BTS: A carefully considered request.

[Formerly Babra]

This is neither rant nor grogification, just a request based on a repeated observation:

I have noted the tendency of bailed crews to do two things after exiting their vehicle. 1) To "move" to cover; and 2) to move to the nearest cover.

On the surface, this makes sense. I'm willing to accept that they don't run, since they could very well be assisting an injured comrade. However, it too often happens that the nearest cover is toward the threat that just killed the vehicle. Even when I cancel the a/i generated movement order and issue a move/run/crawl or whatever command toward cover away from the enemy, the a/i immediately countermands my order and resumes movement toward the nearest cover, and consequently into disaster.

The problem is that they are invariably picked off by the enemy's supporting infantry concealed in that same cover. I've lost a fearful number of crews this way.

It is my opinion that, except in cases where cover is VERY close, bailed crews ought to be heading back to their own lines. The chances of a crew moving across open ground toward the rear being shot down are, I think, significantly reduced if the enemy is shooting at other units moving toward them. If the enemy has no other targets anyway, then the crew's chances aren't hurt any by moving away.

I really think this would bump up realism one more tiny notch.

------------------

There are two things you should never do:

Never get involved in a land war in Asia; and

Never try to out-kill a Sicilian!

[von Lucke]

Yes, this has been mentioned before (by myself and others). The TacAI likes to bolt for the nearest cover when under fire --- irregardless of whether or not said cover is occupied by the enemy.

Agreed, this could use a little tweaking...

[Jarmo]

I've lost a fearful number of crews this way.

I've lost fearful amount of schrecks this way

------------------

Now, would this brilliant plan involve us climbing out of

our trenches and walking slowly towards the enemy sir?

[Guest Germanboy]

Babra, the title of your thread makes for a pleasant change.

Part of the problem here may be that the knocked out vehicle does not give cover behind which the crew could hide (not directly, but have it between itself and OPFOR as shield), so they dash for the closest cover in game terms, which may just be where the tank-killer is waiting with infantry back-up. I can also imagine that there would be complaints if cover is close-by, and the crew starts legging it over open terrain b/c that is how the map looks like. But in principle you are right, I guess.

------------------

Andreas

[Chupacabra]

See, problems like this would be completely eliminated by my IGO/UNOGO system.

------------------

Soy super bien, soy super super bien, soy bien bien super bien bien bien super super.

[WendellM]

Babra,

That's a nice, well-written suggestion. I know that I'd tend to run away from whatever just blew up my vehicle (unless cover toward it were very close by). I also would like to see this checked out (whenever BTS gets a chance).

[TankDawg]

This goes back to the problem that vehicles do NOT block LOS. The first logical place the crew would run (unless the tank is blazing) would be behind it to catch their breath. Then they would retrograde back to enemy lines.

I think all crews should just automatically rout off the map once their tank is toast. Too many of them are being used for way too many non historical uses.

"Hey, you crew, go scout ahead of our company and see who is in that village."

Jeff

[Jadayne]

would a possible solution be to have the tank crews that are not routed or panicked to go into a "hiding" stance directly behind their vehicle toward the friendly side? The protection from the "hiding" stance would mimic (slightly) them taking cover behind their vehicles. Then the player could move them wherever he wanted on his next turn.

[iggi]

Even HQs do this. They run towards the nearest cover which might have enemy there. Even leaving thier squads behind them.

If the HQ thinks that was a good idea, why did it leave it's squads behind?

[Guest Michael emrys]

I would hope that the routing priorities would be rewritten so that units (of any sort) would rout towards the nearest cover (1) away from the nearest confirmed enemy unit in its LOS, and (2) towards the friendly map edge as long as that does not violate (1).

Michael

[This message has been edited by Michael emrys (edited 08-17-2000).]

[L.Tankersley]

<BLOCKQUOTE>quote:</font><HR>Originally posted by Germanboy:

Part of the problem here may be that the knocked out vehicle does not give cover behind which the crew could hide...<HR></BLOCKQUOTE>

This struck a thought in my head. I know people have asked for vehicles to block LOS and provide cover to infantry and the response has been that it's too difficult/expensive to do the calculations, and a single vehicle is probably insufficient to shield a squad anyway if they're spread out. ::makes note to check FAQ workshop for this question::

But how about a compromise? Say, infantry within 10m (or whatever distance) of a vehicle/wreck (or even a stationary vehicle/wreck) receive some amount of cover so they're not considered to be in "open ground." This shouldn't be great protection for several reasons (doesn't protect from all sides, men are probably spread out), but might be roughly equivalent to "scattered trees" without the airburst potential.

Advantages of a compromise like this:

* Wrecks (and maybe vehicles) now provide some amount of cover, hopefully making crews a bit less likely to charge off to their deaths.

* Makes it possible to conduct an "armored assault" with infantry gaining a small degree of protection by advancing with armored vehicles.

* Processing cost should be modest compared to doing LOS evaluations - this could be another part of the standard terrain check, if a target is not located in hard cover (foxhole, trees, building) and IS located within 10m of a vehicle, give a small bonus to cover percentage.

* No more an abstraction than many other things that go on in proximity to vehicles (e.g. close assaults).

Something for consideration, anyway.

------------------

Leland J. Tankersley

[Guest Germanboy]

<BLOCKQUOTE>quote:</font><HR>Originally posted by L.Tankersley:

This struck a thought in my head.<HR></BLOCKQUOTE>

I response to something I wrote?! Wow, that's more than I ever achieved with my undergraduate students.

I like the idea, not sure how practical it is, but I like it.

------------------

Andreas

[tss]

Germanboy wrote:

I response to something I wrote?! Wow, that's more than I ever achieved with my undergraduate students.

If someone knows a fool-proof (or should that be student-proof) method for getting some response out of undergraduate students, please tell it to me.

- One who has to teach computation theory to 300 future electric engineers who are _not_ interested.

[Guest Michael emrys]

The best way to engage the interest of students (at any level) in a topic is to first show how it is related to something they are already interested/curious about. If that doesn't fire them up, suggest a different career path.

Michael

[David Aitken]

L.Tankersley wrote:

> Say, infantry within 10m (or whatever distance) of a vehicle/wreck (or even a stationary vehicle/wreck) receive some amount of cover

I do agree that vehicles should provide some cover, but don't think it would be a good idea to abstract it to this extent. For example, you could advance men in front of a vehicle, in order to flush out AT crews - but they'd get the same cover as though they were behind, using the tank for protection. A lot of things in CM are abstracted, but they're always the kind of things that make sense to abstract - whereas the shortcomings of this idea would be all too obvious, and would also encourage gamey behaviour.

I might also point out that this doesn't escape the problems which led Charles to omit vehicle cover in the first place. In the same way that the path of a projectile can't currently be traced in relation to units other than the target unit, it would also be the case that the cover afforded by a unit can't be calculated dependent upon the position of another unit. If this were possible, then vehicle cover itself would be too.

It's a pain being unable to advance men in the cover of tanks, since this was such a common practice in reality - but if it's going to be done, it needs to be done properly. Until CM2 (or whatever) we'll just have to do without. =)

David

------------------

There's a splinter in your eye, and it reads REACT

[Dan Weaver]

<BLOCKQUOTE>quote:</font><HR>Originally posted by tss:

If someone knows a fool-proof (or should that be student-proof) method for getting some response out of undergraduate students, please tell it to me.

- One who has to teach computation theory to 300 future electric engineers who are _not_ interested. <HR></BLOCKQUOTE>

I'm interested, send it to me by email.

Dan

[L.Tankersley]

Originally posted by David Aitken:

I do agree that vehicles should provide some cover, but don't think it would be a good idea to abstract it to this extent. For example, you could advance men in front of a vehicle, in order to flush out AT crews - but they'd get the same cover as though they were behind, using the tank for protection. A lot of things in CM are abstracted, but they're always the kind of things that make sense to abstract - whereas the shortcomings of this idea would be all too obvious, and would also encourage gamey behaviour.

You're right that my suggestion doesn't take facing or direction of incoming fire into account; this is intentional, to avoid a lot of complex LOS calculation issues. I could argue that even being in front of a tank gives some small measure of protection because you're less likely to be silhouetted against the horizon, or because the guy shooting is likely to be distracted by the tank [let's not debate this; I'm not sure I buy it either].

I think that abstractions "make sense" when they allow you to reasonably model effects or results that would otherwise be too difficult to incorporate. Right now, vehicles provide no cover because it's too difficult to implement the effects realistically. My proposal suggests a way to abstractly implement these effects that (I think) wouldn't be prohibitively expensive. Much as CM models gyrostabilizers in US tanks as everpresent but less effective than in reality, I suggest modeling cover from vehicles as everpresent (within some range gate) but significantly less effective than actually being on the opposite side of the tank from incoming fire.

I might also point out that this doesn't escape the problems which led Charles to omit vehicle cover in the first place. In the same way that the path of a projectile can't currently be traced in relation to units other than the target unit, it would also be the case that the cover afforded by a unit can't be calculated dependent upon the position of another unit. If this were possible, then vehicle cover itself would be too.

Well, it depends how much of an abstraction you're willing to accept. If you use a simple range gate (target to vehicle) to determine whether to apply this cover modifier, then you don't need to check LOS from the firer to the vehicle or check whether the vehicle obstructs or lies close to the LOS (you could argue that you shouldn't use a point-to-point LOS because infantry squads aren't really point targets). My proposal would allow a vehicle to provide cover even if it were out of sight of the firer -- obviously suboptimal, but it keeps the computational requirements down. The only check you need to make is whether the target is within X distance of a vehicle (presumably armored, possibly friendly, maybe stationary), and my gut feeling is that this shouldn't be much more expensive than the LOS and C&C tests that are going on now during turn calculation.

If the radius of protection is small enough (and I think it should be quite small -- I just pulled 10m out of my hat, maybe it should be smaller) then I don't think there are all that many opportunities for taking advantage of this in a really gamey manner. I guess maybe you could drive a tank along one side of a ridge and have infantry moving along the crest get protection from firers on the opposite side, or park a vehicle just behind a crest to provide cover for troops ON the crest. Hmmm. The question is, is the possibility for abuse like this overmatched by the benefits?

Interesting discussion, in any event.

------------------

Leland J. Tankersley

[This message has been edited by L.Tankersley (edited 08-17-2000).]

[Guest Big Time Software]

The TacAI is programmed to run away from known enemy units. HOWEVER, checking for such a thing is CPU intensive. There is only so much that can be done here. For every meter away from the vehicle CM has to search (360 degrees mind you!) it takes exponential time to search the area. The game benefit does not outweigh the hit to the CPU at the moment. This could change in the future with faster CPUs, even though there is a whole list of uses for the extra processing power that are probably more important. Like...

LOS blocking for vehicles is NOT going to happen any time soon. It is a similar issue as above. Far too many CPU cycles needed to give back too little benefit.

Leland, your suggestion is not bad from a conceptual standpoint, but it doesn't work from a coding one. Units are not terrain as far as the code is concerned. To make a knocked out tank suddenly behave like a wall would be a non-trivial programming task.

Thanks,

Steve

[This message has been edited by Big Time Software (edited 08-17-2000).]

[Formerly Babra]

I'm glad my request generated such interesting and thought provoking responses. I'm the first to admit that implementing what I ask may be more difficult in practice than in theory. There will inevitably be those situations where the crew in question is technically surrounded for instance. Which way should they move then? Not so easy for the a/i to implement.

I still think the answer lies in DRPs, Designated Rally Points. These would be markers placed by the player like a Target Reference Point. Routed, broken, scaredy-cat units would then use these as a point to converge on when they don't want to hang out in the fight any more.

------------------

Sounds like 100% weapons-grade bolonium to me.

[Guest Big Time Software]

HQs act as rallying points. Use your Company HQ or Battalion HQ for this role.

Steve

[Formerly Babra]

Cool! Thanks.

------------------

Sounds like 100% weapons-grade bolonium to me.

[Guest Big Time Software]

One qualifier about the HQs as rally points...

Units *tend* to go towards HQs. However, if the unit is too far away, or simply Hell bent on going home, it will go for the nearest friendly map edge. Units also will not go towards a HQ if it means exposing itself to fire (i.e. crossing a big open field to get to the HQ on the other side). So this isn't fool proof.

A tip is to keep the HQ far enough to the rear so that you have time if you need to order it to move towards the fleeing unit. HQs can be moved around fairly quickly so they can act like a catchers mit to some degree. Move slightly left/right to catch the unit trying to get past it.

Steve

[Guest Mortiis]

Good juicy info like what bts just supplied should end up in the faq, I had no idea that routing units headed to their hq's

------------------

<BLOCKQUOTE>quote:</font><HR>History will be kind to me, for I intend to write it.

Winston Churchill<HR></BLOCKQUOTE>

[Hans]

Interesting discussion

I'd say the crew would do one of three things

1. Stay near the vehicle to recover crew/gear and shelter behind it, especially if taking small arms fire.

2. Head away from the fire/towards the rally point start line of the operation.

3. Panic and

a Head almost anywhere

b Break up and head almost anywhere doing of the the three above.

Another point is the experience of crew, especially the surviving the leader.

Of course it would be a pain to actually model this!

As for getting non graduate students to respond. I usually bribe one aggressive student in the class to hit the slow ones.

[L.Tankersley]

<BLOCKQUOTE>quote:</font><HR>Originally posted by Big Time Software:

Leland, your suggestion is not bad from a conceptual standpoint, but it doesn't work from a coding one. Units are not terrain as far as the code is concerned. To make a knocked out tank suddenly behave like a wall would be a non-trivial programming task.<HR></BLOCKQUOTE>

I understand they're not terrain; I was thinking you might be able to insert the code to check for proximate vehicles in or near the code that evaluates the terrain occupied by the target, and you could limit the scope further by only doing this check if the target was in open ground (because otherwise there is probably better cover available anyway). However, I humbly bow to your superior knowledge of the CM code.

------------------

Leland J. Tankersley