect, rate 2" again means each gun under control fires 12 rounds, but this time each gun fires once every 30 seconds, giving - in this case - six minutes of controlled fire. This is useful for preventing known or suspected enemy units from being able to affect your operations.

As for how CMBB handles it ... *shrug* ... Hopefully its one of the things that will get some attention during the [angelic voice]The Rewrite[/angelic voice].

Regarding units switching to sneak when under fire. I've already posted in the other thread on this - my preferred solution would be to see all troops, once they decide to go to ground, go to ground and stay put. The quality of your troops would have an effect on this - in a nutshell; conscripts and greenies would stay upright too long and get cut to shreds*, regs and vets would go ground sooner, crack and elite would press on but go to ground once they started taking cas.

The type of fire would ideally also have some affect: if it was arty, everyone goes to ground, if it was MG or small arms then the above 'rules' would apply.

The point about going to ground is that once down the units morale isn't going to change much**, and when the next orders phase rolls around the player (assuming they aren't panicked or routed) can either order them to return fire from where they are, send them to cover in a location that makes sense to the player***, or give them orders to get moving forward again****. Either way, its up to the player, but he doesn't have to contend with auto-exhausting units on top of everything else.

But, ... I'm not a programmer. I don't know how hard this would be, or what impacts it would have on other parts of the game.

Regards

JonS

* In Real Life™ new troops had a tendency to feel 'invincible', making them useful to callous commanders for assaults. The literature makes reasonably common reference to this. An example off the top of my head is Hastings D-Day: Operation Overlord and the Battle for Normandy.

** At least, not as much as having them running around in the open, or crawling hither and yon which just gets them exhausted. IIRC, being tired or exhausted does bad things to a units morale.

*** Thereby avoiding all the murky decisions that the AI can't handle very well at the moment, including moving to positions out of LOS but otherwise in open ground.

**** Yes, this would mean incurring the command delay, but in this case, IMHO, that's not such a bad thing.

[ October 28, 2002, 03:50 AM: Message edited by: JonS ]

[Sirocco]

Originally posted by Vanir Ausf B:

Situation: my platoon is advancing over about 70 meters of open ground to a stand of scattered trees. They come under fire from a machince gun several hundered meters to their flank. They go to ground and start sneaking towards the objective. About 5 meters away from the cover of the trees one of the squads goes to panic status and turns around and starts sneaking the other direction, over the same 70 meters of open ground they just covered.

I think that's probably the most often seen problem with the TacAI in CM. Your squad or team is almost on top of solid cover, about to break LOS with the threat, and they turn 180 degrees back into the withering fire they want to escape from.

Not having seen the code I can't of course know for sure, but adjusting the code to handle the above situation should be possible in a patch. The cover at the destination, and the distance to it, has to be factored more heavily into the TacAI's relevant calculations.

Edited to add: I'd like to see panicked and routed units stay put more often. When a unit is in good cover, for example a foxhole, I'd expect fire that wasn't causing casualties to pin them in place, even surrender, rather than have them jump up and sprint back across open ground.

[ October 28, 2002, 06:55 AM: Message edited by: Sirocco ]

[Pascal DI FOLCO]

The sneaking/squatting infantry is really a pain in the @ss ... It did not occur in CMBO, now we're stuck with guys sneaking everywhere !

I can't clearly see why it cannot be fixed :confused:

Simple rules could manage the problem :

* Units that TacAI makes seek cover should crawl away from enemy no more than 20-30m, then Run

* Occupied spots should not be considered as "possible" shelters by such units

Not doable ?

[dalem]

Originally posted by Pascal DI FOLCO:

The sneaking/squatting infantry is really a pain in the @ss ... It did not occur in CMBO, now we're stuck with guys sneaking everywhere !

I can't clearly see why it cannot be fixed :confused:

Simple rules could manage the problem :

* Units that TacAI makes seek cover should crawl away from enemy no more than 20-30m, then Run

* Occupied spots should not be considered as "possible" shelters by such units

Not doable ?

Well Pascal-

The occupied foxholes thing is probably doable. I would think. But according to some of the things Steve has posted, some of the behavior we see now is a result of the units having no "memory" of where they came from or where they're going. If that's true (and why wouldn't it be?) then it simply may not be possible for a unit to do a running tally of where it's supposed to be vs. where it is. So we may be living with this for awhile.

Then again, lots of folks appear absolutely nonplussed by it (the need to reset lots of movement orders) and feel it's something that can be well-addressed by simple changes in tactics (i.e. more suppressive fire and cover). For them it's simply not the gamebreaker that it is for me. So either way it looks like we have to learn to live with it for a while at least.

-dale

[Pascal DI FOLCO]

dalem,

Note that the "fixes" I suggested made NO use of any "memory", they just needed to statically assess the situation : the hole is/isn't occupied, and the sneak order I propose is 20m long and that's all.

I agree that replanning the move is no great deal, but why would we have to do that in CMBB where it was OK in CMBO ?

[TSword]

Hehehe Steve,

It seems that your little Silicon Soldiers really have a mind of their own already..... I'd only wonder what might happen when they receive memory and can learn, my God (Hopefully you implement an RAM-allocation limit to keep them down, otherwise..) !!?

Maybe all Players should first disable the Powermanagmentfunctions on their PCs and lock the BIOS prior to installing of CM3...hehehe, or otherwise the little devils will take control of the whole PC !

Greets

Daniel

[dalem]

Originally posted by Pascal DI FOLCO:

dalem,

Note that the "fixes" I suggested made NO use of any "memory", they just needed to statically assess the situation : the hole is/isn't occupied, and the sneak order I propose is 20m long and that's all.

I agree that replanning the move is no great deal, but why would we have to do that in CMBB where it was OK in CMBO ?

Well, I agree about the foxholes issue but I think it was a conscious design decision to introduce the "replotting" by cautious troops, or, if it was always in there, then other infantry modelling design decisions have made it more noticeable in BB. But I also think that since it's in there "on purpose", so to speak, it ain't comin' out. But those are just assumptions and guesses on my part.

-dale

[Redwolf]

Dale, I find your comments mildly distracting.

A number of people in that thread proposed solutions which very clearly do not use any memory in TacAI units, and there is no purpose spreading FUD around it.

[dalem]

Originally posted by redwolf:

Dale, I find your comments mildly distracting.

A number of people in that thread proposed solutions which very clearly do not use any memory in TacAI units, and there is no purpose spreading FUD around it.

Umm, okay. I guess I'll stop commenting then. So sorry to bother you.

-dale