[ADV] Advanced Play System - Questions & Commentary

[Radoye]

The preceding notwithstanding, I'm mystified by the programming underlying the MOV = 0 impact on purchased unit placement possibilities. I just can't see any logical connection here !

Yeah it's a weird one, if it was just a plain overflow one would expect it to become 255

[Lettos]

Elsewhere in this forum, # Lettos # suggested employing an "interim" scenario in which Disbandment of selected Core units would be possible.

SSI was the first to propose such scenarios.

There are three odd scenarios in the classic 39 PG1 campaign: Norway, Crete, Sevastopol.

It is clear that Crete was introduced mainly from a commercial point of view.

But Norway and Sevastopol are examples of "interim" scenarios.
In the PG1 engine there was no upgrade option between scenarios. And it was required to fully simulate the upgrade of the Wehrmacht in the period after the Polish campaign and before the French campaign, as well as the upgrade of tank units in the spring of 1942.
Norway and Sevastopol give a large amount of PP, EXP and allow the necessary upgrades to be made.

It is possible to make a tank range with plywood targets. Interim Scenario for 5-10 turns. Player units gain some EXP by destroying plywood targets with SA=0 HA=0. At the same time upgrades are carried out, units are disbanded and others are created according to certain agreed rules. Cheats are used to add PP to the disbanded units and transfer EXP to the new units. All of this can be implemented in the PGF2100 UI as well. But as long as this is not available, you can get by with destroying plywood targets.

[Radoye]

I believe aircraft carriers can initiate attacks and attack after move. It's been a while since i looked into that so i don't remember the details but for some carriers i added their proper HA / SA / NA values according to the armament they carried (guns up to 8 inch - basically equivalent to a heavy cruiser; some were also armored to the same degree as battleships or battlecruisers, so quite capable in surface-to-surface combat had it come to that).

I haven't experimented with movable forts though.

Another interesting but unrelated thing that i just remembered that i experimented with ages ago - if you have Fuel = 0 for all ground transport units but then add fuel to all transportable units (the lighter the unit the more fuel it gets), you can simulate the differences in transport speed when different types of units are being carried by how often you'd need to resupply them for fuel, since the transport unit "borrows" the fuel value from the unit it's transporting.

Which is all nice, but it breaks down with naval and air transports - if you run out of fuel while on naval transport while out on the ocean you can't resupply; if you run out of fuel while embarked on an air transport, the game crashes (at least it used to in PG(DOS), when i last looked into these things). So unfortunately it was a no-go in the end.

[Lettos]

I believe aircraft carriers can initiate attacks and attack after move. It's been a while since i looked into that so i don't remember the details but for some carriers i added their proper HA / SA / NA values according to the armament they carried (guns up to 8 inch - basically equivalent to a heavy cruiser; some were also armored to the same degree as battleships or battlecruisers, so quite capable in surface-to-surface combat had it come to that).

I haven't experimented with movable forts though.

Another interesting but unrelated thing that i just remembered that i experimented with ages ago - if you have Fuel = 0 for all ground transport units but then add fuel to all transportable units (the lighter the unit the more fuel it gets), you can simulate the differences in transport speed when different types of units are being carried by how often you'd need to resupply them for fuel, since the transport unit "borrows" the fuel value from the unit it's transporting.

Which is all nice, but it breaks down with naval and air transports - if you run out of fuel while on naval transport while out on the ocean you can't resupply; if you run out of fuel while embarked on an air transport, the game crashes (at least it used to in PG(DOS), when i last looked into these things). So unfortunately it was a no-go in the end.

Aircraft Carrier

I put some Aircraft Carriers with SA=10 HA=10 Range=5 in the puddle. Nearby I put some Battleships with same fire range. Infantry on shore at distance one hex from shore. Infantry sees Recon.
The Battleship sails up to fire range and attacks. The Aircraft Carrier stands in place.

Place the Aircraft Carrier next to the shore, and the enemy infantry next to it. The Aircraft Carrier just swam away from the infantry in a puddle. A kind of herbivore animal with big horns and teeth that does not attack anyone.

Fort

Assigned the fort MA=3 MT=4 (default)

Fort moves and shoots, shoots and moves after Fighters. Behaving very aggressively . Behavior is similar to the Capital Ships class.

[Radoye]

Are you sure your recollections don't date back to some SSI wargame ?

It's possible that i'm thinking PacGen

[Lettos]

Great!

The electronic toy soldiers will learn to dig, and I will get used to my new weapon, the hex-editor!

Now they dig like this:


I think ideas on how they should dig would be better expressed in the Ephemeral posts thread.

[Lettos]

Yes, after saving you can upload a new equipment file.

After loading a new file, the unit's AMMO and FUEL parameters remain the same as they were at the time of save. That is, these parameters may be more than MAX AMMO and MAX FUEL.
But Supply (using button) or Re-Supply will already happen to the new (e.g. reduced) AMMO and FUEL values.

Do in the H2H scenario with Home rules with supply lines emulation?
It's very easy!

Put some units simulating warehouses or supply lines on the map. Prepare in advance several versions of the Equipment file, containing progressively reduced AMMO and FUEL values.

For example, some unit has AMMO = 8, FUEL = 40 as standard.
There are 4 unit depots on the map. Destroying each warehouse results in
AMMO minus 2, FUEL minus 10.
Prepare 4 "downgraded" equipment files with the values of the parameters of the selected unit:
AMMO = 6 / 4 / 2 / 0
FUEL = 30 / 20 / 10 / 0.
If at the moment of destroying even the fourth warehouse the unit has, for example, AMMO = 5, then the unit will spend this AMMO (and/or FUEL) as usual. Only then, when AMMO/FUEL becomes "0", the unit can no longer get either.

Funny, isn't it?
(I'm developing "Army Exercise 2" now, and I'll be sure to include this trick in the game conditions).

Or, for example, for H2H we make a scenario for 8-16 two-hour turns (1-2 days) with no option to buy units, with a starting equipment file, and after running the scenario we copy in the \Scenario\ equipment file with zero AMMO and FUEL values.
And play until the units' batteries are depleted.

By the way, in case of successful combat operations, the side that lost this unit-stock and then recaptured the territory can buy such a unit. And put it on the same place where it was at the beginning of the scenario.
And thus gain the right to load an "upgraded" version of the equipment file.

[Lettos]

I experimented a bit with editing the SAVE file.

Thanks to the way Hexcode works viewtopic.php?f=100&t=553#p9068

even an inexperienced amateur like me can recognize the contents of a file.


One very strange thing came up.Editing the file - so it seems, any byte in it - leads to the inscription "REPLAY FINISHED" at the end of the scenario.

I have tried changing bytes in the following sections:

Section 2.T -- NDS 2.T.3

AI Posture & Unit Icons Orientation
This data segment's is 8 bytes long.

Code: Select all
0T 00 00 00

AI STANCE. Boolean ("T" is either 0 or 1). Intended values and their meanings follow:

"00 00 00 00 00" ==> AI IS "DEFENDING"
"01 00 00 00 00" ==> AI IS "ATTACKING"

Minor correction: I think it should be SIDE STANCE here, not AI STANCE. But that's the little things.
Another strange thing: the behavior of the AI does not change in scenario from changing 00 to 01 and vice versa.

UNIT ICONS ORIENTATION.

Important. In Section 2.0, the value is always "00 00 00 00". In Section 2.1, the value is either "01 00 00 00 00" or "FF FF FF FF FF FF". There are only two "East-West" unit icons orientation options which are mirror images of each other.

Changing the direction of the icons even on one side results in the same "REPLAY FINISHED" caption.

I tried replacing the unit with a different one.Everything works fine, the new unit appears on the map instead of the old one, gets AMMO and FUEL from the replaced old unit, but at the end of the script we see the same "REPLAY FINISHED"...

It turns out that the SAVE file is somehow protected from hex editing? For example, a mismatch between the computer time when the SAVE was made and the save time of the modified file? Or something more complicated?

[Lettos]

I don't believe this is due to edits. Kindly consult

What ? Did I Not Just Win ?
viewtopic.php?f=100&t=554#p9099

for a... descriptive explanation. It's a minor "wart".

... the behavior of the AI does not change in scenario from changing 00 to 01 and vice versa.

It so happens that the Game-State file's internal structure contains redundancies which are, well, duds !! To change the AI Module's behavior you'll have to hex-edit the pertinent values here:

*.PGSAV (SECTION 15)
AI Behavioral Instructions
viewtopic.php?f=100&t=553#p9092

I was afraid to assume yesterday that this was the case. But assumed

I did some experiments. These three parameters:

UNKNOWN ASSIGNMENT. "T" can be 0, 1 or 2. Seems to affect which city centers will be "surrounded" by defensive units.

UNKNOWN ASSIGNMENT. Probably a Boolean value ("T" can be 0 or 1). Always the opposite of the value that affects the unit's movement behavior.

AI "RECKLESSNESS" ATTACK. "T" may be 0, 1 or 2.

either do not affect anything, or affect only under some specific conditions, which I was not able to recreate in the experiments.

But changing this parameter

AI DOES NOT RUSH ITS UNITS FORWARD. Boolean value ("T" equals 0 or 1).

immediately manifests itself in the next half-turn.

You can manipulate the AI in some scenario by editing the Save file.
For example, in the briefing to tell the player: after your 7 half turn save the game, find at the end of the Save-file
03 00 00 00 02 00 00 00 01 00 00 00 and replace with "00".

The AI will immediately rush forward at the Player with almost all available units. So called Counter-offensive

If after a few turns you change "00" to "01" again, all AI units will run back to their cities.

[Lettos]

1) Personally, I choose to ignore the word "Cheat". As I recently wrote elsewhere in this forum, "playing" is a strictly private affair.

Agree. "Scenario tweaking codes"?

[Lettos]

I decided to test how many times a unit's real power increases with Over-STR and EXP.

1st Experiment "Tanks vs. tanks"

The experiment involved the same type of tank for AXIS and ALLIED.
Terrain is Clear.
The ALLIED side had 100 units on the field with standard STR=10.
I simulated a tank battle in which the tanks of each side both attack and defend. ALLIED tanks are inserted into the battle in portions of 10-20 units.
AMMO = unlimited.
I did not attack surrounded units. I tried to move the units so that the ENT was not greater than "1".

AXIS side in the experiment had 10 units, for which I changed STR and EXP when deployed.
The data below shows how many ALLIED units were destroyed before the 10 AXIS units were completely destroyed:

AXIS (EXP Cap in Scenario = 0):
10 x STR=10 EXP=0 ---------------- 10 ALLIED (quite logical)
10 x STR=15 EXP=0 ---------------- 18-19 ALLIED (almost 20, but slightly less)
10 x STR=20 EXP=0 ---------------- 30-40 ALLIED (about 35)

In the case of AXIS STR=15 theoretically should be 1.5, for STR=20 - 2.0, but in fact we have for STR=15 only 1.2, and for STR=20 - slightly less than 2.0.

Adding Experience (EXP Cap in Scenario = 100):

AXIS:
10 x STR=10 EXP=1 -------- 15-18 ALLIED
10 x STR=15 EXP=1 -------- 50-55 ALLIED
10 x STR=20 EXP=1 -------- 70-80 ALLIED

Imagine a scenario where you want to increase the density of one side's tanks on one hex. As Middeldorf wrote: maximum 200-250 tanks per 1 km of frontline. This is realistic in PGF even without Stacking!
Based on experimental results, it turns out that using such scalability tools as EXP and STR, we can create a Stack of up to 7-8 times the strength of a standard unit. Of course, we won't then be able to split this Stack into smaller units. As always, the usual patch on the old dress But it's better than nothing.
At least it became clear from a mathematical point of view, what the author of some script is doing, arranging the "experience" asterisks to his favorite units.
Want to know my personal opinion? Sometimes this over-EXP can help to show something in a scenario that wouldn't otherwise be shown in PGF.
But regularly making units 3-5 times stronger than the enemy is sometimes not a good way to win a game


2nd Experiment "Guns vs. Infantry"

10 AXIS gun units, Class=ATY, with SA=6, fire 4 times on infantry units ALLIED STR=10 GD=6 ENT=1.

Looking at ALLIED losses:

AXIS:
10 x STR=10 EXP=0 ------- ALLIED losses = 16-17
10 x STR=15 EXP=0 ------- ALLIED losses = 29-40
10 x STR=20 EXP=0 ------- ALLIED losses = 32-38

Jokes from Mr. Random But all in all, everything is expected and acceptable.

Let's add EXP to the ATY units.

Table of losses of 30 infantry units ALLIED, depending on STR and EXP of AXIS guns:

-----------EXP=0-----EXP=1-----EXP=3-----EXP=5
STR=10----16---------40----------67---------122
STR=15----34---------67---------128---------176
STR=20----35---------77---------174---------230

Mr. Random kept joking around with STR=20 for a while, but then he stopped.

I always thought EXP for ATY was a bonus to INI, which is of very little importance to ATY, and the ability to over-STR for guns.
But I've seen in games that guns with EXP>1-2 become truly killer for the enemy. I didn't understand why that was the case. Thought it was only because of STR>10.

It turns out that EXP for ATY is something like fire accuracy. In this particular case, it gives you the ability to scale the amount of artillery on one hex by a factor of 15!
That is, from the usual battalion of guns (12-18 pcs.) you can make the same 300 guns per kilometer of front, which were often used in the offensive operations of 1943-45 in the most critical sections of the breakthrough?

Perfect!

===========
Added:

EXp values >599 not affects ATY unit's real strenght.

Updated table:

-----------EXP=0-----EXP=1-----EXP=3-----EXP=5-----EXP=599
STR=10----16---------40----------67---------122--------136
STR=15----34---------67---------128---------176--------210
STR=20----35---------77---------174---------230--------250

[Lettos]

The 3rd experiment "Ranged fire on Soft/Hard target"

Soft target: 40 INF Units, GD=6, ENT=1
Hard target: 40 Tank Units, GD=6, ENT=1

All unit classes shoot 4 times. Firing/bombing units (10 units) always have STR=20, only their EXP changes.
The losses inflicted on targets are shown in the table which is a very useful tool for comparing SA and HA of different unit classes:


--------------------EXP=0-----EXP=1-----EXP=3-----EXP=5-----EXP=599
ATY------------------35---------77---------174---------230--------250
Capital Ship--------?-----------?----------168---------236
Destroyer-----------?-----------?-----------?-----------244
FORT-----------------?----------80---------151---------239
Level Bomber-----40----------?------------?-----------201
-------------------------------------------------------------------------------
INF-----------------307---------?-------------?----------463
Recon--------------270--------279---------406--------486
TANK---------------274---------?------------?----------474
AA------------------262---------?------------?----------484
Land Tr-t------------?-----------?------------?----------500
ATG----------------206---------251---------322-------417 (ATG attacked INF units! with EXP=4 losses=348)
ATG----------------272----------?------------?-----------? (ATG attacked ATY units!)
Fighter------------237----------?------------?-----------?
Tac Bomber-------229----------?-------------?--------447

The table clearly shows that we have in PGF two different methods of calculating target loss.

1 - for ATY, Capital Ships, Destroyer, Fort and Level Bombers

2 - for INF, Tank, Recon, AA, Land Tr-t, ATG and Tac Bombers and Fighters

Only the ATG Class has its own special feature - some penalty for combat contact with an INF unit.

Hard Attack units are fully matched to the Soft Attack table. I did not notice any exceptions.


I've checked how target's EXP level affects their losses in case of Ranged fire to target.

If the target has EXP=500, its losses are reduced by about half. I have seen various results that range from 1.5 even to 4 times less than the losses of the same unit with EXP=0. On average, losses are reduced by something like "2" times.


And some small observations:


I don't know if anyone has investigated the Bombers Spec parameter for Level Bombers before.

This parameter can be any number, e.g. "33", "66". It has no effect on target losses. But it does affect how much AMMO and FUEL the attacked unit has left after a Level Bomber attack.


About the "Ignores ENT" parameter.

Only ATY Class cannot ignore ENT. All other classes can. For battleships and forts with large-caliber guns, as well as in the case of large-caliber artillery relocation to another class, ignoring ENT is very important.

[Radoye]

I don't know if anyone has investigated the Bombers Spec parameter for Level Bombers before.

This parameter can be any number, e.g. "33", "66". It has no effect on target losses. But it does affect how much AMMO and FUEL the attacked unit has left after a Level Bomber attack.

Yes, in PacGen this is actually called "Bomber Size" and corresponds to the level bomber's bombload. Beside taking off ammo and fuel it causes long lasting suppression (it lasts for the entire duration of the turn) so making level bombers ideal for "softening up" stubborn heavily entrenched units - after they're attacked by a level bomber every subsequent attack during that same turn will start with the unit already partially suppressed. Casual players don't know this and only look for direct kills so often level bombers get overlooked, but this feature actually makes them very powerful and very useful.

[Lettos]

We are unlikely to know the formula for calculating the target's attack losses. But it is not very necessary if we see the results of this formula.

A large test has been performed. 40 attacks on units with STR=10, ENT=1.
I changed the SA parameter (it could have been HA) of the attacking unit, and the GD of the target. Total losses of attacked units for 40 attacks are shown in the table.



The green field is, conventionally, the maximum losses an attacking unit can inflict. The variability here is in the +50% range. Even if a unit with GD=0 is attacked, it won't be completely destroyed in the stats even if a super-strong unit with Attack=18 is attacked. Well, that's logical - the attacked just run away somewhere.
You can see that when attacking units with low GD, the game mathematical model almost doesn't care what attack parameter the attacking unit has. That makes sense, too. How many shots does it take to blow up a truck with a load of fuel?

The yellow field with the pink diagonal is the EQUIPMENT file designer's field. Only in this space does the loss factor change a lot. The yellow field is tightly constrained to the smaller and larger sides of the diagonal Attack=GD:
-1 --- Attack=GD --- +5

Going outside of this field will result in the creation of a Uber-unit.
For example, a tank with GD=20 will be practically unkillable if its opponents' HA is 13 (i.e. going beyond the "6" difference)

Entrenchment is the same units as Ground Defense.
It makes no difference to the game model whether you attack a unit with GD=8 ENT=4, or GD=4 ENT=8. The formula only uses the sum
8 + 4 = 4 + 8 = "12"

If the attacking unit has an attack parameter of 12 - 6 = "6", it will be able to inflict tangible losses. If the attack parameter is less than "6", then both the 8 + 4 and the 4 + 8 will go beyond the yellow field into orange, and the attacked unit's losses will be minimal.

Experience

The EXP level of an attacking unit is equivalent to attack units. That is,

Attack Strength = Attack + EXP Level

The EXP level of a defending unit is equivalent to Ground Defense units. That is, the defender unit's actual

Ground Defense = Ground Defense + ENT + EXP Level


Minor Observations

Formula in PGF vs PG1

I compared the action of the formula in PGF and PG1. I didn't investigate in depth, but I didn't see any difference on a specific example. It is very likely that the formula is identical.

ATG vs INF Penalty

Applied even in case when ATG attack INF with ranged fire.

AMMO = 0

5 identical tanks STR=10 attack the exact same enemy tank STR=20 and AMMO=0. It took an attack (depending on the EXP level of the tank with AMMO=0) to destroy the enemy tank:

EXP=0 - 6 attacks
EXP=1 - 8 attacks
EXP=2 - 8 attacks
EXP=3 - 17 attacks
EXP=4 - 13 attacks
EXP=5 - 43 attacks

With an EXP difference of "500" the mathematical model of the game becomes illogical. However, scenario designers have long known this from practice and do not make scenarios with a level difference greater than 3 stars.

[Lettos]

The foregoing initial value (except when it is ZERO (0)) is then modified by PLUS ONE (+1) for every Experience Level (EL) the effectively shooting unit enjoys.

Tested. In PGF, even if a unit has Ground Defense = 0, EXP Level = 5 gives a 5 point increment to Ground Defense.

Only Air Attack/Air Defense remain untested.

If the targeted unit is on a River Hex or the effectively shooting unit puts up a Rugged Defense, ADD FOUR (+4) on top in each case.

For Rivers: checked and approved.
For Rugged Defense: It remains to be believed that, too, "-4". It is impossible to check. It does not have much practical value. It is enough to know that "minus something"

1) Any unit other than a Tactical Bomber or Destroyer Class one attempting to shoot at an enemy Submarine Class unit.
2) Any unit other than an Air Defense, Anti-Aircraft, Fighter or (Fighter-)Bomber Class attempting to shoot at an enemy Air Target.
3) Either unit being an Air Super-Class one during Rain or Snow.
4) Any unit which is out of Ammo attempting to shoot.
5) An Air Defense Class unit attempting to initiate combat against an enemy non-Air Target.
6) Any unit being the target of a Ranged Attack unless this an engagement between opposing Capital Ship Class units.

I hope you don't doubt that the PGF has it all?

B) When it comes to PGF, well, there're no guarantees...

Now there is.

I found out how much the penalty is for ATG attacking INF. "-2"

[Lettos]

A) PLUS EIGHT (+8) if a Ground Super-Class unit is being situationally shot at by a Naval Super-Class unit.
E) PLUS THREE (+3) if the situationally shot at target is an Artillery Class unit.

Answered here: viewtopic.php?f=95&t=470&start=50#p10726

The table clearly shows that we have in PGF two different methods of calculating target loss.

And at the same time here is one formula for ATY and Capital Ships.

That is, there can not be +8 and +3 to GD if the practical firing shows the complete identity of ATY=Capital Ships.
And there is no one common "bonus" for GD of a fired target compared to the Tank-Recon-ATG-AA-TacB-Fighter classes that would turn one table into the other. These are two different formulas.

B) PLUS THREE (+3) if the situationally shot at target is subjected to Ranged shots under Rain or Snow.

It's a very difficult question. So far, according to the experiments, it turns out that:
Rain = +3
Rain + Muddy = about +7
Clear + Muddy = +4 (or even +5).
So here is GD bonus for Muddy and very probably also for Snow.
I will continue my experiments.

C) PLUS ONE (+1) for each Experience Level the situationally shot at target enjoys.

Already APPROVED above.

D) PLUS FOUR (+4) if the situationally shooting unit is situated on River terrain.

APPROVED!

F) PLUS FOUR (+4) if the situationally shot at target is favored by a Rugged Defense event.

Very hard to check it. From practice: yes, there is a serious bonus to GD. Maybe even much more than +4.

G) PLUS the target's current Entrenchment Level if the situationally shooting unit is NOT an Infantry Class one.

Already APPROVED above.

H) PLUS ONE HALF (+ 1/2) the target's current Entrenchment Level if the situationally shooting unit IS an Infantry Class one OTHER than a Rugged Defense avoiding type.

APPROVED!

I) PLUS TWO (+2) if the situationally shot at target is an Infantry Class unit AND the situationally shooting unit is an Anti-Tank Class one.

Already APPROVED above.

[Lettos]

B) PLUS THREE (+3) if the situationally shot at target is subjected to Ranged shots under Rain or Snow.

It's a very difficult question. So far, according to the experiments, it turns out that:
Rain = +3
Rain + Muddy = about +7
Clear + Muddy = +4 (or even +5).
So here is GD bonus for Muddy and very probably also for Snow.
I will continue my experiments.

Rain = +3 to GD (including Naval attack to Naval target)
Muddy = +4 GD
So Rain + Muddy = +7 GD (sometimes due to Random from +6 to +8)

Snowing = +2 GD
Frozen not affects GD

APPROVED

[Lettos]

F) PLUS FOUR (+4) if the situationally shot at target is favored by a Rugged Defense event.

Very hard to check it. From practice: yes, there is a serious bonus to GD. Maybe even much more than +4.

Much more like a "+5" than a "+4". Almost APPROVED, since the data is only for 20 collisions with Rugged Defense.

[Lettos]

Overcast Weather Conditions.

- DOES NOT AFFECTS the air unit's air defense if it is attacked by an air unit.

- gives +4 to the ground target's GD. I guess it should apply to Naval targets as well.
- gives +4 to Air unit's GD if it attacked by Air Defense unit. I guess it should apply for Anti-aircrafts attacks as well.

[Lettos]

I'll summarize my research.

Supplemented tables of total losses of 40 units STR=10 after 40 strikes.
The defending unit has ENT=0. EXP=0 for both defending and attacking units. EXP Cap is set to "0" in the Scenario. Weather "Clear (Dry)"





The table for ATY can of course be supplemented. But so far I am tired of clicking, and there is no point in clicking anymore. Thanks to the table for Tank etc, everything is very clear for ATY.

The attentive analyst already sees in the green field of the table possible break points (salient points) in the charts.
The graph is a stepped line on the green field, an unbroken oblique line on the yellow field, and a horizontal line on the orange field. Everything is heavily spiced up with a random number generator.
But I don't see the point in taking the time now to reconstruct the combat collision formula exactly as it exists in the PGF.
Mathematics allows you to draw the same thing on a graph using different formulas. It is not so important now, if we see the result of the formula.

Free Bonus:
Comparison of combat formulas in PGF, Panzer General 1, and Open General


PGF is almost identical to PG1. OG is very different, especially in the "Tank" table.

I didn't set out to study OG and PG1. I only needed the data to compare the three mathematical models.

But it turned out in passing that:
- In PG1 Capital Ship shoots the same formula as ATY, meaning that those "+8" and "+3" to GD from Prima Guide don't even match PG1.
- In OG Recon is also in the "Tank" table
- in OG AI ver.0.91 the formula is the same as in 2.0
- the OG formula has a weird big step, in the GD table 12-13 (losses 141 vs 32). The same step would be at SA=6 and GD=6-7.

I have completed statistical analytical experiments. What is required is a historical and philosophical reflection on the results of the formulas in terms of real and assumed historical military reality.

[Lettos]

Frankly, Prima's text is woefully deficient; quite a few unambiguous technical descriptions are sorely needed here to render it just passable...

Absolutely right!

The Prima Guide attempts to write the specification literary. But the result is neither mathematical clarity nor an easy-to-read literary text.

The spec should be clear, concise in individual sentences, and yet boring to all readers except game engineers.

The Guide should contain what is of interest to the game reader. And the player is only interested in two criterions: the loss of an enemy unit in certain player actions, and, naturally, what price you will pay for the expected losses of the enemy.

For example, a combat formula can be described as follows:
If your unit's attack parameter is greater than its defense parameter by 6 units or more, you will strike hard with a target loss of 2/3 or more.
If the difference between your unit's attack parameter and your opponent's unit's defense parameter is less than 6 and up to zero, the enemy unit's losses will decrease linearly as your unit's attack parameter drops.
If the defense parameter of the enemy unit is greater than the attack parameter of your unit, then the losses inflicted will be extremely insignificant (no more than 1/20).

A unit's EXP stars give its own bonuses to each side, both in attack and defense.
The enemy unit's ENT always works against you.

All of the above applies to your unit that will be attacked by your opponent.

It's just an impromptu sketch, nothing more.

[Lettos]

I'd like to be able to edit integer constants encountered within active code subroutines, whenever appropriate. That's "Deep Modding" for you. When it comes to Prima's technical information, it may or not be perfectly applicable to PGF. Only careful, controlled experimentation can provide reasonably solid guarantees. Once definitive conclusions are reached, then and only then it would make sense to attempt to identify the relevant hexadecimal code segments.

We'll definitely talk about that. About constants, what they can and cannot change in principle. About experiments, too. But a little later, in a few days.
I have a time-out this weekend. After the weekend, Monday-Wednesday, I want to publish all my observations, comments, and thoughts on these very sacred formulas.

[Lettos]

Example: A Tank Class unit sporting two (2) Experience stars initiates an attack against an enemy Infantry Class target sporting one (1) Experience star and being the beneficiary of four (4) Entrenchment Levels.

<...>
EntRR = (3 + 1) / (1 + 1) = 2

It follows from your practical example that EntR is a UCER
viewtopic.php?f=100&t=544#p8987
Is this true, or am I wrong?

[Lettos]

Look, from a statistical analysis standpoint, 20, 50 or even 100 trials aren't enough for one to conclude anything, especially without controlling for the pseudo-random "seeds"; hence, the observed extreme variability in small samples...

Before I forget, PGF's predictions are quite "primitive" in that they... unabashedly depict RD chances higher than 100% ! As I've mentioned many times in this forum, for the most part, PGF's programming isn't anything to write home about when it comes to all kinds of details; PGF's engine is anything but "polished"...

You may have noticed that I was checking the extreme unstable values. For such tests, 50-100 measurements are enough.

But, if that's not enough... and it feels like it is, because I can't convince you with measurements at the unstable boundary values of the formula, then... the courageous clicker Lettos made a test of 1000 attacks.

INF TEntL=1 and TExpL=1
Tank AExpL=1
It mean 10% in PG1, and PGF also predicted same 10% with CTRL+LClick.
20 tests with 50 attacks in each, with PC restart after 10 tests.

Important note: UCER for INF Class changed in exe from "3" to "5".

Test number - Rugged Defense cases
1 - 4
2 - 6
3 - 3
4 - 3 (33 attacks without RD)
5 - 2
6 - 10
7 - 4
8 - 5
9 - 5
10 - 9 (4 RD during 12 attacks!)
--------
Total in tests 1-10 : 51 RD in 500 attacks

11 - 4 (36 attacks without RD)
12 - 5 (38 attacks without RD)
13 - 1 (!) (40 attacks without RD)
14 - 5
15 - 4 (about 40 attacks without RD)
16 - 6
17 - 5
18 - 6
19 - 8
20 - 5 (3 RD in 3 attacks!)
------------
Total in tests 11-20: 49 RD in 500 attacks

I didn't fake anything. I just clicked.

51 + 49 = 100. 10%.


I predicted for myself something like 96, 105... But to be that accurate?! Clearly, it's a fluke within a given boundary.

Will you try to draw a graph of the Gaussians? It will be something between the red and orange curves. It's too uniform a distribution to be unaggressive

[Lettos]

However, my sole experiment so far (100 trials) came close to PGF's prediction. I'll experiment some more and I'll let you know.

There is no need to continue with these tests. Everything is very clear: if TEntR in formula will be "5" instead of "3", we'll see 10% predicted.
I replaced the clean labware with dirty labware, and I got the wrong result.

I apologize for your wasted time!

"1". Correct is 2/3!

[Radoye]

Limited Screen Display Space

Aiming at full, unobstructed visibility, UT alphanumeric descriptors CANNOT be longer THAN NINETEEN (19) characters. In fact, to the extent that a unit enjoys Organic Transport, its alphanumeric descriptor CANNOT be longer THAN FIFTEEN (15) characters; reason being, the requisite, tiny icon to the immediate right of the descriptor is also displayed in the Scenario Panel.

There are ways around this, if one is willing to get their hands dirty

It is possible to edit the game UI so that it provides more room for longer unit names. At the moment, i have a working version (not yet released) which allows a full unobstructed display of unit names 21 chars long, including the little transport icon.

UI files are HTML with CSS, fairly easy to modify, and it is possible to modify the dimensions of the various panels shown on the screen. The one of a particular interest here the file main.htm. You can go crazy and experiment, some quite interesting (if not always desirable) effects can be achieved.

[Radoye]

The hypothesis is absolutely correct - check my Torch PGF WAW scenario for an example, where i have (Vichy) French owned cities which can be captured by both sides (nation #7 is not listed either as Allied or Axis in the Nations section) but where i placed some units with Flag ID #7 and assigned them to Allies (to represent the Tunisian Vichy troops which resisted the Axis order to disarm). Since nation #7 is not listed as either Axis or Allied, neither side is capable of purchasing units for it but the pre-placed "Allied-Vichy" units work just as any other Allied unit in the sense they can initiate combat against Axis units and capture Axis owned hexes but when they do so they flip the flag to #7 and the said city later behaves like the other nation #7 owned cities already on the map - neither side can "see" them and both sides can capture it (as long as the unit is not listed under nation #7). One additional quirk in all this, which IMO in this case works quite appropriately for the scenario, is that the #7 owned victory hexes are showing the white Axis ownership marker when viewed on the strategic map even though as i mentioned they actually behave as neither Axis nor Allied.

I was planning to follow up on this with further experimentation (like, what happens if we have Neutral owned victory hex which can't be captured by either side? Or what happens with the hex ownership if we have the same nation listed as both Axis and Allied? What happens if we only have units on the map which belong to nations not listed as either Axis or Allied for the particular scenario but they're assigned to their respective side?) but there's never enough time...

[Radoye]

Yep, it appears the ownership / capture entirely depends on the flag that is set on a particular hex, and this can be set on any hex regardless of the type.

[HexCode]

TEMPORARY APPEARANCE...

Intended Audience: # Lettos # & # Radoye #

Introduction

In the past, our Moderator and I have exchanged findings and opinions regarding certain innovative uses of the Structure Class. In particular, much of our interaction has revolved around the "construction" of Unit Types which never voluntarily move or involuntarily retreat as a result of combat.

Relevant Play System Feature Description

Such Unit Types CANNOT take Strength Factor (SF) Replacements of any kind unless they are located on friendly, NON-NAVAL CITY hexes. It's very important to note that Port Cities / Facilities and Embarkation Cities simply won't do. They HAVE to be Non-Naval Cities or else...

Is the Feature Relevant to Custom Content Design ?

Only in PGF-CDP territory !

[Radoye]

Interesting...

[HexCode]

TEMPORARY APPEARANCE...

Intended Audience: # Lettos # & # Radoye #

Introduction

In the rather distant past, our Moderator and I exchanged findings and opinions regarding certain innovative uses of a Unit Type's Spotting Range (SpR). Initially, the discussion revolved around the possibility of somehow "constructing" Unit Types which don't exhibit ANY spotting capabilities whatsoever. The conclusion reached was that PGF engine's programming just CANNOT support this designer's potential wish.

The discussion then moved on to Unit Types sporting SpR ZERO (0). SSI's "flagship" PG1 / AG content does NOT feature any such Unit Types.

Relevant Play System Feature Description

1) Under CLEAR SKIES, the Unit Types spot enemy units ON / OVER the hex in which they are situated ONLY. When they move, their spotting path / trail consists of a ribbon, ONE (1) hex wide, which strictly coincides with just the hexes they traverse.

2) Under OVERCAST SKIES, the Unit Types behave as per preceding point (1) with ONE NOTABLE EXCEPTION. Namely, Unit Types which are AIR TARGETS spot enemy units which are ADJACENT to them as well. When they move, their spotting path / trail consists of a ribbon, THREE (3) hexes wide.

3) Under PRECIPITATION (i.e., RAIN or SNOW), the Unit Types spot enemy units which are ADJACENT to them as well. When they move, their spotting path / trail consists of a ribbon, THREE (3) hexes wide.

Is the Feature Relevant to Custom Content Design ?

Only in PGF-CDP territory ! Well, seriously, it depends. There's an obvious issue of reality misrepresentation here. How can it be that increasingly inclement atmospheric conditions are accompanied by improved spotting capabilities ?

Okay, this feature is a splendid example of what happens when a wargame Developer / Programmer codes his engine to specifically support his pet content as opposed to preserving logico-mathematical consistency. Elsewhere in THESE OG Forums I've made the case for PGF de facto serving as an accidental Content Design Platform (CDP) as opposed to a consciously intended one...

[Radoye]

Ok, this feature is a great example of what happens when a wargame Developer / Programmer codes his engine to specifically support his pet content as opposed to preserving logico-mathematical consistency.

Indeed.

[HexCode]

Kindly delete !

[Lettos]

TEMPORARY APPEARANCE...

Play System Feature Description

1) The last data section featured in scenario definition file *.PGSCN is entitled "# Units". It hosts a column entitled "Ammo". The column's entries adhere to an integer format. Each value represents the Ammo Points (APs) the corresponding unit is granted at the very start of the scenario.

2) There are no restrictions as to the magnitudes or algebraic signs of AP values. They can be negative.

3) PGF's engine interprets an empty AP cell as having value ZERO (0).

4) AP value ZERO (0) instructs PGF's engine to assign to the unit APs equal to the unit's Listed Ammo Capacity (LAC). Consequently, it's technically impossible to proactively assign actual ZERO (0) APs to a unit.

5) An AP value can be less or greater than the unit's LAC. No matter, in-game, the relevant Resupply / Replacement algorithms work like an algebraic... charm. This includes negative AP values.

6) A unit which enters a scenario with a NON-NEGATIVE AP value can NEVER achieve NEGATIVE AP status in-game.

7) A unit which enters a scenario with a NEGATIVE AP value is able to unobstructedly expend APs "on credit" in-game until its AP status turns non-negative due to sufficient Resupply / Replacement. Once that happens, points (5) and (6) above apply until the end of the scenario.

Does the Feature Exhibit Differentiation ?

Yes, it does.

Is the Feature Relevant to Custom Content Design ?

On the face of it, a unit which practically never runs out of ammo could be put to some good use by an "adventurous" content designer. Nevertheless, such a content designer could achieve essentially the same effect by grossly oversupplying the unit...

What happens if LAC is ZERO (0) though ?

In case if AMMO in pgscn is set to zero, AMMO will be set by PGF engine to unit AMMO as in eqp.

If you set AMMO=0 in eqp, that zero will still be zero in PGF. In fact, this is the same as setting SA, HA, NA, AA = 0.
If you set AMMO= negative in eqp, you get unlimited AMMO. With each attack and/or Re-supply and/or Replacements the value will become more and more negative. Enemy Level bombing will not affect negative AMMO.

Yeah, also on the subject of negatives. If Level bomber has negative parameter Bomber Special, then Suppressed units will be positive (I don't know what it means for PGF engine), STR of attacked unit from attacks will increase, but FUEL and AMMO will increase. Each such bombardment supplies an enemy unit!
Alas, this is not practically applicable. At least I can't find in reality any analogy to what happens in PGF.

Setting the AMMO parameter in pgscn to be larger than in eqp has a logical analogy to reality. Suppose at the beginning of the scenario my ARTY is located west of the river in its territory. And a few turns ARTY fires to east from its home positions with ammo depots nearby. Why not give it AMMO=10 instead of the standard 6 at the beginning of the scenario? Re-supply does not reduce the amount of extra-AMMO. Replacements reduce AMMO to the value from the eqp file.

The negative EXP of CORE units at the start of the campaign is nothing more than a maths trick too. Non-elite replacements will immediately give a huge EXP boost and make the EXP value positive. Elite replacements will keep EXP negative.

The created universe of PGF is not limitless without programmer intervention...

[Radoye]

What happens if LAC is ZERO (0) though ?

If you set AMMO=0 in eqp, that zero will still be zero in PGF. In fact, this is the same as setting SA, HA, NA, AA = 0.

I think what was being asked here is what happens if an unit has ammo = 0 in eqp but set to some non-zero value (positive or negative) in scn?

[Lettos]

What happens if LAC is ZERO (0) though ?

If you set AMMO=0 in eqp, that zero will still be zero in PGF. In fact, this is the same as setting SA, HA, NA, AA = 0.

I think what was being asked here is what happens if an unit has ammo = 0 in eqp but set to some non-zero value (positive or negative) in scn?

That's a good point!

If a unit has AMMO=0 in eqp but is assigned some AMMO value in pgscn, the unit can use the assigned AMMO value.
Re-supply will increase the unit's FUEL but leave the AMMO unchanged. Replacements will reset the AMMO value to zero.

In general, the situation with minimum AMMO (and sometimes FUEL) values in eqp and increased AMMO(FUEL) values in pgscn has practical application not only in the ARTY case I described above.

Some examples:

- Rocket. In eqp AMMO=0 and FUEL=1. In pgscn AMMO=1, FUEL is at the option of the scenario designer, depending on the MVT of the missile. Disposable unit. Unfortunately, nobody will prevent this missile from flying back to its airfield and standing there until the end of the scenario.

- Partisans, saboteurs, paratroopers. In eqp AMMO=1 or even better if = 0. At the beginning of the scenario AMMO=4-6 is assigned. These specific units have nowhere to take AMMO other than the stocks they have at the start of the action. Oh, and Replacements for these units were often not actually possible. If a player makes Replacements, he gets a non-capable unit with zero AMMO. Such units are best not to be put in the hands of AI!

- Fortification on an island or at a remote location on the shore. Shell logistics are not possible. You have to rely only on stockpiles. Replacements are not possible. You can allow AIs to use these units.

[HexCode]

TEMPORARY APPEARANCE...

Intended Audience: # Lettos # & # Radoye #

The last data section featured in scenario definition file *.PGSCN is entitled "# Units". It hosts a column entitled "Ammo". The column's entries adhere to an integer format. Each value represents the APs the corresponding unit is granted at the very start of the scenario (SAPs).

File EQUIPMENT.PGEQP contains a column entitled "Max Ammo". The column's entries adhere to an integer format. Each value represents the corresponding unit's Listed Ammo Capacity (LAC).

For the purposes of THIS post, it's assumed that BOTH SAP and LAC values are POSITIVE.

Relevant Play System Feature: Basic Description

1) There're no obvious restrictions as to the magnitudes of SAP / LAC values.

2) PGF's engine interprets an empty SAP cell as having value ZERO (0).

3) SAP value ZERO (0) instructs PGF's engine to assign to the unit SAPs equal to the unit's LAC.

4) A SAP value can be less (Undercapacitation) or greater (Overcapacitation) than the unit's LAC. No matter, in-game, the relevant Resupply Only algorithms work like a charm.

[HexCode]

TEMPORARY APPEARANCE...

Intended Audience: # Lettos # & # Radoye #

On occasion, content designers have specified Starting Ammo Point (SAP) values greater than a unit's Listed Ammo Capacity (LAC). That's Overcapacitation.

For the purposes of THIS post, it's still assumed that BOTH SAP and LAC values are POSITIVE.

Relevant Play System Feature: Special Properties

1) In all instances, Replacements Procurement automatically wipes out all APs in excess of a unit's LAC.

2) In all instances, in-game unit Upgrades (Auto-Upgrades as well) automatically wipe out all APs in excess of a unit's LAC.

The algorithms governing Resupply via Replacements Procurement or Upgrades are not Resupply Only ones. They do not conform to strict algebraic logic.

Interestingly enough, loss of APs due to Carpet Bombing does obey strict algebraic logic. The underlying algorithm is a Resupply Only one, in reverse !

Is the Feature Relevant to Custom Content Design ?

This is a very interesting question. On the face of it, Ammo Overcapacitation does confer a fighting advantage to a unit. For how long, though ? Replacements Procurement automatically nullifies the advantage. Plausible situations follow:

A) Units less likely to incur serious Strength Factor (SF) losses for awhile due to their relative distance from the thick of the battle (e.g., units possessing ranged attack capabilities).

B) Units less likely to incur serious SF losses for awhile due to their superior attack and defense capabilities (e.g., super-armor, super-fighters).

C) Units assigned missions requiring them to keep on fighting without respite no matter what. Such missions could be suicidal or, in the case of sufficiently Overstrengthened units, relatively unstoppable walkovers.

[Lettos]

Air Defense or Anti-Aircraft unit fires at Aircraft, that is, the Air Attack(AA) parameter of a ground unit versus the Ground Defense(GD) parameter of an air unit.

Tests show that there is some strange rule in the PGF mechanics of calculating attack results, which I will try to formulate as follows:

If the GD minus AA parameter is greater than "6", the air unit's losses will still occur and will be in the 3-7% range.

For example, in the Air Defense tests, a unit has AA = 14, and the air unit's GD parameter = greater than 20. It does not matter if the parameter is "21" or "41". The loss of 10 units from 10 attacks will be in the statistical range of 3-7% (i.e. 3-7 STR from 100 STR).

By changing the GD parameter, you can see the statistics (% loss from 100 STR).
If AA=14 and:
GD=14 ----- 40-50%
GD=15 ----- 30-40%
GD=16-18 - 15-30%
GD=20 ------8-14%
GD=21 ----- 1-8% (more often 3-7%)
GD=41 ----- 1-8% (more often 3-7%)
GD=81 ----- 1-8% (more often 3-7%)

That is, there is no GD parameter at which the Air Unit has no losses at all!
If you fire an Erlikon (yes even a hunting rifle!) at a spaceship in the PGF model, the spaceship will have a loss of 3-7%.

Logically there is some reasonable explanation for this. For example, airplanes suffer losses from technical incidents both in the air and during takeoff and landing. Of course, we face the convention that if a unit is not attacked, it has no technical losses if it flies even a gazillion hexes.

In vanilla eqp Flak 8.8 has AA=10 and B-29 Superfortress GD=21. Clearly, only if Flak 8.8 has EXP=500 (5 stars) will Flak have an increased chance of inflicting more losses instead of statistical 3-7%. Clearly, having at least one EXP star on a B-29 would only make it vulnerable again by a statistical 3-7%.

What does all this "statistical crap" do for practice?

1) In the form of the AA parameter of the AD/AA unit, we have in hidden form the fire altitude.
Only it should be taken into account that Level Bomber's GD should depend on bombing altitude more than it does now.

A situation where M16 MGMC (Four 12.7mm M2HB heavy machine guns), or SdKfz 7/1 (Four 20mm) have AA=11-12, and Flak 8.8 has AA=10 is illogical.

I think we have here the result of successive strange actions in the creation of parameters. First they gave Tac Bombers too high GD parameter for unknown reasons. Then they thought about it and decided - we have to shoot them down somehow! And gave a high AA value for small anti-aircraft artillery. Another reason - someone really wanted Wehrmacht tanks, protected from air attacks only by small-caliber Anti-Aircrafts, to reach Antwerp in January 1945, despite B-26, A-26 and other dangerous Allied aircraft.
As a result, it became less effective to shoot at Superfortress with Flak 8.8 EXP=500 than to use SdKfz 7/1 or Wirbelwind.

Reduce TacBomber GD -> Reduce AA of small-caliber AA and AD.

Then how do you bomb AD 76-85-88mm with TacBombers? Mainly you don't. TacBombers didn't do such suicidal nonsense nowhere in reality except semi-real Panzer General and its clones. I find it easier to accept the assumption that a battery of Flak 8.8's is also protected from assault attacks by smaller caliber anti-aircraft guns (sort of like an integrated battery) than to see a B-29 hit by 20mm guns.

And after all, the situation was created artificially, and it can be corrected! More often use Level Bombers and Long-Range ARTY to suppress anti-aircraft guns.

2) Creating air aces

Player Aces

The player's aces sort of create themselves as units gain EXP. This increases INI, attack-defense parameters. Fighters gradually turn into quite effective bombers by SA/HA values. It seems that the player doesn't really need aces....

... although the player can be offered the alternative of buying a unit that will be an ace. It will not be a fighter-attacker (SA=0-1, HA=0, NA=0) but will have an additional bonus +1 +2 to AA and INI. And increased GD value (may be even up to 18-20). If the fighter does not fly assault missions, it is engaged in its primary business of fighting enemy aircraft. And he very rarely comes under AD/AA fire.

That is, it is the same model Air Unit, only its parameters have been changed in accordance with the combat missions performed. The Aces didn't like assault missions. If such air hunters hung a bomb or two from the airplane, the pilots tried to drop this bomb as quickly as possible on the first target they saw. The pilots were not interested in the effectiveness of this strike. They realized that the presence of a bomb would reduce the chances of winning the expected air battle.

AI Aces

This is one that likes to spread his best fighters over AD or near AD ace will come in handy!
Really, it doesn't surprise you that in some scenarios most of the AI fighters you destroyed with your AD/AA?

Take away from them any ability to attack ground targets (SA/HA/NA=0).
Give them - at least some of them - GD=21!
Give them so much EXP that they equal or exceed some of your Fighters in INI.
And another trick already mentioned - put a couple of them in TacBomber Class. That way they won't be distracted even when defending their Bombers. Aces are free hunters, not guards. They guard Bombers indirectly by destroying enemy fighters.

How to implement Aces in unit icons? Taking into account that there may not be so many models of fighter planes flown by aces, it is possible to make such icons:

[HexCode]

Intended Audience: # Lettos # & # Radoye #

PGF's programmer wrote:

... a weak unit always has a small chance of inflicting some damage to a much stronger opponent as well as a small chance of escaping total destruction from the opponent's (counter)attack.

[Lettos]

PGF's programmer wrote:

... a weak unit always has a small chance of inflicting some damage to a much stronger opponent as well as a small chance of escaping total destruction from the opponent's (counter)attack.

But not if one unit shoots another from a considerable distance! I mean, there are fields of knowledge like physics, geometry, ballistics....

[Lettos]

PGF's programmer wrote:

... a weak unit always has a small chance of inflicting some damage to a much stronger opponent as well as a small chance of escaping total destruction from the opponent's (counter)attack.

But not if one unit shoots another from a considerable distance! I mean, there are fields of knowledge like physics, geometry, ballistics....

Let's see how the spiritual founding fathers of the SSI-Panzer General sect looked at the problem.

Vanilla Panzer General:
Flak 8.8 AA = 10
Wirbelwind AA=11
B-29 Ground Defense = 21

Version 1.2 fixed the problem:

Vanilla Panzer General 1.2:
Flak 8.8 AA = 15
Wirbelwind AA = 11
B-29 Ground Defense = 21

Then there were heretical movements against the classic canons.
And Flak 8.8 AA = 10 again.
Good thing it didn't come to the Inquisition and bonfires!

[Radoye]

PacGen had this problem solved by removing the possibility for AA class to engage LB class units. There was also a unit special ability which when selected enabled the unit to provide air defense support fire, mainly to be used on ships which were to provide AA protection, but worked elsewhere too. So a solution was found in the modding community by moving the light AD guns to AA class and adding them the "ADA Support" special so that they provide anti air cover. In the game they still worked as usual, but were unable to hit the B-29s and other Lancasters and Flying Fortresses

[Lettos]

PacGen had this problem solved by removing the possibility for AA class to engage LB class units. There was also a unit special ability which when selected enabled the unit to provide air defense support fire, mainly to be used on ships which were to provide AA protection, but worked elsewhere too. So a solution was found in the modding community by moving the light AD guns to AA class and adding them the "ADA Support" special so that they provide anti air cover. In the game they still worked as usual, but were unable to hit the B-29s and other Lancasters and Flying Fortresses

In our universe, we too can - and already do - transfer AD Units of 20-37mm caliber to Class AA. But we don't have any other options besides adjusting Air Attack and GD.
Fortunately, the range of adjustments is very wide!

A unit with Air Attack = 1 will do exactly the same damage (statistically) to an aircraft with GD = 6 as an AD unit with Air Attack 16 will do to an aircraft with GD = 21.
Here, for each dual/combined machine gun and pom-pom, you can pick the appropriate AA parameter from "1" to "14" for each dual/combined machine gun and pom-pom. And for all Air Class, the GD range from "3-4" to "21" is enough for fine adjustment.

[Radoye]

Of course.

We can even go higher than 21 GD for the LBs. There is plenty of wiggle room.

[Lettos]

Of course.

We can even go higher than 21 GD for the LBs. There is plenty of wiggle room.

Yes, of course!

But this scale should also have its own logic for the AA-GD relationship between the three classes of Air Units.
It's not that simple... you can't make some AD units a little dangerous for Level Bombers, and at the same time absolutely killer for Tac Bombers.
It shouldn't be a table divided into three AA-GD areas. The areas should overlap each other.

I'm off to test something along those lines. By the way, I have a suspicion that all of this has already been done in SSI. And I'm unlikely to invent something new. Unless PGF has some other, different loss calculation mechanism than Panzer General built in.
Most likely, I'll repeat the old results. Only now I will do it not just by intuition, but understanding why and what I am doing.

[Radoye]

This is a problem, because large caliber AA that were capable of hitting the high flying strategic bombers were usually useless against low flying fighters and ground attack planes - unable to traverse fast enough to track them efficiently.

If in PGF we would have the same capabilities as in PacGen and AA class was prevented from engaging LB class, we could then (somewhat counter-intuitively) actually reduce the GD for LBs and AA for AD units while increasing drastically the GD for FTR / TB and AA for AA class.

If AA class is by default prevented from attacking LBs, it can have air attack = 255 for all that matters, it won't be able to do any damage to LBs even if their GD is 0. Then an AD class (like the 88) could have AA=1 and still damage the LBs, but not really TBs and FTR that have GD = 255. Where an AA with AA=255 would be able to damage the same FTRs and TBs, but class rules would prevent it from touching LBs...

(i have used exaggerated numbers in my example above, of course, but you get the idea!)

[Lettos]

This is a problem, because large caliber AA that were capable of hitting the high flying strategic bombers were usually useless against low flying fighters and ground attack planes - unable to traverse fast enough to track them efficiently.

I'm well aware of that.
And at the same time 70 Tac Bombers Fairey Battle, sent to bomb the bridge over the Meuse, were shot down by Flak 8.8 on the same day in May 1940.
But we don't have the ability to make a layered unit that simulates multiple AD batteries in a single hex. Unless we make a DUAL unit for the AI from two different ADs, but that doesn't solve the problem, it just creates new ones.
We can't emulate the faustpatrons, flamethrowers, bazookas, mortars, small caliber guns that the infantry had. We just give the infantry some weird HA, and let them fight the tanks as best they can.

That's why I wrote:

Lettos: I find it easier to accept the assumption that a battery of Flak 8.8's is also protected from assault attacks by smaller caliber anti-aircraft guns (sort of like an integrated battery) than to see a B-29 hit by 20mm guns.

In reality, the air defense was a very dense layer cake of AD guns on one hex... several guns of different caliber could stand tens of meters away from each other. And something has to be done about that. Assumptions are inevitable, even if we're talking company and battalion scale. We're limited by the hex.

[Lettos]

Here's the full picture of what happens in the case of a shooting.

Air Attack minus GD gives the following Loses % approximate statistical result (previous results are also included):

19-15 ---83-85%
13-12 -------70%(72%)
10-7 --------65%-69%
6 ------------50%
5-4 ------------?
3 ------------47%(50%)
2-1 ---------?
0 ------- 40-50%
-1 ------ 30-40%
-2 -4 -- 15-30%
-6 -------8-14%
-7 and more minuses ----- 1-8% (more often 3-7%)

We need a reference point in the form of at least some sort of Air Attack scale to get started work with it.

Suppose the most primitive scale could look like this:
Type-----------------AA
7.7mm --------------1
12.7mm-------------2
2x7.7mm------------3
2x12.7 or 4x7.7----4
1x20mm-------------5
4x12.7---------------6
2x20 or 1x37-------7
4x20-----------------9
2x37-----------------10
4x37-----------------11
76--------------------12
88--------------------14
105mm and larger --15-16 and more

I warned you that I would repeat the SSI discoveries... Wirbelwind 4x37 AA=11, Flak 8.8 AA=15.
Yes, that's exactly what we've been playing with for 30 years!

And now we need to pin three air classes to these two tables. Do it in such a way that the logic is not affected and that the players are satisfied. I assume that in principle there is no perfect solution that satisfies everyone. But it is also clear that these tables are not a reason for religious wars

[Lettos]

This is a problem, because large caliber AA that were capable of hitting the high flying strategic bombers were usually useless against low flying fighters and ground attack planes - unable to traverse fast enough to track them efficiently.

I find it easier to accept the assumption that a battery of Flak 8.8's is also protected from assault attacks by smaller caliber anti-aircraft guns (sort of like an integrated battery) than to see a B-29 hit by 20mm guns.

Let's look, for example, at the distribution of AA-guns in the German army on Eastern front in Barbarossa on June 22, 1941.

616 8.8cm Flak 18/36 in 41 Luftwaffe mixed flak battalions and 10 Army ("heavy") motorized battalions of AA-guns.
Light anti-aircraft companies (at least 97 were created, may be more) were armed with 20-37 mm guns. Fourteen battalions were created from 46 companies. The remaining 51 companies were used as separate units attached to panzer and infantry divisions.

Army ("heavy AD guns") Battalion composition:
8-12 Flak 8.8
9-18 20mm transportable
12 20mm on trucks (this separate company was sometimes part of the Army Heavy-AD battalion)
Transport: 10-15 Sd.Kfz.7 for Flak 8.8, 56-65 Trucks and others.

Luftwaffe Bat:
12 Flak 8.8
30 Light AD guns

That is, even at battalion scale, we get a comprehensive AD unit effectively firing at different altitudes. Well, that's fine!