HoI4 DevDiary - Tank Designer

[Jutland]

Oh boi, I can't wait to see production/reliability bonuses for t-34 modules heh

 

[podcat]

Regarding the criticism that you guys have been getting lately, I know how frustrating it can be to get ridiculously high criticism for something you've put way too much work on, I've been there. I really appreciate what you guys have been done for this game over the last few years. Just wanna say that right off the bat.

To talk about specifically what happened last week, some of the criticism was rightfully there IMO. There were some things one could consider "flaws" in that dev diary, and it's our right to talk about those things as the community. But some of us were way too harsh on you guys and you didn't deserve it. I just hope we can get past that and look forward to what's coming.

Keep up the good work, keep throwing those content on us!

We shouldnt really discuss it in this diary, but just to be clear. We invite feedback and criticism on our content (and there was good points brought up!). We dont like it when it targets developers. If you dont like a thing you can say that "I dont like X, it doesnt seem accurate!" vs "You are an idiot for putting X in the game" which... doesnt make us want to listen, and makes devs feel like its a hostile environment.

 

[Juha689]

This does look interesting but I'm still hoping that there is an Finnish focus tree to be announced.

 

[Svenkomigen]

I like although I would like two things added:

-armored cars created the same way,

- Flamethrower tanks

They did say that light tanks could be given wheels or half-track designs, so I assume that's how we'll get armored cars now?

 

[Archangel85]

Amazing work, I can't wait to play this.
New mechanics being introduced to the game makes me very happy, but also a bit concerned. How will the AI handle all of this? You said it yourselves that you are changing really core mechanics of Land Warfare, if you don't spend some time making the AI use it well enough, SP games are going to become even easier. Years after the release of other DLCs, the AI still doesn't use its Intel Agency and navies well enough. I believe that, by gut feeling, at least 70% of the community is with me on this, can you guys at least tell us, from a developer perspective, how you see the current state of the AI? Will you guys take some time to make the AI good at managing things better? But so far, I loved this new designer!

We are well aware that we need to get the AI to play well with the changes we are making. I think part of the issue here is that there are different expectations of what "good AI" looks like and how it feels to play against it. I would question that 70% of the community thinks the AI needs to be better, considering that almost half of the community plays on easy or very easy difficulties. What we do see is that a lot of people are disappointed when they see the AI make obvious mistakes, and we want to fix that as much as possible. Some of it is "the AI didn't do the thing I would have done and is therefore making a mistake", which is hard to work around when you have thousands of people having ideas about what the correct approach to a situation is. We do hope to show you some improvements to the AI later in development, when the features are more settled down, but I'm not going to use a loaded term like "we fixed the AI", because we probably disagree on what that means exactly.

I don't think it is very viable or even desirable to have an AI that plays optimally, or the current meta at all times. For the tank designer in particular, my goals are in order of priority:
1. The AI builds competitive tanks, meaning it researches tech at a good pace and puts modern tanks in the field
2. There is a degree of historical flavor to the AI's tank design and building approach, particularly if historical focuses are on
3. The AI adapts to circumstances to a degree (don't design heavy tanks if you have no industry base, for example)

There are more things I would like to do, like making the AI adjust its build strategy based on what the player does (if player builds heavily armored tanks, prioritize piercing), but while it is fairly easy to make an AI that appears competent in 90% of cases, it is really hard to make an AI that appears competent in 95% of cases. Humans are really good at eyeballing situations when making decisions, and AIs are really bad at it.

Will you be improving the way the AI interacts with naval templates as well?

While working on the tank designer AI I did spot a few issues that allowed me to go back and adjust some AI behavior in the ship designs in particular. We also fixed some underlying code bugs. I can't tell you if that will end up feeling like a massive improvement.

It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

This isn't good game design as it basically adds complexity without depth. This is a major issue that the HOI4 designers struggle with (a recent example was the addition of different types of recon).
Especially taking into account what happened with the MTG ship designer. But I understand the need to add some shiny features to put on the DLC product page or 3D models (such as the armored car in previous DLC).

A module system is only interesting if:
A) several significantly different templates for the same unit are viable. Otherwise —and this is what happened with MTG— you're simply adding extra clicks to achieve what could be already done before. In MTG, there is for example only 1 optimal detection ship template, so why force us to click on all the modules? Same for other ships.
B) the choices you make on modules actually have an impact. Making people choose between Armor A, Armor B or Armor C is useless if difference between those modules is negligible in the final fight calculations.

Additionally, no true effort was made to balance modules MTG in subsequent patches. Balance in a module system is essential, otherwise there are optimal templates which defeat the very goal of having a module system.

Question to the devs:
Currently, we already have many different tanks, however people overwhelmingly use the same ones (medium tanks).
What makes you think that adding a template designer is beneficial to the gameplay?
What's going to happen is that people will build the same optimal design repeatedly, with extra clicks now. It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

When I set out to design this feature, this is exactly the scenario I was worried about, although I would argue that there being a design optimized for spotting and another optimized for combat is kinda the point of having different options in a designer. I have no doubt that people will come up with "optimal" designs for a given situation, but that is part of the point of making a strategy game. If it keeps people playing for another 500 hours until they have solved the optimization problem, I think the system is doing its job.

I also think that tanks benefit a bit more from wanting to optimize more for different values. A tank that is good for Japan in China is not necessarily a tank that is good for Germany in Russia. A tank that works against the Russians might have to spend too much on armor and hard attack to be optimal in another theather where it only faces infantry, but a tank optimized for combat against infantry suffers if it runs into enemy tanks. I made an effort to give every weapon a specific niche where it is best at ("Highest soft attack for a light vehicle" or "best hard attack per IC spent"), because I want people to think about what they need the vehicle to do and where they are willing to compromise.

 

[DeadHeat16]

Planes when?

Also

Look at the tank the wrong way and it’ll fall apart XD

 

[bitmode]

I know I'll probably be in the minority on this, but I'm really not sure this was necessary and it'll just end up being more tedious clicks.

Isn't that the point of DLC though? Necessary things go in the base game, so DLC will necessarily have unnecessary things in it.

 

[boost_1214]

Plz increase research slot!!!!!!!!!!!!!!!!!!!!

 

[FlakPak]

Can you somehow make it possible to disable the designer for ships? It's just too much work (in my opinion) and I always forget to research new modules.

 

[Voigt]

Also the best way to keep a friendly discussion but also drill the person a bit, is simply asking questions.

Not you are stupid that you have included branch x, but not branch y, and using that name is wrong, but instead
I feel this person x come somehow out from nowhere, and I am surprised that person y wasn't included. How come that you designed it this way? Personally I really would like person y included, would there still be time to make changes for that?

If enough people post the same critique, and optimally in a friendly way, and it isn't a complete month like rework, more often than not Paradox listens and does changes. Not always, but often.

I mean the devs said, that there was valid criticism in the last dev diary, just that it was drowned in other aggressive postings. Maybe which also had the same core critique, but still making it less likely overall to read and act on thr feedback.

 

[Tdizzl3d]

Yo this looks sick! Are there any plans for more granular 3d model support than we've got already, or is the 3d model selection shown here just choosing from the models already available?

 

[LL0rd]

It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

This is questionable game design as it basically adds complexity without depth. This is a major issue that the HOI4 designers struggle with (a recent example was the addition of different types of recon).
Especially taking into account what happened with the MTG ship designer. But I understand the need to add some shiny features to put on the DLC product page or 3D models (such as the armored car in previous DLC).

A module system is only interesting if:

1. several significantly different templates for the same unit are viable. Otherwise —and this is what happened with MTG— you're simply adding extra clicks to achieve what could be already done before. In MTG, there is for example only 1 optimal detection ship template, so why force us to click on all the modules? Same for other ships.
2. the choices you make on modules actually have an impact. Making people choose between Armor A, Armor B or Armor C is useless if difference between those modules is negligible in the final fight calculations.
3. (applies to single player only): the AI is competent enough that designing templates actually has a payoff. In MTG, that isn't the case; any template will stomp the AI, so, after the first few hours of discovery, why bother designing templates? (navies are actually not needed in single-player as the AI kills all it ships purposefully on naval bombers, but that's another topic).

Additionally, no true effort was made to balance modules MTG in subsequent patches. Balance in a module system is essential, otherwise there are optimal templates which defeat the very goal of having a module system.

Question to the devs:
Currently, we already have many different tanks, however people overwhelmingly use the same ones (medium tanks).
What makes you think that adding a template designer is beneficial to the gameplay?
What's going to happen is that people will build the same optimal design repeatedly, with extra clicks now. It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

I disagree that the overwhelming majority of people use medium tanks. Lots of people use heavy and even light tanks (I mainly use heavy tanks even when im not playing MP).

 

[Rachidelson]

We are well aware that we need to get the AI to play well with the changes we are making. I think part of the issue here is that there are different expectations of what "good AI" looks like and how it feels to play against it. I would question that 70% of the community thinks the AI needs to be better, considering that almost half of the community plays on easy or very easy difficulties. What we do see is that a lot of people are disappointed when they see the AI make obvious mistakes, and we want to fix that as much as possible. Some of it is "the AI didn't do the thing I would have done and is therefore making a mistake", which is hard to work around when you have thousands of people having ideas about what the correct approach to a situation is. We do hope to show you some improvements to the AI later in development, when the features are more settled down, but I'm not going to use a loaded term like "we fixed the AI", because we probably disagree on what that means exactly.

I don't think it is very viable or even desirable to have an AI that plays optimally, or the current meta at all times. For the tank designer in particular, my goals are in order of priority:
1. The AI builds competitive tanks, meaning it researches tech at a good pace and puts modern tanks in the field
2. There is a degree of historical flavor to the AI's tank design and building approach, particularly if historical focuses are on
3. The AI adapts to circumstances to a degree (don't design heavy tanks if you have no industry base, for example)

Thank you for the kind reply! I definitely didn't know so many people played on easier difficulties, so that actually surprised me. Thank you ( once again ), it definitely answered my questions, hope you guys continue to make each new announcement better than the previous one, this DLC is definitely getting me on the hype train more and more each week!

 

[Markodds]

Nice

 

[ShortButThick]

"Equipment Designer" seems like an interesting name choice for a tank designer

Can't think of any other major arm out of the three which the designer could possibly be applied to...

 

[Almas]

Good Diary.
On the different Armor types: Historically the industrial base of countries was set up to use specific kinds of armor, England for example was quite famous for using riveted armor only. Will that somehow be represented in the game? For example only a set amount of production lines can build tanks with welded armor, so that the player would have to be conscious about what to use those lines on or investing recources to build more welding lines.

 

[Brownsnoot44]

When will horse design overhaul be added

 

[minimouse007]

It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

This is questionable game design as it basically adds complexity without depth. This is a major issue that the HOI4 designers struggle with (a recent example was the addition of different types of recon).
Especially taking into account what happened with the MTG ship designer. But I understand the need to add some shiny features to put on the DLC product page or 3D models (such as the armored car in previous DLC).

A module system is only interesting if:

1. several significantly different templates for the same unit are viable. Otherwise —and this is what happened with MTG— you're simply adding extra clicks to achieve what could be already done before. In MTG, there is for example only 1 optimal detection ship template, so why force us to click on all the modules? Same for other ships.
2. the choices you make on modules actually have an impact. Making people choose between Armor A, Armor B or Armor C is useless if difference between those modules is negligible in the final fight calculations.
3. (applies to single player only): the AI is competent enough that designing templates actually has a payoff. In MTG, that isn't the case; any template will stomp the AI, so, after the first few hours of discovery, why bother designing templates? (navies are actually not needed in single-player as the AI kills all it ships purposefully on naval bombers, but that's another topic).

Additionally, no true effort was made to balance modules MTG in subsequent patches. Balance in a module system is essential, otherwise there are optimal templates which defeat the very goal of having a module system.

Question to the devs:
Currently, we already have many different tanks, however people overwhelmingly use the same ones (medium tanks).
What makes you think that adding a template designer is beneficial to the gameplay?
What's going to happen is that people will build the same optimal design repeatedly, with extra clicks now. It would have been much better to work on finding use-cases for the existing tanks rather than adding new tanks with no use-cases!

while i do agree that mtg modules' poor balance actually removes the need for the whole system, the truth is, everybody focused on medium tanks... historically. It was just the good jack of all trades option armies needed in main armoured divisions. everybody still had light tanks or heavies, but focused on mediums. Actually there aren t many optimal ways to armour a tank: look at the t-14 armata, looks a lot more like a western-armoured tank, than the t-90 did. Also, the way stuff designed should tie in heavily with doctrines, infantry/cavalry tank distinction came from doctrines, just as the germans spamming the same medium tanks in independent divisions.

 

[Markodds]

One question thou - will I be able to move into one of those tanks when my wife eventually will divorce me?

 

[Beserker]

I honestly don't know how I feel about this. As someone already pointed out before, there'll end up being an optimal design and most people will just stick with that, rendering all of this choice somewhat useless. I was really excited for the naval designer, but at some point researching everything separately and playing around with design just becomes tedious, because every time you unlock something new, you have to go back and change it up to improve it. Also, what's the point of having the ability to make a bunch of different designs, when you'll probably just end up being able to have the manufacturing capacity to field significant numbers of one.

I know I'll probably be in the minority on this, but I'm really not sure this was necessary and it'll just end up being more tedious clicks and something for people that like to min/max.

As an aside though, it would be nice that whenever I unlock something new (say a turret) for a battleship, that I can fit it to the ship while it's still under construction, without losing all of the manufacturing progress. I don't like having to create a new class of ship or having to wait till I have all the modules researched before I construct the ship I want, since they take a long time to build. I will concede though, perhaps I just missed the ability to "refit" a ship still under construction.

Personally, watching naval battles and the like was really fun for me, up until MtG - to me, that's when the naval game became unplayable. But the reason is not the DLC, or the ship-designer, but the changes in naval costs - if you start with a navy, you already won, because you can set up naval invasions easily, and, if you didn't, you'll most likely just mass produce subs, because trying to do anything else will take forever, due to the cap on dockyards for larger ships. The designer is SUPER fun, but the fact that you can't get any kind of heavier ship as a minor just removes any interaction with the system.

I believe that comparing this tank designer to the naval designer is a fallacy, because the above point doesn't apply to tanks. They are much much cheaper than ships, so you can actually use the designer a lot more. What will probably happen is that, instead of having a single Tank template that you mass produce, you will have a couple of tank divisions of each different type.

This is just speculation, of course, but I am honestly very positive about this change, overall!