In a series of updates around the end of 2021/beginning of 2022, off-map artillery in the game was completely overhauled. To paraphrase the developers: "It's no longer represented by mages casting fireballs". To some - it manifested as an annoying artillery cursor change, to some - the artillery all of sudden became somewhat useless, and there had been at least 4 or 5 were lengthy arguments around "god-like" precision and lack of "realism".
I will try to the best of my ability to explain why the changes are in fact "realistic" (I keep using quotes, because there is always an infinite number of things in the simulation that are not in fact realistic at all), and why the artillery's performance feels on one hand too good and on the other hand completely off.

Unless I'm linking or quoting something, this is all my understanding of the matter, and it may be incorrect in some or all aspects.

In this awesome compilation[www.britishartillery.co.uk] of info about UK artillery (which I will continue referencing throughout the guide, there is a section describing various factors[www.britishartillery.co.uk] that contribute to the artillery inability to place every single shell into the exact spot where the target is located.

There are two aspects to it: accuracy and consistency. Figure 1 demonstrates the difference between the two, and GT developers drew an even more illustrative example[imgur.com]

Originally posted by Nigel F Evans:

Accuracy means the closeness of the mean outcome to where it should be.

Mean - because every shell will hit a slightly (or not so slightly) different spot. As the explanation proceeds to say, as long as the fire is observed, the observer can correct the aiming error. For unobserved there are many factors involved - accurate maps, surveys, registration fires and naturally the correctness of target data. For those interested, there is always US Artillery Fire manual (FM 6-40).

Consistency (aka Dispersion) is

Originally posted by Nigel F Evans:

the size of the spread of outcomes (eg shells) around their MPI

In other words, once someone declares "fire for effect", and the guns aim is no longer adjusted - what happens to the shells fired? It's reasonable to assume they won't fall into exactly the same spot :) Next section will show some examples. In the meanwhile - is there one takeaway, it should be distinction between the two concepts. Accuracy is largely "external" to the gun itself: how well it can be aimed? Consistency/Dispersion is largely internal: how "repeatable" are the consecutive gunshots (not 100% internal of course - weather is an external factor)

I apologize for stealing someone's quote, but it was too good to pass up!

Back to consistency/dispersion. What was it like? Surely terrible - we are talking about WW2. But this is where to my surprise I found how my expectations did not line up with reality.

If one lobs a sufficient amount of shell towards the target to get a decent distribution of impacts, that distribution for a "regular" gun (i.e. not a tube mortar or MLRS) will look like a very skinny ellipse stretched along the vertical (range) axis. Particularly, if we take 25-pdr QF, thanks to that same awesome compilation, we learn that

Originally posted by Nigel F Evans:

For a 25-pdr firing charge 3 at 8000 yds range the PER was 30 yds (PE range - along the line of fire) and PEL was 4 yds (PE line - across the line of fire)

Looking at Figure 3 on that page, roughly 2/3 of the shells will fall into an area of 120x16 yards - when the gun is 8000 yards away. Pretty impressive, huh! Was 25-pdr a marvel of contemporary engineering? No, other guns may be better or worse, but not by an order of magnitude. For example, here is ZIS-3 Soviet 76mm gun [imgur.com] at 8km: PER = 34m, PEL = 4.2m

OK, fine, but what happens when we throw in inconsistent charges, barrels wearing out, terrible weather and other calamities? British did a study, with results in "Table 4 - Errors which affect dispersion in predicted and observed fire", and they figured about 1.8x increase in PER. Seems pretty significant, yet given very small base PER (there no detailed table for PEL, but the page cites 3 minutes ~ 7yards), it’s really not. So the ellipse is still very skinny, even after taking field conditions into account.

With all this info, if we are to look at a 25-pdr in the game, after the artillery update went it



This roughly in line with the data we just reviewed, so the game is after all "realistic".


Keep in mind, that the most significant factor for the dispersion increase is the barrel wearing out, and despite the best efforts of the developers who meticulously research each operation, I doubt that such details are easy to come by. Which means range table (firing table in US lingo) is likely to be the source of their data.

Prior to the overhaul: fireballs struck here[i.imgur.com]

By this point, you may be wondering - what in the world would poor artillerymen do if they receive an order to target an area? Their equipment is too consistent - maybe let a few screws loose to increase dispersion? Fortunately, there is a solution that does not involve sabotaging your own guns :)

Enter a concept of a sheaf - a relative position of aiming point for each gun in a battery. I am too lazy to make drawings, so I'll reuse these[www.globalsecurity.org] you'd have to use your imagination and pretend that the lines from the guns in the illustrations are parallel for all intents and purposes because the distances between these guns are negligible compared to the distance to the target

US manual from 1939[www.ibiblio.org] (FM 6-40) does not cover "standard" sheaf and has the following definition of an open sheaf:

Originally posted by FM 6-40:

An open sheaf is one effectively covering a maximum of front without sweeping

They are basically saying that given a line dispersion of a 105mm gun and the fragmentation field, 40 yard seems like a good distance between two aiming points. Any more, and you start getting less than desired effect. With 4-gun battery, the regulation says one can cover 160 yards. Narrower target would be OK too, but what happens when a target exceeds 160 yards?

Originally posted by FM 6-40:

Sweeping fire.-When a target is too wide to be covered effectively with an open sheaf, a wider sheaf is used' and each piece is traversed a definite amount after each round, usually to the left, until the front is covered. This is termed sweeping fire
...
The data necessary are distribution: number of rounds sweeping; amount of sweep or traverse
after each round

You turn the guns by a predetermined amount and cover the desired front in one or more turns. There is a very rough calculation There is a rough calculation for it:

Originally posted by FM 6-40:

amount of each shift is determined by dividing the difference between the width of the target and the width of area covered by an open sheaf, by one less than the number of sheafs required to cover the target. The result is converted to mils at the target range.

Similar idea for a deep target that covers depths of 100 yards and more. For 105mm gun, taking the approximate dispersion into account, the regulation prescribes 3 "ranges" for 100m (i.e. shift the aim 50 yards 2 more times from the initial target line).

This is meant to be "neutralization" against troops in the open, hence such optimistic "cover the area twice is enough" assumption. The definition of neutralization:

Originally posted by FM 6-40:

Neutralization. Fire delivered on areas to destroy the combat efficiency of enemy personnel by causing severe losses and interrupting movement or action. Neutralization is established by deliver- ing surprise fire in intense masses. It is maintained by intermittent bursts of fire in lesser amounts.

There is a passage indicating how more missions would be fired if the effect needs to be maintained. Arguing whether the prescribed amount of fire meets “intense masses” or not is outside of the scope for this guide

And there you have it: to cover larger area, guns are shifting aim points at set intervals.

Armed with the knowledge, let's look at the game now.

Here is me doing some divine gardening:


Naturally, I am annoyed, because I've not yet read through a 200-pages artillery manual and the game (as it always does) let me fail miserably. It does not help that just before the update I would get this instead in a similar situation:




But now I know! So let's set up a 200 yards concentration mission with 105mm howitzers. That involves 2 sweeps (shifted by around 30m) and going over 5 ranges (50m each). So 10 missions in total just to cover one area. Not at all streamlined, but possible.


I assume we first sweep then move the ranges, not sure that's accurate, but it does not matter for the sake of the exercise. It's impossible to order less than 3 shells per gun - again, close enough.
Here is what ends up happening:


Somewhat sparse, but since the mission's goal was to scare a bunch of infantrymen who were assumed to be staying out in the open, this is fine. For a better effect, more shells are needed.

An interesting observation from this experiment is that a battery is expected to complete this mission in 5 minutes, whereas in game it took 20 (which seems too much if I take 5 as gospel, even with 50% more rounds fired). However the regulation lists "the times are based on the maximum permissible rate of fire for short bursts"

I would also question whether the computer opponent is capable of doing something like this. It would be hard to test though.

If I had my wish, this would be streamlined - setting the area, desired effect and … profit. In which case this entire section would be rendered moot :)