Vanilla AI in 9.2 beta patch now builds very close to 40-width divisions.

[Gefallener_Held]

From the gameplay side, 40 width vs 20 width is in an OK spot. 40 width is superior in more places, obviously, but 20w (and 10w) definitely have niches. 40w armor divisions paired with 20w pinning infantry is a pretty standard combo, for instance.

This is precisely the problem. And everything I have read indicates 40 width armor really cannot be stopped without also usingn40 width meta garbage which I do jot wa t to do and others do not want to do and should not have to do! The rest of your post proves the combat dynamics cannot be fixed otherwise, which is why 40mwidth divisions, particularly armor division, must be banned. I would endorse altering the width to maybe 60 so that 20 and 30 w are doable, as so someone else suggested....

 

[SchwarzKatze]

For what it's worth, your analysis is totally correct. "Fixing" the 40 width meta would require a lot of changes to many systems. Things that would need to be adjusted include:

• The combat targeting algorithm
• The cost of support companies
• The benefit of support companies, especially how some provide a % bonus (e.g. logistics) while others provide base stats (e.g. support artillery)
• Total organization
• Organization regain rate, especially the boost when at low organization
• Command limit of generals
• The benefits of granularity for smaller divisions
• Reinforce rate
• Probably at least a half-dozen more that I'm forgetting

That's... a LOT of stuff that would need to be changed for something that would have a negligible benefit. In a game like HoI4 where there's still so much that needs to be improved (e.g. peace conferences), it's hard to justify this as a pressing need.

From the gameplay side, 40 width vs 20 width is in an OK spot. 40 width is superior in more places, obviously, but 20w (and 10w) definitely have niches. 40w armor divisions paired with 20w pinning infantry is a pretty standard combo, for instance.

Not really? There are really only 2 problems:

1) Extremely harsh over combat width penalty applied to *all* divisions in the fight.
2) The way defense is calculated that makes a large division better than the sum of its parts.

1) is what limits effective division widths to factors of 80 (i.e. 40, 20, 10, 8, ...), and 2) is what make the largest of them all the best.

14/4 is a subset of the problem, with the addition of
3) line artillery takes up 3 width, limiting the combinations possible to achieve 40 width.

If 1) and 2) don't hold, then 3) isn't a problem.

 

[Harin]

I mean I haven't looked that deep into the mechanics of it and reflected on it, and my brain is fried atm so im just heading out for the day to play some Divinity.

What I can say though from just taking a peek at it, just for fun (this isn't a proper assessment or a statement so don't take it as such) is that if I am looking at it in terms from a game development perspective (deadlines, resources yaddayadda) such a change would probably require many systems to change along with it.

If you think that you are thinking that you are changing 1 singular system like you are here you would probably quite quickly into a lot of surprises like "We now have to rework air combat because the simulation doesn't work anymore" and "terrain has broken" or "some bonuses from doctrines are now stacking in weird ways and become op". Not saying that it is impossible to fix, but it is more likely easier to implement it in a whole new game built for the system than nudging an existing system for it. In the same way you can't just take HoI4s combat and slap it onto CK2 since is changes fundamentals about the game.

For what its worth I don't think that 40 width is always optimal, such as fighting in low supply terrain and its not like you instantly loose because you are using 20 Width instead of 40 Width. Although I don't think that it is particular fun to always have a "40 width or 20 width is the only choice" since it isn't an interesting choice, it isn't a choice at all. But that is my personal opinion though, and players will always find an optimal template so I guess its a cursed problem

Thanks for your reply. It helps everyone to know that you are listening, but more importantly it helps if you can let us know what is in the realm of possibilities and what you are thinking.

For example of what is feasible in the game, or what you think is probable, first consider this single player scenario.

If a player playing Germany was only allowed to use the division templates the AI uses now, but the computer controlled Allies were allowed to use 40w armored divisions and other META templates, these forums might blow up in disbelief and some rage. I suspect there would be few who would consider that game balancing, fair, historical, realistic, etc... Arguments for these powerful METAs would most likely be turned on their heads. Some would probably say the AI is cheating at that point.​

If someone believes that this scenario is ok, then they would probably say there is nothing to fix. For those who thinks this scenario would harm the game, what is in the realm of reason that could be suggested to fix it?

Personally, I see problems with:

Battalion, Division, and Battlefield width​

How divisions have to fight alone in defense, or pairs (or so) in offense, instead of a cohesive whole​

How battalions, the building blocks of divisions, are defaulted the same across every country, from the US to Tibet.​

I do understand that no matter what Paradox does, a META will be found. That cannot be avoided. I would suggest that today, the METAs are simply to powerful. The space between the AIs combat power and the players is to great.

What is in the realm of possibility to reduce the difference? One quick answer might be to let the AI start with and default to building METAs, but I do not know how much work that would be. It would at least reduce the power gap, but if the three problems I listed above are legitimate problems, it does not address them at all.

Is there any advice you can give us in how we should frame our thoughts for solutions so that is it not a waste of your time to read them?

 

[SophieX]

The problem is this META itself:

The base template has 5 x 5 = 25 battalions. This is written in the defines; but it could easily be modified.
let's say to 3x3, 3x4 or 4x4
and
My idea is to add a new define ( I couldn't do that ) that varies the front-width randomly. Now the width is 80.
With a random modifier of, lets say 20%, it varies between 64 and 96. ( only as an example )

If you want to make it more complicated, this random modifier changes all 2 or 4 days ( only as an example too )

 

[him_15]

That's great to hear that.

 

[Emren]

The problem is this META itself:

The base template has 5 x 5 = 25 battalions. This is written in the defines; but it could easily be modified.
let's say to 3x3, 3x4 or 4x4
and
My idea is to add a new define ( I couldn't do that ) that varies the front-width randomly. Now the width is 80.
With a random modifier of, lets say 20%, it varies between 64 and 96. ( only as an example )

If you want to make it more complicated, this random modifier changes all 2 or 4 days ( only as an example too )

Variable battlefield width is an old suggestion, which Paradox has not shown any interest in. Hopefully that might change.

 

[Reman]

Not really? There are really only 2 problems:

1) Extremely harsh over combat width penalty applied to *all* divisions in the fight.
2) The way defense is calculated that makes a large division better than the sum of its parts.

1) is what limits effective division widths to factors of 80 (i.e. 40, 20, 10, 8, ...), and 2) is what make the largest of them all the best.

14/4 is a subset of the problem, with the addition of
3) line artillery takes up 3 width, limiting the combinations possible to achieve 40 width.

If 1) and 2) don't hold, then 3) isn't a problem.

By "the way defense is calculated", I'm assuming you mean how concentrated attacks can overwhelm a division's breakthrough/defense and deal 4x damage, and how each division targets a random enemy division in each hour of combat? That's what I was referring to with "the combat targeting algorithm". It's the biggest factor leading to 40w dominating, but it's far from the only factor that would need to be changed to make combat width irrelevant in division design.

For instance, assume #1 and #2 in your post are fixed. What's to stop people from just spamming tiny 10w divisions then? Organization is calculated as an average of the division's battalions, so more divisions = more org and more org regain. Now small divisions dominate since they let you double your org for free.

 

[SchwarzKatze]

By "the way defense is calculated", I'm assuming you mean how concentrated attacks can overwhelm a division's breakthrough/defense and deal 4x damage, and how each division targets a random enemy division in each hour of combat? That's what I was referring to with "the combat targeting algorithm". It's the biggest factor leading to 40w dominating, but it's far from the only factor that would need to be changed to make combat width irrelevant in division design.

For instance, assume #1 and #2 in your post are fixed. What's to stop people from just spamming tiny 10w divisions then? Organization is calculated as an average of the division's battalions, so more divisions = more org and more org regain. Now small divisions dominate since they let you double your org for free.

Isn't that the same issue? That a division's combat performance is not equal to the sum of its parts.

And what was the last time the organization issue was tested? Because if that's true, then in the case that neither side has more attack than any divisions' defense value, then smaller divisions are several times as resilient as the bigger ones? That sounds like a serious bug.

I'm asking because I was in the dispersed versus concentrated industry thread, and it turned out that the entire first page of the discussion was based on a faulty premise (18 months break-even)

 

[Reman]

Isn't that the same issue? That a division's combat performance is not equal to the sum of its parts.

And what was the last time the organization issue was tested? Because if that's true, then in the case that neither side has more attack than any divisions' defense value, then smaller divisions are several times as resilient as the bigger ones? That sounds like a serious bug.

I'm asking because I was in the dispersed versus concentrated industry thread, and it turned out that the entire first page of the discussion was based on a faulty premise (18 months break-even)

My point is that making it so a division's combat performance is the same no matter what size it is (e.g. making 2 20w divisions equivalent to 1 40w division in all respects) would be a difficult task. Either the devs would have to do a full rework of how land combat works, or they'd have to try to bandage-over dozens of ways the current system interacts with width and division size. Solving any single one of these issues wouldn't be that much work, but doing them all would be a significant investment of the devs' time. Heck, even *finding* all the obscure places width and division size make a nontrivial difference would be challenging. And in the end, what would be gained from all this effort? A slightly more historical experience? The game has much more pressing issues that should be solved first.

The organization issue can be trivially tested in the division designer in game, or here. A 20w pure infantry division has 60 org, as does a 10w, as does a 40w. Using 2 10w divisions in place of 1 20w therefore nets 120 organization instead of 60, letting the smaller divisions hold out far longer ceteris paribus.

 

[noobermenschen]

It is not just ahistorical, but highly exploitative precisely because one 40 width is qualitatively superior to two 20 width, etc, for no rational reason.

Exploits should be everyone's concern, especially when the direction is to force players to use them. I do not like the experience exploit people were using, or surrounding a capital to cut supply off, space marines etc.

I am sorry to see this irritate your gameplay or anyone else's, but I don't think AI 40 width ruins the game for most people. It does help reduce division spam (somewhat, I noticed my Columbian ally has 96 smol Infantry brigades in the field), and gives the player a challenge from a stronger and more efficient AI. I would rather the majors not put these in the field as it would then make 20 width completely ineffective, and the Japanese 10/2 with AT and German 9/3 with AA are sound templates that are tough to beat in large numbers.

 

[blahmaster6k]

Isn't that the same issue? That a division's combat performance is not equal to the sum of its parts.

And what was the last time the organization issue was tested? Because if that's true, then in the case that neither side has more attack than any divisions' defense value, then smaller divisions are several times as resilient as the bigger ones? That sounds like a serious bug.

I'm asking because I was in the dispersed versus concentrated industry thread, and it turned out that the entire first page of the discussion was based on a faulty premise (18 months break-even)

This is pretty much the reason 10 width divisions are banned in many multiplayer servers. 10 width divisions allow you to have so many divisions on the front line and in reserve that if you tech into enough reinforce rate and org recovery, you can infinitely stall an attack by cycling your divisions back into the battle. No amount of 40 width tanks can break through a province if you have fifteen 10w divisions cycling in the battle. if you're attacking with 2 40w tank divisions on a one province front, you make 2 divisions retreat, and there are still 6 in the battle. Two more reinforce. You make the next two retreat. two more reinforce. You make the next two retreat. two more reinforce. You make the next two retreat. The first two divisions you caused to retreat have now fully regained their organization and are ready to reinforce. etc etc. Most multiplayer games are played on speed two during war since players need to micromanage so much without being able to pause (with players being allowed to request speed one if they're trying to hold a defense together), and at slow speeds manually cycling divisions like this even in several places at once is not difficult.

Mass spam of small divisions also causes the game to lag, but that's another issue entirely.

 

[SchwarzKatze]

My point is that making it so a division's combat performance is the same no matter what size it is (e.g. making 2 20w divisions equivalent to 1 40w division in all respects) would be a difficult task. Either the devs would have to do a full rework of how land combat works, or they'd have to try to bandage-over dozens of ways the current system interacts with width and division size. Solving any single one of these issues wouldn't be that much work, but doing them all would be a significant investment of the devs' time. Heck, even *finding* all the obscure places width and division size make a nontrivial difference would be challenging. And in the end, what would be gained from all this effort? A slightly more historical experience? The game has much more pressing issues that should be solved first.

So you think that "all templates but 10 Inf / 7 Inf 2 Art / 14 Inf 4 Art are heavily penalized" is a non-important issue?

And this is already a bandage if what you said is true: Two silly problems, one where smaller divisions takes disproportionately more damage, and one where smaller divisions has more org in total, partially cancelling each other out. Sounds like a major problem that needs to be fixed.

How many issues are more pressing than this? Actually I can name a few: AI going full Luigi Cardona and gets million of troops killed in a front-wide losing offensive. AI guarding neutral borders while active fronts are melting down. Paradropping randomly not executing. Merging supply zones drops the total supply. Just to name a few that are important but still aren't addressed.

Did they stop Paradox from adding anything new and fixing less pressing issues? No, so that's really a non-argument.

Personally, I think this is the single biggest outstanding issue with "land warfare mechanics", now that we have fuel in the game.

The organization issue can be trivially tested in the division designer in game, or here. A 20w pure infantry division has 60 org, as does a 10w, as does a 40w. Using 2 10w divisions in place of 1 20w therefore nets 120 organization instead of 60, letting the smaller divisions hold out far longer ceteris paribus.

That's not verification. Verification is to actually run the game and record the actual behavior. It was how the 40 width meta was discovered.
And even if verified, it just means that the current combat system is barely holding by two nonsensical oddities counteracting each other.

This is pretty much the reason 10 width divisions are banned in many multiplayer servers. 10 width divisions allow you to have so many divisions on the front line and in reserve that if you tech into enough reinforce rate and org recovery, you can infinitely stall an attack by cycling your divisions back into the battle. No amount of 40 width tanks can break through a province if you have fifteen 10w divisions cycling in the battle. if you're attacking with 2 40w tank divisions on a one province front, you make 2 divisions retreat, and there are still 6 in the battle. Two more reinforce. You make the next two retreat. two more reinforce. You make the next two retreat. two more reinforce. You make the next two retreat. The first two divisions you caused to retreat have now fully regained their organization and are ready to reinforce. etc etc. Most multiplayer games are played on speed two during war since players need to micromanage so much without being able to pause (with players being allowed to request speed one if they're trying to hold a defense together), and at slow speeds manually cycling divisions like this even in several places at once is not difficult.

Mass spam of small divisions also causes the game to lag, but that's another issue entirely.

Alright, so now it's definitely a serious issue that needs to be addressed. I thought the 10 width one is mainly based on reinforcement chance instead of broken organization.

 

[Reman]

So you think that "all templates but 10 Inf / 7 Inf 2 Art / 14 Inf 4 Art are heavily penalized" is a non-important issue?

And this is already a bandage if what you said is true: Two silly problems, one where smaller divisions takes disproportionately more damage, and one where smaller divisions has more org in total, partially cancelling each other out. Sounds like a major problem that needs to be fixed.

How many issues are more pressing than this? Actually I can name a few: AI going full Luigi Cardona and gets million of troops killed in a front-wide losing offensive. AI guarding neutral borders while active fronts are melting down. Paradropping randomly not executing. Merging supply zones drops the total supply. Just to name a few that are important but still aren't addressed.

Did they stop Paradox from adding anything new and fixing less pressing issues? No, so that's really a non-argument.

Personally, I think this is the single biggest outstanding issue with "land warfare mechanics", now that we have fuel in the game.

Not sure what you mean by "all templates but 10 Inf / 7 Inf 2 Art / 14 Inf 4 Art are heavily penalized", because there are lots of templates beyond those that are both useful and not penalized by the system as it currently stands. The system can force some slightly awkward setups, e.g. if you want to mix a bit of SPAA into your tanks you'll probably need to use 2 battalions per division instead of 1 since SPAA only has 1 combat width. In practice, however, this type of thing is only ever a very mild problem.

The fact that smaller divisions have some benefits that partially cancel out their drawbacks is what makes the current system tolerable. It goes far beyond the targeting algorithm and org differences, with stuff like reinforce checks and support company concentration/efficiency also being factors. All of this makes choosing between 40w and 20w (or 10w) a nontrivial exercise. Some people believe that 40w divs are best in every situation, but that's simply not the case. You contend that this is all indicative of a "major problem", but I really don't see why. The combat width system is arbitrary, but functional and relatively balanced. This puts it ahead of a lot of mechanics in the game, like peace conferences, which are arbitrary, nonfunctional, and unbalanced.

A problem's severity absolutely *should* be a factor in how highly it gets prioritized to be fixed. PDS definitely doesn't do this in many cases, but that should only be held as a mark against how they prioritize things. Small problems should be fixed if they only require a small amount of dev time, while big problems can be justified to be fixed even if they require a large amount of dev time. But spending lots of dev time on small problems, like combat width, simply doesn't make sense when there are much more pressing concerns.

 

[Gefallener_Held]

I am sorry to see this irritate your gameplay or anyone else's, but I don't think AI 40 width ruins the game for most people. It does help reduce division spam (somewhat, I noticed my Columbian ally has 96 smol Infantry brigades in the field), and gives the player a challenge from a stronger and more efficient AI. I would rather the majors not put these in the field as it would then make 20 width completely ineffective, and the Japanese 10/2 with AT and German 9/3 with AA are sound templates that are tough to beat in large numbers.

None of these considerations you outline (reducing division spam, giving the player an unnatural challenge) are excuses for this. It is not a challenge more than it is obligating people to use 40 width divisions. The game should be made more challenging by improving the AI.

 

[Ketchup & friends]

So why can't artillery be 2 or 4 width
3 is a bad number to make 20 or 40 since it obliges you to pair them

We have previously proposed changing the width of all artillery units to 1 and reduce all combat stats to 1/3 of the current values.
Making artillery 2 width could work as well.
But width 4 would allow creating extremely wide divisions by adding lots of artillery. That would be extremely gamey.

With a random modifier of, lets say 20%, it varies between 64 and 96. ( only as an example )

If you want to make it more complicated, this random modifier changes all 2 or 4 days ( only as an example too )

Making this completely random would be a bad idea. It would make it impossible to design divisions according to strategic plans.
But if the permanent reduction or increase was based on Terrain and a temporary on Weather it could work.



One thing rarely discussed in threads like this is the number of support company slots. Historical divisions often had more than 5 such units.
IRL they also used more manpower than in the game.

We think this needs to be rebalanced somehow (but we couldn't agree how exactly this should be done).

 

[Zauberelefant]

Well, there is a downside, if the AI builds 40w I guess historical divisions will no longer be viable

Only if you think that 27 width is actually historical. There are good reasons it's not.

 

[SchwarzKatze]

Only if you think that 27 width is actually historical. There are good reasons it's not.

How so?

 

[SophieX]

With a random modifier of, lets say 20%, it varies between 64 and 96. ( only as an example )

If you want to make it more complicated, this random modifier changes all 2 or 4 days ( only as an example too )

Making this completely random would be a bad idea. It would make it impossible to design divisions according to strategic plans.

In my humble opinion, in this game there aren't any strategic-plans, only tactical ones.
In those tactical plans the AI only count the number of frontline provinces, and not the quality and quantity of enemy deployment. For me as a player; who never used "battle-plans" such a random factor would reflect on the frictions, which every commander has to deal with.

But if the permanent reduction or increase was based on Terrain and a temporary on Weather it could work.

Permanent factor for terrain and weather as the random factor is quite a very good idea! ( I am sure I cannot modify the 80w front-width; but I will take a look whether I can modify the division combat width regarding terrain and weather )

 

[Zauberelefant]

How so?

There was a lengthy and ultimately toxic debate about this, but the gist is that a standard western ww2 division had 4 artillery Battalions or equivalent, at least 1 AT battalion and 1 AA battalion, giving a minimum 32 width. If you added corps or army assets like Independent tank brigades/battalions, heavy artillery, additional pioneers, AT, AA, TD, you end up close to 40w.

 

[Jafkka]

There was a lengthy and ultimately toxic debate about this, but the gist is that a standard western ww2 division had 4 artillery Battalions or equivalent, at least 1 AT battalion and 1 AA battalion, giving a minimum 32 width. If you added corps or army assets like Independent tank brigades/battalions, heavy artillery, additional pioneers, AT, AA, TD, you end up close to 40w.

Totally reasonable. And the game's divisional manpower is still lower than a historical counterpart because a lot of support staff - cooks, drivers, veterinarians, etc, are simply not modeled. Veterinary staff alone can be close to 400-600 people in a ww2 non-cavalry division.