MG team recommendation

[Maastrictian]

Michael emrys wrote:

<BLOCKQUOTE>quote:</font><HR>

Actually, Maastrictian, I'm afraid that your industry in producing the calculations (which I removed for the sake of brevity) was founded on a misconception, namely that the gun barrel is pointed parallel to the sight line so that the bullets immediately begin to fall below it. If that were the case, the gun would only be effective at very short ranges. Instead, all guns are aligned so that the barrel points somewhat above the target, the angle above increasing with range to compensate for bullet drop.

<HR></BLOCKQUOTE>

You are of course correct (I even mentioned this was an assumption I was making in my post). I also seem to have underestimated muzzel velocities (as other's point out). But given air restance I think my calculation for a muzzel velocity of 500m/s is a reasonable lower bound for the maximum distance grazning fire is effective at.

In anycase, regarding elivating the barrel. It should be noted that by elivating the barrel you no longer have a "line" of fire but rather two lines.

(the dots below are placeholders and should be thought of as blank)

.............--------...............

........----/........\----........ head

mg ----/..................\----...|| body

-------------------------------------ground

...|........|........|.........|

....effective..........effective

............ineffective

Especially considering that the men under fire are likely crouching (and so presenting a smaller vertical target) fire that affects a group of men at (say) 300m may miss men at 150m. And the further away a group of men are the smaller the line of effective fire is (because the arc the bullets are taking is higher)

I'm not sure how much effect this has. I don't have time to do the math now (which is not terribly nice), but if I get a chance durring the day I will.

[Wreck]

Steve: if I understand what you are saying here correctly, you are pointing out that during any given LOS check the only units which count for anything are the firer and target; that is, that other units simply don't enter in to it.

That is the same understanding as I had about LOS when I proposed my idea. I want to clarify that my proposed solution for grazing does *not* assume anything other than the LOS that exists. It does use it somewhat more, which I see (from you and tom_w) may be more of a problem than I thought.

In the abstract, what I proposed was that every time a burst is fired, first angles would be checked to each enemy unit, then those within a critical angle get LOS checks. If both checks are passed, then the unit is grazed.

Contrary to what Stellar Rat said, I do not think the angle check needs to be very costly. Yes, if you want to know the exact angle between two vectors, you need to use a trig function (arccos), and sqrt, which are not cheap. But we don't need the exact angle; all we need is to know whether or not the angle is less than some particular angle.

[Note: I had a proof of this here, but this bboard does not seem to like inequality signs and so I removed it. If anyone wants it just email me.]

Summary for those that hate math: the vector math is not bad. It should be quite fast on any CPU of pentium class or better. Checking angles to all of say 100 enemy units should be a reasonable thing to do.

Also note that if one is clever, one can cut off a lot of angle checking simply by keeping track of the last angle result, and using it if neither unit has moved. The same idea would also work for LOS, and in fact is likely already implemented. So most likely, for any given MG burst only a small subset of enemy units would need to be checked for angle, since by and large most of them would have been checked in the last burst and neither the MG nor they would have moved.

The LOS check, though, appears to be more costly than I had hoped. About the only thing to say about that is, that for most circumstances the number of units within the critical arc for any given MG should be pretty small. So I am not really proposing adding that many LOS checks in the vast majority of cases. Perhaps one or two extras per average MG burst; something like that.

Whether or not that is affordable, well, that's BTS business.

[ 04-17-2001: Message edited by: Wreck ]

[ 04-17-2001: Message edited by: Wreck ]

[ASL Veteran]

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

Thanks for the informative read ASL Vet. I am curious when the FPL-Grazing Fire of MGs would be applicable in CM terms. Would it only be possible in defensive, prepared positions, ie stationary from setup? Or could a platoon of MGs prepare these fires in short order after moving to new positions, ie in a meeting engagement?

Ron<HR></BLOCKQUOTE>

Well, in game terms I think you should be able to set one up without having prepared positions. In real life it would require a man walking the FPL to find dead space, but in the game you could just look at the map to determine the dead space - there is no need to have a man 'walk' the FPL. Interesting question though. Even in real life you could always position the MG to perform grazing fire, it is just that if someone wasn't walking the FPL you wouldn't know where your dead space was. That wouldn't make it any less effective if no dead space was present though. Walking the FPL in CM is performed by the LOS tool though. I guess it would just depend. Personally I wouldn't see any reason why it would require prepared positions to be effective.

[Marlow]

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

Michael emrys wrote:

You are of course correct (I even mentioned this was an assumption I was making in my post). I also seem to have underestimated muzzel velocities (as other's point out). But given air restance I think my calculation for a muzzel velocity of 500m/s is a reasonable lower bound for the maximum distance grazning fire is effective at.

In anycase, regarding elivating the barrel. It should be noted that by elivating the barrel you no longer have a "line" of fire but rather two lines.

(the dots below are placeholders and should be thought of as blank)

.............--------...............

........----/........\----........ head

mg ----/..................\----...|| body

-------------------------------------ground

...|........|........|.........|

....effective..........effective

............ineffective

Especially considering that the men under fire are likely crouching (and so presenting a smaller vertical target) fire that affects a group of men at (say) 300m may miss men at 150m. And the further away a group of men are the smaller the line of effective fire is (because the arc the bullets are taking is higher)

I'm not sure how much effect this has. I don't have time to do the math now (which is not terribly nice), but if I get a chance durring the day I will.<HR></BLOCKQUOTE>

OK, I looked it up last night. For a 5.56 mm bullet fired from an M-16 (which does not have as flat a trajectory as a 30 cal. bullet):

With the rear sight set for long range firing, the bullet rises to a maximum height of 11 inches above the shooters line of sight to the target at 225 meters, and crosses the line of sight again at 375 meters. With the rear sight set for short range firing, the bullet rises to a maximum height of 5 inches above the shooters line of sight to the target at 175 meters, and crosses the line of sight again at 250 meters.

[JasonC]

ASL Vet, I know all about MGs and how they really work, I made MG defense plans for U.S. artillery battery positions, and had them set up on exercises, OK?

But a lack of the complexity involved in all of that is not the problem in CM today. Is there such complexity? Yes. Does CM lack it? Yes. Is that lack the cause of infantry's ability to rush MGs in CM? - *no*.

That one "approximation" or abstraction exists, does not on its own mean that said abstraction, is the cause of some unrealistic result. And lack of fire lanes is not what prevents MGs in CM from denying open ground areas.

I can prove it. It is simple to understand. The squads run right through the fire lanes.

MG firepower in CM does not stop a squad from crossing the MG's lane of fire. You can set up all the interlocked lanes you want - all you will do it force the attacker to cross an MG fire lane. Where it will receive the fire of a CM MG team. *Which will not stop that squad*.

MG teams do not have the firepower to prevent movement in CM today, even in open ground. They do not have the firepower to prevent movement across long stretches of open ground when the target is *continuously* under their fire. In your fire lane solution, they would have to *cross* a lane somewhere. But right now, they can run clear *up* one, *lengthwise*, and do fine.

In order to have fire lanes that actually prevented anyone crossing them, you would not only need the fire lane. You would also need to have crossing that lane be much more dangerous to a moving squad, than it is now.

Clear? Just lanes, will only mean squads hopping merrily through lanes. Near invunerability to MG fire in the open, is quite enough to defeat your hoped-for realistic effects, from all the complexity of fire lanes.

And that shows that it is a misdiagnosis. The problem is not that the effect of MGs cannot be shown abstractly without fire-lane complexity. The problem is that the effect of MGs in the open, flat is not there. Regardless of how the MG comes to apply its firepower to the target. The result of those firepower applications, simply leaves the squad less hurt, than it would really be.

Now, there is no way to fix this actual problem, without increasing the impact that a firing MG team does to a squad that is, for whatever reason (normal, lane, whatever), hit by it while moving in the open.

Notice also, that solutions that split the fire of the team over many targets also do not solve this problem. The problem is not that the MGs can't hit "all of them". The problem is that they can't hurt even the ones they do hit. If you sprayed the same firepower over 3 units, you'd muss the hair of 3 units even less than you hurt one now, and would not come close to stopping a rush.

But any change to the effective firepower can cause other realism problems. Whether it causes other problems, depends on how this increase is brought about. There are not that many factors that go into combat resolution, so there are not that many ways to alter the current set up.

1. - you can raise MG firepower, everywhere. This makes them uberweapons, because it unduly increases their effectiveness against non-moving units, units in good cover, etc.

2. - you can increase rate of fire. If you do this everywhere, same result as before plus a lowering of ammo, that is all. If you make this range dependent (short), then MGs cease to be longer-ranged weapons compared to squads, and become just squads with fewer men and slower speed.

3. - you can put in firelanes, hoping that they will often catch 2 units at once and thus effectively increase their firepower. But the firepower on the units hit, is unchanged. They again, waltz through the firelanes.

Aside - notice, however, that this "super realism" solution attempt, predicts 2x overall firepower with decent MG lane placement, to "skewer" two targets at once. Skewering more than 2 will be rare, especially if the attacker uses loose formations. Sometimes 3, sometimes 1, often 2.

4. We've covered fp, we've covered shot frequency, both overall and range-varying, we've covered targets hit (spray with same fp doesn't help as already noted; spray with fp applied to each is like "lanes" above). Obviously quality still needs to vary, whatever solution we come up with. What else is left that effects the outcome of CM shots? Only one thing - %exposed ratings, aka "cover".

Therefore, to solve the problem of MGs not killing the squads they do manage to shoot during rushes, the factor that needs to be changed in % exposed, or something like it. What are the advantages of this approach, compared to the others?

1. it allows the effect to depend on whether the target is in the open. Cover stops enfilade fire, so this matters for lane realism. It also matters to avoid uber-weapon results If the movement-exposed link and delay is also included, then it further allows the effect to depend on the targets movement state. Crossing a fire-lane is dangerous in reality, because each man in the squad faces attack as he does so. This effect is not seen for non-moving troops, most of whom are out of the fire lane on this side or that. Open and movement related effects also do not alter the range role of MGs vs. small arms.

I saw one potential problem with a solution geered only to the movement state and the cover, open or not. That potential problem is an unrealistically high effectiveness for MGs, engaged in mere "snapshots" at any desired patch of open ground, especially ones a long way away, that in reality they could not cover simultaneously. This is somewhat related to lanes, in that an MG cannot be sighted for good lane fire on a dozen areas at once, without adjustments etc.

Thus, it seemed to me the missing factor could be captured by time of exposure. Instead of the MG transitioning to much higher firepower against just any moving target, or just any target in the open, let it get that bonus only if the target is exposed long enough to shot at it repeatedly. This discriminates wide areas of open ground from small patches of it.

The total output of fire can double in my proposal. The total output of fire with lanes, if the AI picks them reasonably well, would not be appreciably different, since usually the MG would succeed in skewering 2 targets at once in a wide open area. Indeed, the effect can be viewed, if you like, as skewering two portions (or 25% 1, 50% 2, 25% 3, etc) of the same unit shot at.

There is no need to burden the tac AI with the selection of fire lanes. And by having the increased effect hit one unit, it might actually do something to that unit, which MGs at present do not do, and would not do with just lanes.

Also, if its increased effect drives a unit to the ground, then the ordinary target-selection procedure of the AI will incline it to pick a different target for the next burst, if all the others are still closing - since it favors the closest target. Thus, if the increased fp is enough to pin a unit with a few shots, then over a couple of minutes the MG ought to engage a couple of targets, and with higher average effect than today.

Which will tend to stop rushes. Which is the point. Adding complexity is not the point. And lane complexity will not stop rushes on its own.

The kinds of effects you would see, would be akin to what lanes and skewered targets and the difficulty of crossing a lane of fire in the open, actually do in reality - multiple the firepower of MGs somewhat, and make them especially effective against units moving in the open.

But without the player needing to worry about such minutae. And without the tac AI, having to be programmed to solve such minute and complicated optimization schemes on the fly ("skewer this one and that one? Or that closer one, and maybe this one if it keeps moving?" A programming nightmare).

Once again, the problem is that squads walk through MG fire in the open. Not that the MGs don't get to fire at them because they can't spread their fp over tons of targets - because the present firepower can't hurt one target let alone three. And not that they can't lay down fire patterns - because the squads would just walk right through those fire patterns, the same as they walk right through the direct fire now.

The solution is to increase the firepower of MGs against troops moving in the open, a cover like, exposure effect in CM's present scheme of things. Such effects can be easily modeled in CM and do not require large increases in tactical complexity, either to the player or to the tac AI.

The MG gunner's ability to skewer two targets sometimes, simply shows up as doubled firepower against the target shot at, in the sorts of situations in which the MG gunner would be able to pull it off. The complexities of the MG gunner's job are his own concern, are handled abstractly by an approximate overall result, and variance in it is just packed off to the combat resolution "die roll".

[JasonC]

Originally posted by Wreck:

[QB]I am a bit surprised by this.

If you guys think it is hard to impliment the effects of grazing fire once the direction is settled, wait until you get to the AI, and get to try to backsolve the expected results over all possible directions of fire, including the effects of targets currently in motion, in order to pick the angle of fire.

For every MG. For players solving such problems would be an interesting exercise in competitive draftsmenship and architecture, but for the tac AI it would be some minimum principle calculus of variations do-hickie in a spikey peg-board phase space of angle choices. It is ridiculous.

And unneeded, because the real problem with the way rushes happen now is a cover effect, not a fire-lane effect, and the role of firelanes can be abstractly and on-average included in the solution. See my previous response to ASL Vet.

[Michael Dorosh]

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

ASL Vet, I know all about MGs and how they really work, I made MG defense plans for U.S. artillery battery positions, and had them set up on exercises, OK? <HR></BLOCKQUOTE>

Wow.

[Username]

Is JasonC related to Rexford?

I notice the same long winded self serving style in his posts. Also the always lovely "follow my own post with yet another important post by myself".

BTS has said all there is to say:

1. No fix for CMBO

2. THEY are going to show you the fix for CM2.

Cant wait.

Lewis

[Battlefront.com]

Wreck wrote:

<BLOCKQUOTE>quote:</font><HR>In the abstract, what I proposed was that every time a burst is fired, first angles would be checked to each enemy unit, then those within a critical angle get LOS checks. If both checks are passed, then the unit is grazed.<HR></BLOCKQUOTE>

This is a manditory part of any type of intersection calcuations. If you don't give boundaries then the system will just check every location on the map I think you are just woefully underestimating how difficult it is for a CPU to crunch these sorts of numbers. Because you are not a programmer (and neither am I, just an experienced developer) you are just going to have to trust me that there is absolutely no easy/slick way around the problem. We will see what we can do but I very much doubt we will be able to do anything more than an abstracted solution.

Jason,

I generally agree with your basic thinking, but let me emphasize that there are at least three inter related problems as I see it:

1. Run is too fast for what it is supposed to simulate. Time is one of the worst enemies of a unit in the open, so moving too fast is going to reduce the effectiveness of the firing unit. Slowing down the unit will increase the chance of getting pinned and/or taking casualties.

2. Run provides too much cover. Reduce cover, increase exposure to fire. This means greater chance of getting pinned and/or taking casualties.

3. Run overrides adverse behavior too much. Units that should be altering course or getting pinned down are probably being driven forward due to a sort of programmed "determination" for the unit to remain moving to the destination. If we lower this the unit will have a greater chance of faltering, which increases the chance of being pinned and/or taking casualties.

Notice that all three of these possible problem areas might (in theory) yield the same end behavior changes. And they will be compounding too. And that is one of the problems with just getting in there and messing around with this stuff. Plus, tweaking the Run command will likely mean limiting reasonable offensive motion.

Without claiming to have the final answer here, I think we find the solution by doing a couple of fairly major code changes. Tweaking isn't likely going to cut it. So I suggest that we might solve the problem this way...

1. Changing the character of the Run Move to be a very risky order when under fire. Minimum cover, weak return fire, and maximum speed.

2. Adding an Assault Move. This will offer decent cover, decent return fire, and a speed somewhere inbetween Run and Move. The order will only be available to units which are in good shape (like passing a morale check of sorts). Success of the Assault Move will depend heavily on Experience and Leadership bonuses.

3. Adding "go for broke" MG last ditch defensive fire. Once an enemy target gets too close the MG will increase the number of times it can fire in a given slice of time. We will have to code up some logic to prevent this from happening in any and all circumstances or MGs could quickly run out of ammo after just one mad rush by the attacker.

Note that I have said nothing about grazing fire and fire lanes. We honestly are not sure what we can do with these things yet, although firelanes are probably not that hard to add.

Remember folks... the only important thing is the outcome. It is totally irrelevant if we simulate true grazing fire if the overall behavior is realistic. As I have said, the weak spot in CM right now is lone MGs vs. Running squads.

You guys can mess around with this to see roughly Run might do after we tweak stuff. I just did a single quick test using the following setup.

Defender (Foxholes)

2xMMG .30cal M1919A4

Attacker

2xRegular Pattern 44 Infantry

I took away all the ammo for the Attacker and positioned them 200m, bunched up, in front of the two US MGs which were about 20m apart (left to right). I had all the Attackers do Area fire where there was nothing to shoot at. Because of the range the Attacker gets to the MGs in one turn, so that is all I looked at.

The results were as follows...

2x44 Platoons (Regular)

- 30 Casualties (all units one casualty or more, except for one trailing HQ)

- 1xBroken

- 1xRouted

- 3xShaken

- 2xAlerted

- 1xOK (the HQ towards the rear)

I did two quick repeats and found similar results, although casualties were in the lower 20s both times.

So it is very clear to me that MG firepower is not too weak and the lack of grazing fire, fire lanes, and even final defensive "go for broke" behavior is not the main problem to be looked at here. Notice...

All German units suffered at least one casualty and some sort of Morale penalty. This in spite of the fact that the attacker had four times as many units as the defender did. So even though we don't have Grazing Fire simulated to the nines, our "area" effect does in fact spread out damage and penalties to units not being specifically targeted. And this is with the current system of cover and morale "calming" effect built into the Run order.

So imagine this situation where the Attacker has a reduced cover rating and ability to keep on going inspite of enemy fire. Hopefully you can see that without ANY changes to MGs the proposed changes to Run will do a whole lot all by itself. Toss in "go for broke" MG fire and possibly an increase in our current simulation of Grazing Fire (spreading out fire to nearby units) and I think we might be all set, but it is too soon to say one way or the other.

We shall see

Steve

[Michael Emrys]

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

Especially considering that the men under fire are likely crouching (and so presenting a smaller vertical target) fire that affects a group of men at (say) 300m may miss men at 150m.<HR></BLOCKQUOTE>

That would be fine as far as I am concerned. I take that as a more or less realistic result. If MGs using grazing fire can force infantry crossing open ground to stop and take cover (and have to take a morale check), I consider that perfectly adequate. I don't regard it as necessary that half the squad fall down dead.

Michael

[aka_tom_w]

Great Post Steve!

I'm sure some realistic solution to this issue will evolve with CM2. In my opinion it will have to, otherwise the "hords" of russians will ALWAYS win by using the RUSH against MG's.

I think this is an especially sensitive issue for the Russian front, (Side note, My old wargemer buddies ALWAYS stuck me with defending with the Russians in World In Flames, or that "other" BIG WW II boardgame game, as they wanted to prove over and over again that a competent Dicotator could have excuted Barbarossa successfully and won the Eastern Front) as the Russians were known to have lots and lots of manpower and soldiers if NOTHING else, so this issue of dealing with the multiple unit assault rush MUST be dealt with in some realistic way or the Russians with all their "extra" manpower will always RUSH to take out emeny MG positions and it will always work.

I'm sure Steve and Charles are on the right track and I'm sure at some point there will be a CM2 Beta release and I bet it will have at least one scenario that will let us test the new "anti MG rush protocol" with the changes Steve has mentioned.

Keep up the Good work BTS!!!

-tom w

[aka_tom_w]

Steve writes:

"So imagine this situation where the Attacker has a reduced cover rating and ability to keep on going

inspite of enemy fire. Hopefully you can see that without ANY changes to MGs the proposed changes to

Run will do a whole lot all by itself. Toss in "go for broke" MG fire and possibly an increase in our current

simulation of Grazing Fire (spreading out fire to nearby units) and I think we might be all set, but it is too

soon to say one way or the other."

I would like to suggest that if grazing fire is abstracted in the game as "simulation of Grazing Fire (spreading out fire to nearby units)" (I'm assuming grazing fire is to impact units nearby the "targeted" unit) that we should suggest if grazing fire has NOTHING to do with units that are intersecting the actual LOF then grazing fire should ALSO be applied to non-targeted units in close proximity to the MG, if it could be determined that they were generally "in front" of the MG.

This is assuming they are in the open and they are rushing to assault the MG position.

So I think that if grazing fire is modeled as a firepower combat result impacting infantry units nearby the targeted unit then grazing fire should ALSO be a firepower combat result impacting units in close (10-15 meters?) proximity to the MG that is firing. In this way the firepower of the MG could impact multiple units which is important in the solution to this problem as it is the multiple unit RUSH in the open, that is the problem here, I think.

-tom w

[ 04-17-2001: Message edited by: aka_tom_w ]

[BloodyBucket]

I am impressed that this much consideration is being given by BTS to this issue. My Customer Loyalty Meter is just about pegged to the right.

Agree or disagree with the posts, the fact that someone from the production side is listening is deeply gratifying. Thanks.

[ 04-17-2001: Message edited by: BloodyBucket ]

[Michael Emrys]

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

1. Run is too fast for what it is supposed to simulate. Time is one of the worst enemies of a unit in the open, so moving too fast is going to reduce the effectiveness of the firing unit. Slowing down the unit will increase the chance of getting pinned and/or taking casualties.<HR></BLOCKQUOTE>

Agreed. On the other hand, I feel the need to retain some kind of all-out go for broke dash order. Maybe it should be limited to short distances of, say, 20m. Or maybe the exhaustion penalty should be raised for using it. But I still think the game needs to have that in.

<BLOCKQUOTE>quote:</font><HR>2. Run provides too much cover. Reduce cover, increase exposure to fire. This means greater chance of getting pinned and/or taking casualties.<HR></BLOCKQUOTE>

Excellent observation!

<BLOCKQUOTE>quote:</font><HR>3. Run overrides adverse behavior too much. Units that should be altering course or getting pinned down are probably being driven forward due to a sort of programmed "determination" for the unit to remain moving to the destination. If we lower this the unit will have a greater chance of faltering, which increases the chance of being pinned and/or taking casualties.<HR></BLOCKQUOTE>

Hoorah!

<BLOCKQUOTE>quote:</font><HR>Without claiming to have the final answer here, I think we find the solution by doing a couple of fairly major code changes. Tweaking isn't likely going to cut it. So I suggest that we might solve the problem this way...

1. Changing the character of the Run Move to be a very risky order when under fire. Minimum cover, weak return fire, and maximum speed.<HR></BLOCKQUOTE>

Sounds right.

<BLOCKQUOTE>quote:</font><HR>2. Adding an Assault Move. This will offer decent cover, decent return fire, and a speed somewhere inbetween Run and Move. The order will only be available to units which are in good shape (like passing a morale check of sorts). Success of the Assault Move will depend heavily on Experience and Leadership bonuses.<HR></BLOCKQUOTE>

Any chance this Assault Move might also have a Stop and Bring Under Fire Any Spotted Enemy Units component as has been frequently requested? Just asking...

<BLOCKQUOTE>quote:</font><HR>3. Adding "go for broke" MG last ditch defensive fire. Once an enemy target gets too close the MG will increase the number of times it can fire in a given slice of time. We will have to code up some logic to prevent this from happening in any and all circumstances or MGs could quickly run out of ammo after just one mad rush by the attacker.<HR></BLOCKQUOTE>

Okay, but this sounds a tad tricky to implement and debug to me.

<BLOCKQUOTE>quote:</font><HR>Note that I have said nothing about grazing fire and fire lanes. We honestly are not sure what we can do with these things yet, although firelanes are probably not that hard to add.<HR></BLOCKQUOTE>

Oh good! That's some of the best news yet.

<BLOCKQUOTE>quote:</font><HR>Remember folks... the only important thing is the outcome. It is totally irrelevant if we simulate true grazing fire if the overall behavior is realistic.<HR></BLOCKQUOTE>

Yes, but... The trouble with a kludge vs. a true simulation is that sooner or later, someone will always find a way to break the kludge. That is, there will be discovered some situation where it exhibits egregiously unrealistic behavior.

<BLOCKQUOTE>quote:</font><HR>...possibly an increase in our current simulation of Grazing Fire (spreading out fire to nearby units)...<HR></BLOCKQUOTE>

Go for it!

Michael

[Battlefront.com]

Tom wrote:

<BLOCKQUOTE>quote:</font><HR>So I think that if grazing fire is modeled as a firepower combat result impacting infantry units nearby the targeted unit then grazing fire should ALSO be a firepower combat result impacting units in close (10-15 meters?) proximity to the MG that is firing. <HR></BLOCKQUOTE>

I am not sure we are understanding each other. Currently, from I don't know what version (1.0?), a MG has a simulated Grazing Fire effect. This basically applies the firepower of the MG to units imediately around (front, side, rear) of the targeted unit. You can see this in the game now, and is in fact what allowed the two MMGs in the above example to pop at least 7 of the 8 rushing units. We also allow MGs to switch targets more easily than other units, which also sorta simulates Grazing Fire. Range is irrelevant to the area effect thing, but I am pretty sure it does increase target switching as range decreases.

BloodyBucket, thanks

Michael:

<BLOCKQUOTE>quote:</font><HR>On the other hand, I feel the need to retain some kind of all-out go for broke dash order. <HR></BLOCKQUOTE>

This is what Run will turn into. It will become a "dash" order and little more than that.

<BLOCKQUOTE>quote:</font><HR>Any chance this Assault Move might also have a Stop and Bring Under Fire Any Spotted Enemy Units component as has been frequently requested?<HR></BLOCKQUOTE>

Nope. Even better, we are planning on introducing a "Move to Contact" order. It will work similar to the way Hunt does for vehicles, although the nitty gritty of the behavior will not be the same since Infantry has different needs than an AFV does.

<BLOCKQUOTE>quote:</font><HR>Okay, but this sounds a tad tricky to implement and debug to me.<HR></BLOCKQUOTE>

True. The tricky part is to somehow let the MG know that situation A it should open up and risk jams/ammo, but in situation B it should just keep on squirting bursts. I'm sure we can come up with something though. We have tackled far more complex issues plenty of times

<BLOCKQUOTE>quote:</font><HR> The trouble with a kludge vs. a true simulation is that sooner or later, someone will always find a way to break the kludge. That is, there will be discovered some situation where it exhibits egregiously unrealistic behavior.<HR></BLOCKQUOTE>

Correct. And that is why we are discussing MGs right now in the first place The problem is that I do not think it is possible to simulate Grazing Fire the "right" or "complete" way because of the CPU demands. Therefore, any solution to the current soft spots in MG modeling need to be overcome without it.

As I demonstrated above, I think the majority of the problems now have to do with the Run order itself, so the "kludge" for Grazing Fire will be less likely to become a significant way to undermine the system as a whole.

One has to keep in mind that there is probably NOTHING in Combat Mission that, when really looked at, 100% simulates reality. The sum of all of its parts certainly doesn't do this 100% of the time in 100% of the situations that can pop up. Therefore, putting in an abstraction (or "kludge" if you like) is not even an issue. What is is HOW that abstraction is modled and HOW it interacts with the other abstractions.

Steve

[Username]

(Wasnt Steve bellyaching and crying about how bizzy he was an there was gamillions of things he had to do and he had no time for this and that WE wouldnt hear from HIM further on this issue and that it was THEIRS to fix and locking down threads about it, etc? I mean, what the hell is going on? He cant stop posting about it now. Little do many people here know, but I was instrumental in making changes in the initial infantry firepower handling debacle. It was like everyone was firing blanks. Ive been suggesting run limitations, assault moves,etc, since way back when. Its rapidly becoming apparant that my ideas are slowly transmogrifying into BTS "ideas". My subliminal messages are slowly taking hold. Steve doesnt know it but he wakes up in the middle of the night, zombified , and does searches under my name and whatever topic has been troubling him lately. He then goes to work the next day and says "Hey, you know what would be a great idea? Skipping High Explosive shells". Or something like that.)

[Michael Emrys]

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

quote:

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

On the other hand, I feel the need to retain some kind of all-out go for broke dash order.

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

This is what Run will turn into. It will become a "dash" order and little more than that.<HR></BLOCKQUOTE>

Okay. :cool:

quote:

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

Any chance this Assault Move might also have a Stop and Bring Under Fire Any Spotted Enemy Units component as has been frequently requested?

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

<BLOCKQUOTE>quote:</font><HR>Nope. Even better, we are planning on introducing a "Move to Contact" order. It will work similar to the way Hunt does for vehicles, although the nitty gritty of the behavior will not be the same since Infantry has different needs than an AFV does.<HR></BLOCKQUOTE>

Fine! I'm sure many people will be happy to hear this. Even :cool:er.

<BLOCKQUOTE>quote:</font><HR>One has to keep in mind that there is probably NOTHING in Combat Mission that, when really looked at, 100% simulates reality. The sum of all of its parts certainly doesn't do this 100% of the time in 100% of the situations that can pop up.<HR></BLOCKQUOTE>

Yes, that is true, and is inevitably true of any game that attempts to portray any part of the real world. Which is why I have urged that criticism of the game be tempered by the realization that certain things may just be beyond realization at this time.

Michael

[Michael Emrys]

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

[edited for preposterous content]<HR></BLOCKQUOTE>

[snort]

Sure, Lewis. And I'm sure that you not only discovered electricity, invented the digital computer and created the internet, I'll bet you were the very first one to say, "Let there be light."

LOL!

For my part, I'm happy to have Steve resume participation in a give and take discussion. Perhaps he may even have felt that it might stimulate his own creative processes. Who knows? But if it leads to an improved product, I'm happy to see even you providing input.

Michael

[StellarRat]

<BLOCKQUOTE>quote:</font><HR>Originally posted by Michael emrys:

[snort]<HR></BLOCKQUOTE>

I second that snort and also claim that I was instrumental in getting other people to realize that a snort was necessary.

[Vanir Ausf B]

This is just classic. First Lewis rags on Jason (and Rexford) saying:

<BLOCKQUOTE>quote:</font><HR>

Is JasonC related to Rexford?

I notice the same long winded self serving style in his posts. Also the always lovely "follow my own post with yet another important post by myself".<HR></BLOCKQUOTE>

And then follows that with:

<BLOCKQUOTE>quote:</font><HR>

Little do many people here know, but I was instrumental in making changes in the initial infantry firepower handling debacle. It was like everyone was firing blanks. Ive been suggesting run limitations, assault moves,etc, since way back when. Its rapidly becoming apparant that my ideas are slowly transmogrifying into BTS "ideas".<HR></BLOCKQUOTE>

Sometimes all you can do is laugh.

[Wreck]

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

If you guys think it is hard to impliment the effects of grazing fire once the direction is settled, wait until you get to the AI, and get to try to backsolve the expected results

<HR></BLOCKQUOTE>

Simple. I don't expect that. As I said in my post, this is not a perfect simulation. It is an implementation of grazing that is good enough that at least some of the time it will produce more firelane-like results. But primarily it is good because it is computationally light.

<BLOCKQUOTE>quote:</font><HR>

for the tac AI [solving that problem] is ridiculous.

<HR></BLOCKQUOTE>

Well then, let's not try to have the AI solve it. Since it does not do so now, in fact all we need to do, is nothing. Making no change -- BTS definitely has time to do that.

<BLOCKQUOTE>quote:</font><HR>

And unneeded, because the real problem with the way rushes happen now is a cover effect, not a fire-lane effect, and the role of firelanes can be abstractly and on-average included in the solution.

<HR></BLOCKQUOTE>

I agree that the problem is a cover effect. I proposed a solution myself in the previous thread, easily implementable: just downgrade the move status of units under exposed fire, with the idea that they are trying to get less exposed. This is, IMO, realistic.

However, there is a separate problem (that of the firelane), which IMO is not gonna be solved in approximation by point fire. And I do not think grazing fire is small enough to wave away with abstract, on-average coverage.

I entered the thread to help BTS with firelanes, which they are already going to try to do. Do they need any help? I don't know. Clearly they have an uberhacker on staff, which is good, but even the best don't always think of everything.

As for your solution I don't see it really slowing troops down (except via morale effects), so I don't think it is all there is to say. I certainly don't want to micromanage my rushes across an open field (bursts of running with crawling mixed in to get my exposure down). I want my troops to do that automatically, perhaps even sometimes against my will.

[Username]

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

Defender (Foxholes)

2xMMG .30cal M1919A4

Attacker

2xRegular Pattern 44 Infantry

I took away all the ammo for the Attacker and positioned them 200m, bunched up, in front of the two US MGs which were about 20m apart (left to right). I had all the Attackers do Area fire where there was nothing to shoot at. Because of the range the Attacker gets to the MGs in one turn, so that is all I looked at.

The results were as follows...

2x44 Platoons (Regular)

- 30 Casualties (all units one casualty or more, except for one trailing HQ)

- 1xBroken

- 1xRouted

- 3xShaken

- 2xAlerted

- 1xOK (the HQ towards the rear)

I did two quick repeats and found similar results, although casualties were in the lower 20s both times.

So it is very clear to me that MG firepower is not too weak and the lack of grazing fire, fire lanes, and even final defensive "go for broke" behavior is not the main problem to be looked at here.

So imagine this situation where the Attacker has a reduced cover rating and ability to keep on going inspite of enemy fire. Hopefully you can see that without ANY changes to MGs the proposed changes to Run will do a whole lot all by itself. Toss in "go for broke" MG fire and possibly an increase in our current simulation of Grazing Fire (spreading out fire to nearby units) and I think we might be all set, but it is too soon to say one way or the other.

We shall see

Steve<HR></BLOCKQUOTE>

Steve

Actually you have modeled one of the proposed changes; limiting moving fire from running units and decreasing tired/weary effectiveness of fire to nill.

You dont atribute this to the outcome much but just guage the MG fire. As I suggested in the other thread; reduce the FP of running units and make exhausted states very risky business.

I dont get the "calming of running" effect at all. It tends to get people excited, running that is, and people firing at them doesnt help. I get your point though. This needs to be tweaked of course.

It is not my contention that fire lanes/grazing fire needs to be modeled directly. I like some of the abstractions that achieve the same effect. Having an MG fire quickly at units within a narrow covered arc within a reasonable range is just as good.

I would also like BTS to look at units that run/move through enemy positions on robot orders. It is my opinion that there has been something lost because there are no hexes to show a units "possesion" of terrain. I dont want units to get pinned because they are close to the enemy but rather "halted". A good order state. But it would take an assault command to get units to butt heads with enemy defenders "in their own space". Perhaps the space they command can be a function of their present suppression, etc.

Anyway. Its my idea and you heard it here first..

Lewis

[Maastrictian]

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

OK, I looked it up last night. For a 5.56 mm bullet fired from an M-16 (which does not have as flat a trajectory as a 30 cal. bullet):

With the rear sight set for long range firing, the bullet rises to a maximum height of 11 inches above the shooters line of sight to the target at 225 meters, and crosses the line of sight again at 375 meters. With the rear sight set for short range firing, the bullet rises to a maximum height of 5 inches above the shooters line of sight to the target at 175 meters, and crosses the line of sight again at 250 meters.<HR></BLOCKQUOTE>

I think the discussion has rightfully moved away from my mad calculations, but thanks to Marlow for finding some data to refute my point.

[Wreck]

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

I think you are just woefully underestimating how difficult it is for a CPU to crunch these sorts of numbers.

<HR></BLOCKQUOTE>

Actually I was trying to point out, that contrary to what some were saying, doing a check for angular distance is not as bad as it seems. No trig function is needed. The operations needed turned out to be: 11 floating point multiplies, one floating point divide, six floating point adds, and one if-statement to deal with the case of negative dot-product. A total of less than twenty operations all implemented in hardware on any pentium-I class system or better; you should be able to do millions of angle checks per second.

As for the LOS check, I accepted your assertion that during a turn, LOS checks collectively are heavy. It seems like they should be to me, and in any case I have no way to check this. But for my proposed form of grazing, the angle check cuts off most LOS checks. The number will be further cut down by something I suspect is already in there, namely, cached LOS information for non-moving units. So the number of full LOS checks my solution would be adding is going to be small in the average case. Most likely not even one per MG, but it might go up to several per MG per burst, if they are all moving together across the MG's firelane.

<BLOCKQUOTE>quote:</font><HR>

Because you are not a programmer (and neither am I, just an experienced developer) you are just going to have to trust me that there is absolutely no easy/slick way around the problem.

<HR></BLOCKQUOTE>

Actually, I am a programmer. I am trying to give you a slick way around much of the problem.

Whether or not it really is a slick solution depends on the costliness of a single LOS check. If there are already hundreds or thousands of these per turn resolution (which is what I guess is true), then adding another 2 every MG burst is going to be pretty small. If there are only say tens of LOS checks per turn (eating 20% of the time), then adding a few more is going to be much more costly.

[aka_tom_w]

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

Tom wrote:

quote:

So I think that if grazing fire is modeled as a firepower combat result impacting infantry units

nearby the targeted unit then grazing fire should ALSO be a firepower combat result

impacting units in close (10-15 meters?) proximity to the MG that is firing.

I am not sure we are understanding each other. Currently, from I don't know what version (1.0?), a MG

has a simulated Grazing Fire effect. This basically applies the firepower of the MG to units imediately

around (front, side, rear) of the targeted unit. You can see this in the game now, and is in fact what

allowed the two MMGs in the above example to pop at least 7 of the 8 rushing units. We also allow MGs

to switch targets more easily than other units, which also sorta simulates Grazing Fire. Range is

irrelevant to the area effect thing, but I am pretty sure it does increase target switching as range

decreases.

Steve<HR></BLOCKQUOTE>

Hi Steve

I think I understand the way you have explained grazing fire as having a sort of collateral effect or impact on other infantry units nearby the targeted infantry unit. (is that correct?)

My suggestion was to have the same form of grazing fire effect or impact on all non-targeted units in the open advancing on the MG position once they get within ~10-15m of it if the MG was targeting a more distant unit, I think that MG's usually target the closest unit so maybe this is not a big deal. Also your suggestion for the enhanced, "go for Broke" (expend all rounds) MG protocol would probably work better than the suggestion to try to apply the impact of grazing fire on ALL opposing units close (10-15m) to the MG position.

I think that those suggestions you posted earlier will generate the desired result which I think we all agree is to make the multi unit "bum Rush" or running assualt against MG positions over open terrain a great deal less effective than it currently is.

Thanks!!

-tom w

[ 04-18-2001: Message edited by: aka_tom_w ]