MG team recommendation

[Battlefront.com]

Hey, whadda know! Lewis is actually correct. I owe you an apology there. Looks like Charles tossed it in there at the last minute, and also did the graphic. I kept seeing "M1917 HMG" and was thinking "M1917" .50cal, which was the original designation. Learn something new every day

Steve

[ 04-16-2001: Message edited by: Big Time Software ]

[Username]

And ya know what else? Four of 'em cant stop a company of banzai-ing volkgranadiers!

We dont have to just look at WWI. Theres plenty of insane infantry behaviour regarding MGs and WWII. During the great encirclements in 1941, russian troops literally pushed their way through stretched out MG positions rather then surrender. Japanese like a verry much charge a MG after a night of saki. Marines didnt have many options BUT to charge MGs.

Anyway, back to the taxes.

Lewis

PS Glad it made it into the game.

[Michael Emrys]

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

As for grazing fire, it is only an issue for up close and personal engagement ranges in defender ideal terrain. Even a high velocity MG like the MG42 has a significant arc to its shots. I just saw an image of a chart from a MG1 (postwar adopted MG42) manual showing the angle of shot over distance. To get rounds to hit targets at even medium distances (500m or so) there was a significant arc necessary. And this is with a perfectly flat target range as the basis. Getting maximum grazing fire effects in the real battlefield are a lot harder, and over some range impossible. So yes, this feature is very significant but I bet a lot less so than some think. Close range, optimal ground conditions are necessary for grazing fire to work effectively.<HR></BLOCKQUOTE>

By "arc", Steve, I assume you mean the height of the bullet trajectory above the sight line. I don't have exact figures before me, so I might be off here, but I would guess that for 500m it wouldn't be more than a foot or so high for the 7.92mm Mauser and not much worse for the .30 Browning. From what I have heard, the .50 M2 should be even flatter over that range.

But even that is relatively irrelevant, as for defensive fires we need only be really concerned with bullet performance out to a range of 200-300m, and there the height of trajectory is probably on the order of 6-8 inches.

In other words, grazing fire should be quite operable over the stated ranges. I think the objection you were trying to make to this, though you might not have stated it as clearly as I would have hoped, is that all this matters only so long as there are not areas of dead ground that are many or large, and that is quite true. If the machine gunner can't see the enemy because of intervening obstacles, he can't shoot at them. But then, one of the basic skills taught to junior officers is precisely to locate his machine guns and other weapons so that there are not many large areas of dead ground. This even includes if necessary clearing brush and other obstacles when possible and time permitting.

Not being a programmer, I cannot estimate how difficult it would be to accurately write grazing fire into the game engine. I'm sure we can and will live with whatever limitations are imposed on you by those difficulties. But it really is pointless to deny the importance of the tactic. As ASL Veteran said, this is what machine guns were made to do. Certainly they performed this function with greater and lesser degrees of effectiveness dependent on the usual assortment of variables. But, as you have already acknowledged, the function is presently undermodeled in the game.

That is not to ask that MGs become the next super weapon. Obviously there are ways to attack and suppress them. They are not invulnerable. Further, I really don't think an increase in lethality is called for. That's an important point. But they should definitely be forcing moving units, especially those moving through open ground, to drop to the ground and seek cover, even the cover that is notionally present even in "open" ground. And as a number of players have noted, at the present time the defense in CM is a little anemic. And as Bullethead has asserted earlier, this is largely because the primary function of MGs in the defense is not being modeled in CM.

Michael

[Battlefront.com]

There is, code wise, no difference between an air cooled and a water cooled MG. Obviously this should be some sort of factor when we plug in the "go for broke" firing behavior. Like you say, the one remaining advantage of a water cooled MG (in theory, anyway) is that it can fire sustained longer than an air cooled one. Normally this isn't an issue, which is why everybody uses air cooled MGs now, but in emergency fire it might be a factor. Probably just reduce the chance of a mechanical failure as ammo feeding is just as much a RoF limiting factor for either type of MG.

Steve

[Battlefront.com]

Michael,

<BLOCKQUOTE>quote:</font><HR>Not being a programmer, I cannot estimate how difficult it would be to accurately write grazing fire into the game engine.<HR></BLOCKQUOTE>

Very difficult. There are all sorts of problems with trying to simulate grazing fire. Which is why it isn't in CM1 to the nines. The major problems are as follows:

1. Intersection - as most of you know, there is no exact tracing of a round's flight path from A to B, even for the big guns. This kind of calculation is a CPU killer.

2. Dead Zones - as flat and featureless as "Open Ground" looks in CM, it is not simulating a pool table or shooting range. Therefore, we somehow need to fairly account for dead zones. As you say, gunners try to position themselves so significant dead zones are either not present or are covered by other units. The problem is that this is highly variable and very dependent on the length of prepration the unit/s have for their defense.

3. Terrain Slope - a totally seperate factor is the slope of the terrain and the relation shooter and target have relative to each other. If two units are on the same plane (i.e. terrain height) that is fairly simple. But what if there is a one height dip inbetween the two, and the attacker goes into it? Totally different effects most likely.

The first one is the real killer as it has nothing to do with coding time or skill, but rather hits to the CPU. So I don't know if it is even physically possible to do totally correct Grazing Fire for CM2. We are, however, looking into doing just that and therefore we will just have to see what comes up.

<BLOCKQUOTE>quote:</font><HR>And as a number of players have noted, at the present time the defense in CM is a little anemic.<HR></BLOCKQUOTE>

Yes, but again... not necessarily because of the lack of full Grazing Fire. I think the Run move cover factor and/or speed (one or the other needs to be reduced), the TacAI behavior assigned to the Run move, and the lack of "going for broke" behavior is far more at fault for producing less than reasonable MG results.

As I said in the previous thread, this is where we are going to start when we look at changes for CM2. The non Grazing Fire changes are fairly easy stuff to change (code wise) and fairly easy to tweak (in theory at least). So basically, if we can get the behavioral changes we think are needed, without expanding Grazing Fire from its current area affect around the aimed target, then that lessens the need for full bore Grazing Fire treatment. And we hope that happens since we are not confident that the CPU can handle the demands of a real deal Grazing Fire system.

Steve

[Michael Emrys]

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

Lots of outdated weapons are used during war, but from what we can tell there were very few M1917s in Western Europe 1944-1945. Were there some? Most likely.<HR></BLOCKQUOTE>

Yep. For instance, I have in my possession a photograph of one. The caption places it in Italy. There were a ****load of them being used in various locations in the Pacific during the early part of the war.

Which is not to deny that they were definitely on the way out in the US Army by the time period that CM covers. On the other hand, the Vickers were still certainly soldiering on, as they continued to do even in Korea and later. And the Germans certainly issued them to static troops. And the Soviets certainly used the Maxim in large numbers for the first year or so of their war.

<BLOCKQUOTE>quote:</font><HR>But since the M1919A4 replaced the M1917 in the 1920s as the standard weapon<HR></BLOCKQUOTE>

It began to do so. As they were still useful weapons, they were kept in stock for training purposes, and when the Army was drastically and suddenly enlarged at the start of the war, they were often issued to frontline units in lieu of the aircooled Brownings whose production took a while to catch up with demand.

<BLOCKQUOTE>quote:</font><HR>BTW, we left out craploads of small arms for the Germans using this same "we can only simulate so much" logic.<HR></BLOCKQUOTE>

I understand and agree...up to a point.

Michael

[ 04-16-2001: Message edited by: Michael emrys ]

[ 04-16-2001: Message edited by: Michael emrys ]

[Username]

It will probably be coded similar to getting a hull down position. Some people (obviously great gun emplacers) will state that it can be done with practice and skill and others will not have a clue.

Sometimes you wont know till you fire. The presence of enemy units defines the firelane. If a tree falls in the woods on a bear, does anyone hear it?

[Homba]

...

(edit) let a post sit on your desktop for a few minutes before submitting and it quickly becomes obsolete!

H

[ 04-16-2001: Message edited by: Homba ]

[aka_tom_w]

<BLOCKQUOTE>quote:</font><HR>Originally posted by Adam Lloyd:

Steve,

Can you just program it to affect everyone in the orange "area fire" line, rather than simply at the end of that line?

In effect, a very long "corridor" area target order.

(Hope I'm being clear here)

There would be no need to trace the individual bullets.<HR></BLOCKQUOTE>

This is EXACTLY what the game engine does not do. ("affect everyone in the orange "area fire" line, rather than simply at the end of that line?"). From everything Steve has said this is VERY CPU intensive and it has been stated by Steve and Chrales this would take far too long for the "average" cpu to crunch or correctly compute the result of combat to correctly determine all units that intersect all LOF's in the game.

As it is now there is no intersecting fire in the game. Main Weapons of tanks fire straight thru friendly and enemy AFV's (and Pillboxes and bunkers and Roadblocks and sometimes even Houses ) that are not buring and are not targeted, BECAUSE the action of the shot is ONLY calcuated against the intended target, not what may lie inbetween.

The Line of Fire of an HMG is NOT a stream of bullets or a wall of death, not at all, only the area targeted or the unit targeted "feels" the effect of the HMG firepower on it. (in the form of a combat resolution result inflicted on the target). So obviously walking or running throught the LOF of an HMG is not problem at all (i.e. there is no effect) if the HMG is already targeting a more distant unit, becaue there are no real bullets in that Line of Fire.

And yes, Steve has said this is an issue and it is a VERY difficult problem to resolve realistically in the game.

-tom w

[Wreck]

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

... we are not confident that the CPU can handle the demands of a real deal Grazing Fire system.

<HR></BLOCKQUOTE>

I am a bit surprised by this. Certainly LOS is already being computed quite frequently during a turn resolution, right? So the LOS check cannot be *that* expensive.

Given a vector that the MG is firing, grazing can be approximated using LOS. If the MG has LOS to a unit, and the angle between the vector to that unit and the target vector is below some critical angle, the unit will be "grazed". If you wanted to get fancy about what angle to use you could factor in distance so as to decrease the fire lane's angular dimension w/ range (which is probably the right way to do it).

Obviously this fails to take into account units that are near to the fire but which the MG cannot see. But it would certainly be sufficient for the vast majority of situations. And in any case at least the more experienced soldiers would be unlikely to be affected by fire from an MG that cannot hit them, even if it does happen to be going close by.

Am I incorrect in assuming that LOS checks are happening in great multitudes during resolution?

[StellarRat]

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

I am a bit surprised by this. Certainly LOS is already being computed quite frequently during a turn resolution, right? So the LOS check cannot be *that* expensive.

Given a vector that the MG is firing, grazing can be approximated using LOS. If the MG has LOS to a unit, and the angle between the vector to that unit and the target vector is below some critical angle, the unit will be "grazed". If you wanted to get fancy about what angle to use you could factor in distance so as to decrease the fire lane's angular dimension w/ range (which is probably the right way to do it).

Obviously this fails to take into account units that are near to the fire but which the MG cannot see. But it would certainly be sufficient for the vast majority of situations. And in any case at least the more experienced soldiers would be unlikely to be affected by fire from an MG that cannot hit them, even if it does happen to be going close by.

Am I incorrect in assuming that LOS checks are happening in great multitudes during resolution?<HR></BLOCKQUOTE>

Three dimensional vector calculations with an arc thrown in are not trival. They involve a lot of floating point trignometric calculations and these are CPU hogs.

[aka_tom_w]

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

Am I incorrect in assuming that LOS checks are happening in great multitudes during resolution?<HR></BLOCKQUOTE>

Hi Wreck

I believe you are incorrect.

I say this because it is my understanding that LOF and fire affects units even when they move out of LOS. During the calculation or the "crunch" combat resolution firepower factors and hit/not hit and AFV penetration results are caculated against targets that may be out of LOS, this why some here complain that some rounds from tanks GO straight through buildings AFTER the AFV has moved into the LOS shadow. I'm not sure about MG and small arms fire, but I suspect there is NO LOS check

during the crunch when AFV's fire rounds at a target that did have LOS at the

begining of the turn and then moves out of LOS early in the 1 minute turn. (I

remember this as Steve has theorized and posted his thoughts about an additional

LOS check after the round is fired to see if the target is still in LOS, this after fire LOS

check is not curently modeled for rounds fired from AFVs against other AFV that

move out of LOS).

-tom w

[ 04-16-2001: Message edited by: aka_tom_w ]

[aka_tom_w]

oops

[ 04-16-2001: Message edited by: aka_tom_w ]

[Battlefront.com]

Adam,

What you are saying is the same thing I discussed about being a problem. When you move the LOS tool around the map the CPU is almost 100% available to do the LOS calculation. So it might look like such a calculation is quick and easy, but it is just one instance during a time when nothing else is going on.

Tom, you have been following this problem since the beginning. You have a good understanding of the problem.

Wreck wrote:

<BLOCKQUOTE>quote:</font><HR>Am I incorrect in assuming that LOS checks are happening in great multitudes during resolution? <HR></BLOCKQUOTE>

LOS checks are being performed all the time, true enough. But only to see if a target can be shot at. LOS isn't checking for variable elements at all right now. Only terrain is taken into account, not units. The result of this is a huge savings of CPU usage. To do grazing fire correctly we would have to abandon this abstraction and therefore lose our #1 calculation shortcut. So we aren't sure if we can do that or not.

IIRC, we did a test and found that 20% of a "busy" turn's resolution time was spent doing nothing but LOS checks. Charles gives me the impression that if we were to check for variable units in the LOS path there would be an exponential increase in CPU cycles spent on LOS. Since it is already taking up 20% of the total (nothing else comes close to hogging the CPU this much, BTW) the potential impact on turn resolution times could be unacceptable from a game player standpoint. Truth is, we don't know so we will take a look at it. However, StellarRat is right on the money when he said:

<BLOCKQUOTE>quote:</font><HR>Three dimensional vector calculations with an arc thrown in are not trival. They involve a lot of floating point trignometric calculations and these are CPU hogs.<HR></BLOCKQUOTE>

Steve

[Maastrictian]

Michael emrys wrote:

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

By "arc", Steve, I assume you mean the height of the bullet trajectory above the sight line. I don't have exact figures before me, so I might be off here, but I would guess that for 500m it wouldn't be more than a foot or so high for the 7.92mm Mauser and not much worse for the .30 Browning. From what I have heard, the .50 M2 should be even flatter over that range.

<HR></BLOCKQUOTE>

This may or may not have any relevance whatsoever to the current discussion. But I'd like to refute the above point:

All this assumes no air resistance.

y(the drop of the bullet) = at^2

x(the distance fired) = vt

y = .3m (about 1/3 of a meter, the foot of drop quoted above)

x = 500m

a = 9.8 m/s^2 (gravity)

500 = vt

t = 500/v

.3 = (.5)*9.8 t^2

.3 = (.5)*9.8 (500/v)^2

.3 = (.5)*9.8 * 500^2 / v^2

v^2 = 4,100,000

v = 2000 m/s

(or 4300 miles per hour, or about mach 6 at sea level, way too fast for a bullet).

Here is a calculation of the speed needed for a one meter drop (assume that is the distance from the barrel to the ground) in 500m

y = 1m

x = 500m

a = 9.8 m/s^2 (gravity)

500 = vt

t = 500/v

1 = (.5)*9.8 t^2

1 = (.5)*9.8 (500/v)^2

1 = (.5)*9.8 * 500^2 / v^2

v^2 = 1,200,000

v = 1100 m/s

(or about 2300 miles per hour or about mach 3.3, also way to fast for a bullet)

For fun, lets calculate how long it takes for that one meter drop for a bullet going mach 1.5 (probably a fast muzel velocity, and its not going to stay that fast for long). This is an upper estimate of "grazing fire" range with out doing fancy things like angling the barrel up, just pointing at someone and shooting.

y = 1m

x = ?

a = 9.8m/s^2

v = 500m/s

1 = (.5)*9.8 t^2

t^2 = .204

t = .45

x = 500m/s*.45

x = 225m

Now for a lower bound estimate we do the same calculation with a muzel speed of mach .75 (assuming standard muzel speed is a bit less than mach 1 and air friction slows thigns down, I really don't know what a reasonable muzel speed is, that's for others besides myself.)

y = 1m

x = ?

a = 9.8m/s^2

v = 250m/s

1 = (.5)*9.8 t^2

t^2 = .204

t = .45

x = 250m/s*.45

x = 110m

[CONCLUSION! if you're ignoring the caluculations, look here!]

So grazing fire effective range is between 230m and 110m, probably closer to the lower value.

[Homba]

Im sure "angling the barrel up" is nothing fancy for a dug-in MG... probably standard practice or else you are shooting at ankles.

In any case it must come into the calculation somewhere and to some extent, and this is going to increase effective grazing fire range.

Surely someone can provide a reference on muzzel velocities of WWII machine guns? We have one first hand account from the other thread of MG bullets' 'cracking' sound as they broke the sound barrier passing overhead (though IIRC this was a more modern MG.)

Homba

[StellarRat]

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

Surely someone can provide a reference on muzzel velocities of WWII machine guns? We have one first hand account from the other thread of MG bullets' 'cracking' sound as they broke the sound barrier passing overhead (though IIRC this was a more modern MG.)

Homba<HR></BLOCKQUOTE>

Even a cheesy rifle will have a muzzle velocity of 2000 feet per second. Also, most rifle bullets tend to rise in the first few yards of flight, so the guy with the calculations above is a little off. WW II MGs are definately above 1500 feet per second (900 is supersonic).

Just as an interesting side note a modern SABOT round has a muzzle velocity of about 5000 feet per second. Think of the wallop those carry! Small wonder they literally blow vehicles to pieces!

[ 04-16-2001: Message edited by: StellarRat ]

[Maastrictian]

StellarRat wrote:

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

Even a cheesy rifle will have a muzzle velocity of 2000 feet per second. Also, most rifle bullets tend to rise in the first few yards of flight, so the guy with the calculations above is a little off. WW II MGs are definately above 1500 feet per second (900 is supersonic).

<HR></BLOCKQUOTE>

Does anyone have any sources for MG muzzel velocities as well as down range velocities?

Also, not sure about this "bullets rise in the first few yards" thing... if first few yards means ~10 yards its not going to make a difference to my calculations. If you mean that people fire at distant targets by elivating the barrel then yes, obviously I agree with you and my numbers do not factor that in. But my image of MGs is more of a "point and shoot" weapon rather than a "line up target in range calibrated sight" weapon. This could be wrong, but my impression has been that the procedure for grazing fire is to level gun at firelane and fire bursts, not really acounting much for range.

Note that the point I'm making has nothing to do with the efficacy or existance of grazing fire, merely its effective range.

[ASL Veteran]

Some sample muzzle velocities

MG 34 and 42: 755 m/s

MG 08/15 (German WW1 MG): 900 m/s

US .30 cal M1917: 854 m/s

US .50 cal M2HB: 884 m/s

Soviet PMO 1910 (Maxim): 863 m/s

Gun, Machine, Vickers, 0.303, Mark 1: 744 m/s

Japanese 7.7mm Type 92: 732 m/s

Italian Mitriaglice Fiat 1914/35: 790 m/s

French Hotchkiss: 725 m/s

[ASL Veteran]

Also I might note that the FM says that 1 meter high is about hip height for a standing man (I don't know metrics too well so I can't comment on the accuracy of that) and MGs aren't generally set up with the barrel level at hip height. The barrel would be much closer to the ground than that so if you used ... say barrel at .5 meter height with the trajectory not exceeding 1 meter maybe that would get better results.

[Jager 7]

MG 42 Muzzle velocity = 2,480 fps

M1917A1 .30 MV = 2800 fps

M1919A4 .30 MV = 2800 fps

M2 HB .50 MV = 2930 fps

I knew someone was going to go looking for it and I had it handy. Ref "Small Arms of the World" Ezell (Smith)

[Marlow]

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

StellarRat wrote:

Does anyone have any sources for MG muzzel velocities as well as down range velocities?

Also, not sure about this "bullets rise in the first few yards" thing... if first few yards means ~10 yards its not going to make a difference to my calculations. If you mean that people fire at distant targets by elivating the barrel then yes, obviously I agree with you and my numbers do not factor that in. But my image of MGs is more of a "point and shoot" weapon rather than a "line up target in range calibrated sight" weapon. This could be wrong, but my impression has been that the procedure for grazing fire is to level gun at firelane and fire bursts, not really acounting much for range.

<HR></BLOCKQUOTE>

Rifles and machineguns are zeroed to take bullet drop into account. Generally, they are intended to hit the target at which the firer is aiming at a range of 250 m or so. This results in a bullet that rises for significantly more than 10 yards, and has nothing to do with elevating the barrel for long range firing.

Combined with the fact that most machinegun barrels are fired from less than a meter off the ground (as noted by ASL Vet), there is a significant upward velocity that must be taken into account in any grazing fire range calculation.

[ 04-16-2001: Message edited by: Marlow ]

[Battlefront.com]

The discussion above shows that Grazing Fire is not a simple "I fire my MG, everything in front of it can be hit" sort of thing. It is complex and slapping something in (even if we could) would most likely make MGs too powerfull. As I said, we will do our best to deal with grazing fire for CM2.

Adam wrote:

<BLOCKQUOTE>quote:</font><HR>I'm not a programmer so I have no idea how hard it is to code this stuff. How do FPS's and other games calculate fire like this? Perhaps the solution lies somewhere in there.<HR></BLOCKQUOTE>

If you mean limiting the number of possible variables down to a dozen or so and reducing the size of the shooting area to about 50-100m... sure, there is a solution to be found in FPS games This all has to do with scale, just like with polygons. There are all sorts of things you can do with a small "world" that just aren't practical when you move to a larger one.

Think of how many units are in the average, just AVERAGE, game. Then think of what the average engagement range is for each of those units. Then think of how many times each unit fires per turn multiplied by how many units in the game. Now think of what would happen if you played a large game. Now think about making grazing fire (or shell path for big stuff) fairly applied to all weapons systems. Etc...

Ugh

Steve

[Michael Emrys]

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

This may or may not have any relevance whatsoever to the current discussion. But I'd like to refute the above point:

[snip]

<HR></BLOCKQUOTE>

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.

Therefore, for the entirety of its flight before it reaches the target's range, it is actually above the sight line. For most rifle caliber bullets, it reaches the maximum height above the sight line about 2/3rds. to 3/4ths. of the distance from the gun to the target. One of the proofing tests for new guns accepted into service is to measure bullet trajectory rise at various ranges given various ranges to the final target. I was attempting to quote from memory the results of such tests because I no longer have those tables at hand. My quotes therefore may well be off, but not by nearly such a large margin as your calculations would imply.

Given the breadth of experience and knowledge that turns up on this board from time to time, I do hope that someone who has the correct information will provide it to us, just to satisfy my own personal curiousity.

Michael

[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