81mm mortar dispersion

[Steiner14]

YankeeDog,

the explosive radius is modelled in CM anyway.

In case the unit is not dug in, the effect is stronger, if dug in the effect is less severe.

I used a small area to show how small the probability is, for a direct hit against a dug in unit, while my impression is, we can see direct hits all the time in the game.

The calculation, if correct, supports the thesis, that the mortars in the game are too precise.

But let's use the size of an action spot: roughly 8 x 8 yards = 64 yard².

64 / 512 = 12,5% or every 8th round.

If the formula is correct roughly only 13% of all rounds should fall into the action spot @400m for that mortar type.

[YankeeDog]

If the formula is correct roughly only 13% of all rounds should fall into the action spot @400m for that mortar type.

Unfortunately, my CM-capable computer is down at the moment, so I can't verify, but my recollection is that this is very close the accuracy I see in-game -- that is, for a typical on-map, direct LOS mortar fire mission, something like 1 out of every 8 mortar shells mortar falls into the action spot that is actually targeted.

Note this means that a single minute of fire from a 60mm mortar at <400m (which is usually ~ 10 shells) will pretty reliably take out an MG or gun position, since that single shell detonating within the action spot will almost certainly cause multiple casualties and render the unit combat ineffective, even if not everyone in the unit is a casualty.

[Steiner14]

A test @400m would be very interesting.

[YankeeDog]

Yes, it would... if no one else has done so, I will do some test runs once I have my new computer set up. Unfortunately, this may be a week or so.

(Suffering from CM withdrawal)

[Steiner14]

I'm not at my computer for one week, too.

[Wreck]

I already tested 81mm at 402m. You can see that here. The pattern is circular with CEP=~8m. I expect the pattern for 60mm is the same, but since 60mm don't leave craters you can't easily show it.

[YankeeDog]

Thanks, Wreck.

Any way you could give us a rough idea of scale in that image? Looks like there are pieces of a wall in the center top and center bottom -- maybe you could tell us the approximate distance between the ends of the two walls?

[Wreck]

The top/bottom distance is 6 action spots or 48m. Those two obvious piece of wall you see are "+" shapes (they make it easy to see where the target action spot is). The bits you see are each 4m long, so the distance between their ends is 40m.

[Steiner14]

I only have rudimentary painting skills, so please excuse the not very precise paintjob.

Since the distance between the two walls is 40m i have marked in black roughly 10m to find the 8m range of the action spot.

The red stripe shows the 2x8 m wide dispersion band in range.

The cyan stripe shows the 2x8 m wide dispersion band in deflection.

The numbers in green are the hits within and outside the 2x 8m red stripe.

The numbers in cyan are the hits within and outside the 2x 8m cyan stripe.

If this example is in the statistical norm, then the numbers show the following:

In range (within the red stripe) the hit probability lies at 58%. This is 16% above 50%.

In deflection (within the cyan stripe) the hit probability lies at 69%. This is 38% above 50%.

Uploaded with ImageShack.us

Following the statistical data in the manual of the 60mm unit, with charge 0, 1 and 2 the pecision in range should NOT be better but worse than the precision in deflection.

For example the manual gives for the maximum distance of charge 0 345m and a range : deflection dispersion of 8:1.

With charge 1 @400 yd. 8 range : 4 deflection.

With charge 2 @400 yd. still 7 yd. in range : 6 yd. in deflection.

Specifically the precision in range seems to be way too high (at least @400m and for the 60mm mortar). The pattern is not streched enough.

I don't know if BFC is absolutely satisfied with their model, or if they are open to suggestions.

I have derived a simple model with equations of a line, that allows to get a not so bad dispersion pattern only by using the the minimum and maximum values from manuals:

For an even easier calculation i'll set yard = meter. In case this should be compensated just convert the data from the manuals into m.

Let's say the minimum distance of the modeled mortar should be 50 m.

And the maximum distance is decided to be 1500 m.

Dispersion in range:

The manual gives us roughly

Min. distance: 50m, dispersion in range 1m.

Max. distance: 1500m, dispersion in range 22m.

The simple equation of a line y = k*x + d.

DispersionInRange = kInRange*TargetDistance + DispersionInRangeOffset

kInRange= (DispersionAtMaximumDistance - DispersionAtMinimumDistance) / (MaximumFiringDistance - MinimumFiringDistance)

Equation for the dispersion in range is already finished:

DispersionInRange = (DispersionAtMaximumDistance - DispersionAtMinimumDistance) / (MaximumFiringDistance - MinimumFiringDistance) * TargetDistance + DispersionInRangeOffset

Concrete numbers for the model:

DispersionAtMaximumDistance = 22

DispersionAtMinimumDistance = 1

MaximumFiringDistance = 1500

MinimumFiringDistance = 50

-> kInRange = 21/1450

DispersionInRangeOffset = 0,3

This gives for the dispersion in range:

@50m: 0,3 + 50 * 21/1450 = 1m (manual: 1m)

@100m: 1,7m (manual: 2m)

@400m: 6,1m (manual: 6-7m)

@600m: 9,0m (manual: 8-12m)

@1500m: 22m (manual: 22)

Dispersion in deflection:

DispersionInDeflection = kInDeflection*TargetDistance + DispersionInDeflectionOffset

kInDeflection= (DispersionAtMaximumDistance - DispersionAtMinimumDistance) / (MaximumFiringDistance - MinimumFiringDistance)

Equation for the dispersion in deflection:

DispersionInDeflection = (DispersionAtMaximumDistance - DispersionAtMinimumDistance) / (MaximumFiringDistance - MinimumFiringDistance) * TargetDistance + DispersionInDeflectionOffset

With some concrete numbers:

DispersionAtMaximumDistance = 6

DispersionAtMinimumDistance = 2,5

-> kInDeflection = 3,5/1450

DispersionInRangeOffset = 2,9

Results for dispersion in deflection:

@50m: 2,9 + 50* 3,5/1450 = 3m (manual: 2)

@100m: 3,1m (manual: 2-4)

@400m: 3,9m (manual: 4-8)

@600m: 4,3m (manual: 4-8)

@1500m: 6,5m (manual: 6-9)

This not too complicated model would deliver with growing distances nicely stretching hit patterns that could be easily adopted.

Want less precision over the whole distance? Just increase the offsets.

The mortar is too unprecise in range at higher distances? Decrease k and increase the offset for dispersion in range.

Want less accuracy proportionally with distance because wind is blowing? Multiply k with a certain wind factor.

Want to make wind influence the hit probability overall? Add a wind factor to the offsets.

Want wind even influence the shape of the hit pattern depending from the directions? :cool: Use cos(angle between target direction and wind direction) for increasing the DispersionInRange with the wind factor and sin(angle between target direction and wind direction) for the influence on DispersionInDeflection.

To make the model super-precise, for each charge one equation could be used:

i.e.

from 50m - 300m: equation1: data for charge 0 used.

301m - 450m: equation2: data for charge 1

...

[Rob Deans]

Hi all,

I wanted to add to the conversation a bit more with some clarification. It would seem that the definitions of the "zones" and the amount of rounds that land in them are a little wide of the mark... If I may walk through this by picture....

Here is the 100% box diagram with the associated percentages marked in each PE box (have to move the decimals over two first.)

Now, if we apply a 100 round shoot @ 400yds with a 60mm Mortar (with the data from the references) according to the 100% percent box we get this... (there are 100 dots there)

Important point here..... The so called 50% and 82% zones are marked here..... They do not reflect simply 50 and 82% of the rounds... They reflect 50% of 50% of the rounds (or 25 rounds) and 82% of 82% of the rounds (or 67-68 rounds) respectively. This is because the percentages are applied BOTH for range and deflection..... The numbers don't lie..... Count the dots...

So if we remove the lines we get something like this

and this (adjusted slightly for appearance)

Now compare this to the screen shot of the experimental in game shoot (albeit of an 81mm. I am trying to find PEs for the M43 HE round...) .... Much difference and the dimensions are way off... So you can't say that 50% of the rounds will land within 1 PE of the target.... A bit of a conundrum here as the American reference states this.... But the figures tell another story (and correctly I might add)

So as far as odds go, there is a 25% chance that a perfectly aimed round will land within 1 PE of the target and a 67% chance that a round will land within 2 PE of the target. You simply can't use yards (or metres) to express the distance away from the target the round will land (ie "within 8 yds of the target", because of the nature of the PE. The only way one can use this expression is if the PE for range and deflection are the same and then by definition, the grouping on the ground will be circular)

I hope that this lends some clarity to the situation. I must say that with the references and know-how on their application, one can answer the question definitvely as to the "accuracy" of the accuracy of the mortars in this fine game.

Rob

[Steiner14]

A few more observations:

The four fields around the centre carry each a probability of 0,0625 or 6,25%.

That makes up a probability of 25% for the area around the target spot.

Uploaded with ImageShack.us

According to the given data one field @400m has the size of 8 yd. in range and 4 yd. in deflection.

The inner four fields therefore would measure 16 yd. in range x 8 yd. in deflection.

They are drawn in yellow and should contain 25% of all 48 shots (=12 hits).

In fact the area contains 11 hits. That's impressively close.

The two additional thinner yellow fields in range beside the centre yellow field should contain 8% of the hits each or 16% both of them (7,68 hits). On this picture it are 9 hits (+17%).

The two bigger yellow field in range in front and behind the target should each contain 8% or 16% of the hits combined (7,68).

On this picture it are 1 (-74%) + 5 (+30%) = 6 (-22%).

What make sthis worse, is the 5 hits in the yellow field "too long" are massed on the side close torwards the target.

It would be necessary to check if the dispersion in the picture is a valid statistical representation, but if it is, i think the reason why mortars are that extremely effective lies in the fact, that the distribution of impacts is simply too close to the target:

The dispersion very close to the target seems quite ok (the 4 PE-areas around the target spot). But the dispersion slighly off one PE-area in range is way too small and instead the hits that should go into that area, are occuring way too close left or right from the target.

Because the lenght of the PE-areas is greater and grows faster with distance than the width of the fields, it makes a big difference, if the hits are off one PE-field in deflection (4-8m off) instead being off in range one field (8-16m off).

The suppressive and even more lethal radius of the smaller calibers against dug in units is relatively small. So it should make quite some difference, if the grenades impact too often only 4-8m (in deflection) away or instead are 8-16m too short or long.

[poesel]

To get better statistics I put some 81mm US mortars to the test range. Distance is 400 yards, regular, no wind. Target is the green spot in the middle. The ground is alternating colored so its 8m between green and brown.

I used one unit with 100 shots first and then 5 units with 250 shots (total). All fired from the same spot.

The craters outside the middle are those for ranging. Every unit shoots one over and one under. The under shot for the 100 shot picture is far to the left.

100 shots

250 shots

This looks quite different from what Wreck has. Any ideas why?

[Rob Deans]

Here is my test with the PEs overlaid by grid. Generally the same as your Posel

Apart from the concentration being WAY too compact (as the numbers show), the most glaring thing is the shape of the group. That being sideways to what it should be.

Here is what it should look like

Food for thought, eh?

Rob

[Steiner14]

So far i have found some data of a french 81mm mortar:

"Accuracy of the 81mm Mle1937/31 mortar :

• 8m x 17m square at 460m range

• 9m x 32m square at 995m range

• 17m x 35m square at 1730m range

• 32m x 42m square at 2060m range "

Q: "1) I assume this accuracy data is 50% data? Is it? "

A: "The previous data about accuracy are from "L'enseignement du combat dans le groupe et dans la section d'infanterie - 2e partie" by Commandant N. Bouron. It has been printed in 1938.

It is in an appendix about the 81mm type 1930 Stokes-Brandt mortar (therefore not the 1927/1931 model). It lists examples of dispersion 'rectangles' measured on test range. I have no more details about the number of shells actully fired etc. So far I guessed that these are actual measurement for all the shells. "

http://forum.axishistory.com/viewtopic.php?f=112&t=118276

If Rob Deans is correct (or if i understand it correctly), that the so called "50% data" is the inner four PE-areas, then the data of this french mortar should be:

-------------------Probable Error-------------

------------in range [m]---in deflection [m]------25% of the shots land within

@460m:........8,5...................4,0............................17m x 8m

@995m:.......16,0..................4,5............................32m x 9m

@1730m:......17,5..................8,5............................35m x 17m

@2060m:......21,0................16,0............................42m x 32m

[Steiner14]

Comparison with the US M19 60mm we were discussing prior.

-----------------Probable Error

------------------in range [m]---in deflection [m]

@500 yd (460m).......9,1..................3,7

@1088 yd (995m).....16,5.................4,6

@1891 yd (1730m)....20...................7,3

What i conclude is, that the difference in accuracy between the 60mm US and the 81mm French at the tested distances in the game should be below 10% and the dispersion pattern should be quite similar, too.

So it seems, that the 81mm mortar in the game shows the same problem like the 60mm at the checked distances: it doesn't show a streched pattern like in reality and is too accurate.

[Wreck]

To get better statistics I put some 81mm US mortars to the test range. Distance is 400 yards, regular, no wind. ...

This looks quite different from what Wreck has. Any ideas why?

It looks like you are getting the same "tight" fall of shells that I got up to 320m. I expect that if you try longer ranges you'll see a very similar pattern.

I was testing German 81mm. Possibly US mortars are modeled as having slightly better aiming. Only other possibility I can think of is, are you sure all your crews and leaders had leadership 0? (That's what I used.) If you have some +1 or +2 leadership it might account for this.

[YankeeDog]

Does look like the groups are a bit tight.

In re: circular vs. oval patterns: bear in mind that mortars behave a little differently than lower-angle guns in this regard, and at certain range & charge settings, the pattern is circular, or nearly so -- note in the range/charge setting I posted previously, that some specific entries have an error probable in bearing that is the same as the error probable in range.

But it is true that at many common range/charge settings, the pattern should be clearly oval.

[poesel]

I was testing German 81mm. Possibly US mortars are modeled as having slightly better aiming. Only other possibility I can think of is, are you sure all your crews and leaders had leadership 0? (That's what I used.) If you have some +1 or +2 leadership it might account for this.

I tested again: all unit stats set to normal, +0 etc... 10 mortars firing from one spot, distance 850m (if someone likes to have other parameters please tell me - the firing range is set up )

The US has 500 shots, DR 280 (they carry less ammo per team). The pictures show all craters except those for ranging.

Please note that the firing direction is UP because I could get a better picture this way.

US:

DR:

This is IMHO odd. AFAIK we should have got roughly the same pattern. But the US is much tighter! Given the number of shots it is IMHO enough for the statistics (as said above: raise your hand if you need more).

[Wicky]

Would individuals in squads and crews not being so dispersed in the game and tied to action spots, compared in real life in the field, tend to balance out for the tighter spread of incoming?

[Rob Deans]

Hello all,

Great additions of tests and the like, I must say.

I have seen posed the questions of the quality of troops etc affecting the dispersion of the fall of shot. When we are able to gain the PE data for a given range and a given round, the plotted rounds according to the 100% box percentages are the very best results that can be achieved. The groups shown at 850 are WAY too compact. The PE for the 60mm Mortar (best case - Charge 2) is 6 for deflection and 13 for range.

Here is the test group. I counted 352 craters...

Here are the PEs (albeit from the 60mm)

And here is the way the rounds should have landed...... They simply don't compare...

But if you turn the group on it side........

It is about 2 PE short by range and deflection in this configuration.... Could it be that BF got their axies mixed up??? Not quite the oval shape it should be but better than the sideways group that we've seen as a result of the tests...

If mortars are to be accurately protrayed in this game, there needs to be some work done here....... Then again, there has to be some willingness to do it too..... I, for one, hope so.

Rob

[akd]

Is wind set to "none"?

[c3k]

DR = German?

[poesel]

I can't check the wind setting right now. I left it as standard but it's definetly the same for both tests.

DR = 'Deutsches Reich', so yes

[poesel]

Checked the wind and it is 'Lüftchen' (I've read this one the first time) - meaning the smallest possible wind. Using this word in this setting is quite hilarious - old fashioned word you would nowadays only use in context with farting...

[Steiner14]

@poesel

"Lüfterl" is a very common expression and "Lüftchen" is the correct written word for it. I'm wondering what people are reading, consuming or doing, if they associate normal words always with a pornographic or other disgusting meanings. Or even believe the normal and usual meaning was the wrong one.