• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Hoi4 Dev Diary - Telemetry and data

Hello there, the time has come to do a fun dive into what you've been playing the game as a community. So over this diary we are going to look into some of the trends and statistics we seen from our telemetry and see how we use it to gain insights in current player trends when working on HOI4. The wonderful data team have been working to get us all the information we wanted and I hope you enjoy seeing some of their great work. As a forward point, as with any data collection we do not store any personal information or anything that can used to identify players in accordance with GDPR.

So first of we will start with something fun, can you guess what this graph is telling us.
1645107087765.png


If you guessed this was us spotting the continuous naval production focus exploit being discovered you get a gold star, although it seems some of you are still very much into using the focus.

With telemetry we can work to understand our players by seeing how you play and this lets us form an experience players will enjoy. we see this in many ways, for example we can see by player difficulty settings players really shy away from the harder difficulties.
1645107138178.png


We can also see how players are reacting to new content, here we can see the soviet union and Poland grow and maintain solid new highs in terms of sessions played. Germany still remains you favourite major to play, and for those of you that like playing minors the PRC seems to be the country to go for.
1645107317495.png


Now lets get into some more specific game related data, player play behavior's are something that always interesting to know. You might think you know what most players are doing from discussions on various discussion platforms, streams and videos, however what's even better is knowing for sure what the meta and average player does.

For example many people really need to check their infantry equipment stockpile late game
1645106739776.png

Seriously players really enjoy making infantry equipment 3
1645106760437.png
Alright that's allot of graphs, and we know graphs are not how we get views, so lets take tried and trusted top 5's for spin.

We will start with rapid fire round of nukes so.

Top 5 target nations:
  1. Germany
  2. Soviet union
  3. Japan
  4. USA
  5. England
Top 10 Target provinces:
  1. Tokyo
  2. London
  3. Berlin
  4. Hiroshima
  5. Moscow
  6. Nagasaki
  7. Washington
  8. Osaka
  9. Nagoya
  10. Paris
Top 5 countries nuking:
  1. Germany
  2. Soviet union
  3. USA
  4. England
  5. Italy
As we can see players are really not very imaginative with their nuke targets or they share a strange wish to nuke all the same places. Please also let me know why you really like nuking Paris in the comments below.

We can now take a trip over to something newer, its that new shiny toy the tank designer. I must say if there was a meta with tank design someone forgot to tell players because the only consistent design trend has been making cheap flame tanks.

So here we have your favourite modules for each category over the last month
MODULE_TYPEMODULE_NAME
ARMORtank_welded_armor
CHASSIShv2
ENGINEtank_gasoline_engine
GUNtank_small_cannon_2
SUSPENSIONtank_christie_suspension
TURRETtank_medium_three_man_tank_turret
extra_moduleadditional_machine_guns
extra_modulewet_ammo_storage
extra_moduletank_radio_1
extra_modulesloped_armor

How about your favorite Tank role designs, featuring some monster MBT design that players love.
ROLECHASSISGUNTURRETSUSPENSIONARMORENGINEMODULE1MODULE2MODULE3MODULE4ENGINE_UPGRADEARMOR_UPGRADE
mmbt0hc3mbttorcasdierad3smokeasyslop2020
mlt2csl2mchrrivgasrad1---------32
mlt2sc2l2mchrwelgasrad1---------102
mmt2mh1m3mbogrivdierad2tsctscmg64
mlt2ac1l2mchrcasgasrad1---------32
mlt2csl3mchrrivgasmgmgmgmg90
mlt2sc2l2mchrrivgasrad1---------32
mlt2csl3mchrwelgasrad1---------102
mlt2ac1l2mchrwelgasrad1---------32
mlt2fltl2mchrrivgasrad1---------32

You guys also have some erm "interesting" collective opinions when it comes to amphibious tanks.
ROLECHASSISGUNTURRETSUSPENSIONARMORENGINEMODULE1MODULE2MODULE3MODULE4ENGINE_UPGRADEARMOR_UPGRADE
ammt3mc2m2mchrcasturrad3amphweteasy200
ammt2mc2m3mintwelgasrad3mgmgamph60
ammt3mc2m3mtorcasturmgslopamphrad3110
ammt3mc2m3mchrwelgasrad3slopamphmg99
amhv3hv3h3mchrcasturrad3amphslopsque2020
amhv3hc2h3mintwelgasrad3slopsmokamph88
ammt3hv2m3mtorwelgasrad3slopsmokamph56
ammt3mh2m3mchrcasturrad3easywetamph200
ammt2mc2m3mbogrivgasrad3mgmgamph60
ammt3mc2m2mtorweldieweteasyamphrad31510

We have also seen that about 10% of you and only about 4% of total designs are made with auto design, this is something that reinforces our desire to improve this element of the designers.
1645106778449.png


Alright lets get into some more gritty details. first off we have been seeing how player designs have been changing since release in response to changes in patches and the design meta developing.
1645106789765.png

1645106795449.png

1645106800452.png

Next up lets see how you've been using those handy preferred tactics, it neck and neck over what players think is the best one for the army.
1645106830179.png

But you have a clear favourite for generals.
1645106844393.png

But what about all those new focus trees, well lets take a look at how you traverse focus trees. here we look at what exclusive choices you are making when playing with the new focus trees. As a community players will always trend toward the path that lets them expand and be independent so most of the breakdowns are as expected. however its also very useful to know what specific choices players make within these paths.

sov_focus_selection.jpeg

pol_focus_selection.jpeg

lit_focus_selection.jpeg

lat_focus_selection.jpeg

est_focus_selection.jpeg

axis_allies_baltic_focus.jpeg

sov_industry_focus.jpeg

Construction now, when to switch to MIL's from CIV's is always an topic for debate. as we can see though the change happens on average quite early with the first year averaging at 13.17 CIV's built. This is consistent over most majors outside SOV, with many being much more extreme than the average.
1645106861728.png

Additionally you always need more railways and infrastructure.
1645106873810.png

Finally I'll bring us to Division meta. this might be surprising but meta divisions while definitely popular, are not as ubiquitous as online discussion would have you believe.
1645106884415.png
As you can see we have the usual suspects of the 10-0's, 7-2's and 14-4's but they are nowhere near dominant in the average game.
The situation becomes even less unified once we take support companies into account
1645106896799.png


That wraps it up for this Dev Diary, If you have anything specific you're interested in from this brief overview let me know and maybe we can do a deep dive in future.
 
  • 60
  • 52Like
  • 8Love
  • 3Haha
  • 2
Reactions:
I believe Stellaris has different AI behaviour for different difficulty levels, also the Expert AI mod for HoI4 allows you to toggle AI behaviour settings. For example how much they min-max their divisions.
The main difference for Stellaris difficulty levels is how much AI cheats give them extra resources. This in turn would give different behaviour as they hit their economic plans sooner. The situation is in flux as the Custodian team are re-tooling the Stellaris AI again.

Mods are a separate topic: unpaid developers with no time budget to worry about can achieve great things. Corporate developers with deadlines to meet have to worry about the cost effectiveness of having lots of toggleable AI settings: every permutation ought to be tested as the player base (rightfully) get very cranky when something doesn't get caught during testing. They have to do that while keeping up with new features: both coding said features and then getting the AI to understand what to do with them. Add managers getting cranky for one reason or another, which includes "why is this task taking longer than it ought to take? Wasn't this supposed to be easy?"

That's on top of how we already have toggles for AI behaviour: which focus tree path it should pick for certain nations. That's another way to tweak how difficult a campaign is: change which nations are in the faction you'll be fighting.
 
  • 1
  • 1
Reactions:
Mods are a separate topic: unpaid developers with no time budget to worry about can achieve great things. Corporate developers with deadlines to meet have to worry about the cost effectiveness of having lots of toggleable AI settings: every permutation ought to be tested as the player base (rightfully) get very cranky when something doesn't get caught during testing. They have to do that while keeping up with new features: both coding said features and then getting the AI to understand what to do with them. Add managers getting cranky for one reason or another, which includes "why is this task taking longer than it ought to take? Wasn't this supposed to be easy?"
Obviously, but you can say that about any feature. Scandinavian focus trees? Air rework? Better defensive warfare? No can do, that'd mean more work to balance and bug test, and we can't have that.

That's on top of how we already have toggles for AI behaviour: which focus tree path it should pick for certain nations. That's another way to tweak how difficult a campaign is: change which nations are in the faction you'll be fighting.
No, that's got nothing to do with AI difficulty levels.
 
  • 2Like
Reactions:
  • 4Haha
Reactions:
The reason I don't play at the higher difficulty levels is that I prefer higher diffculty levels to have my opponents buffed, not my country being nerfed. I usually play EAI on hard or very hard.
Disagree because i dislike the powercreep that entails.

I would much rather to fight with half my division than the same foe with twice his divisions.
 
  • 1
Reactions:
The reason for that is not arrogance, but the simple acceptance of the fact, that HOI4 (like every PDX game) has a learning curve. No master just fell from heaven, this is the most normal thing in the world. But to answer a question like "what is my favourite / most used division template" I first and foremost need to know how to design one in the first place. To say that 9inf is the most commonly used is not an information at all, for me at least, because that just means, that a sizeable amount of playes don't know how to change a template.
I certainly won't argue about the learning curve, every major change to the game is a pain, butg that is okay. I disagree that a sizeable amount of players don't know how to change templates. Most of the players, if not all are very knowledgeable about the internet and how to use it. There are so many let's play videos out there, and some directly deal with the nuts and bolts like changeing templates. Of course, you also have this forum which provides all sorts of informtion. I think it is more the fact that they don't want to be bothered dealing with it. It would be interesting if Paradox could do a survey on some of these questions that the telemetric data can't show.

I apologize that I misinterpreted your post, but it sure seemed as if you fell into the hard-core category.
 
  • 1
  • 1Like
Reactions:
I certainly won't argue about the learning curve, every major change to the game is a pain, butg that is okay. I disagree that a sizeable amount of players don't know how to change templates. Most of the players, if not all are very knowledgeable about the internet and how to use it. There are so many let's play videos out there, and some directly deal with the nuts and bolts like changeing templates. Of course, you also have this forum which provides all sorts of informtion. I think it is more the fact that they don't want to be bothered dealing with it. It would be interesting if Paradox could do a survey on some of these questions that the telemetric data can't show.
I doubt that any tutorial/lets-play/whatever available on the internet recommends 9inf. However, what came to my mind in the mean time, that this might be an error on PDX side, for counting build divisions, not used. This would explain the 9inf.
I apologize that I misinterpreted your post, but it sure seemed as if you fell into the hard-core category.
All fine.
 
  • 1
Reactions:
But what about all those new focus trees, well lets take a look at how you traverse focus trees. here we look at what exclusive choices you are making when playing with the new focus trees. As a community players will always trend toward the path that lets them expand and be independent so most of the breakdowns are as expected. however its also very useful to know what specific choices players make within these paths.








It's a bit late, but I wanted to add that when I play the USSR/Russia I almost always go down the path of the Center, because the Opposition Branches mean you have to fight a civil war, which can greatly weaken your country; while with the Exiles paths feel underdeveloped to me. The only time I break away from the Center is go to the Right, because it seems actually reasonable. I feel like the Second Civil War should be easier to achieve if you've opted to stay Marxist-Leninist. For the Rightists, the bloodless coup should be attainable at the start of the game with some reasonable effort, while for Trotsky the civil war should be an uphill battle; but should be able to get momentum within the Party.

The exiles war just feels frustrating. When I did it, it took about 2 years. By that point the Allies were already at war with Germany, and Germany was poised to invade me within a year or two. I meant that I would have to take my 40-60 divisions and hold the Germans with a ravaged nation, its difficult and pointless. Not to mention that your probably missing some states that broke away. In short, the Exiles need some buffs to assist them in getting back to power, this includes some generals. They start with a ton of Field Marshals, but not a lot of generals. Its underwhelming to play the exiles, because once you get to picking a genuine political path, the main war has started and you've shifted focus to winning that war. In short, the exiles just need a bit more. It feels like the only really developed part of them is getting to power. Not to mention it's hard to simulate the division within the White's camp, which could be modeled via the Faction system (BftB). I feel like the Whites should be able to become an enormous powerhouse, but would be kneecapped by internal divisions over what political direction Russia should take. Also redo the Third Rome path. I feel like the Patriarch wouldn't take direct power, but have a Tsar as his puppet and doing his political work.
 
  • 2
Reactions:
I'm slightly surprised at the nuke statistics. In some games I drop a lot of nukes and the enemy capital is never the target. The usual targets are all the airfields with strategic bombers. I guess that is because reaching "nuke time" usually means I'm playing the Soviets and still having air superiority problems, nukes fix that sharpish if you have a few of them. In the game I just finished I think I must have nuked at least 25,000 aircraft and probably a lot more.
 
  • 1Like
  • 1
Reactions:
Considering fact that there were several threads and many posts grumping about soviet opposition paths and still nothing then... Abadon your hope and try find some good mods / mod game yourself. Devs abadoned all alt-history paths for Soviet.
 
  • 1
Reactions:
Considering fact that there were several threads and many posts grumping about soviet opposition paths and still nothing then... Abadon your hope and try find some good mods / mod game yourself. Devs abadoned all alt-history paths for Soviet.
No, I won't refuse. I bought a product and expect a certain quality from it. It's like buying a new car with a broken tire and putting up with it and going looking for other ways instead of having the dealer fix the broken wheel. If that doesn't happen, I won't buy another DLC.
 
  • 1Like
Reactions:
That would mean either having multiple AIs, which would need separate support and development, or force the AI to take decisions that itself would consider bad. Neither sound like a good scenario.
Erhh. Not at all. You can selectively disable/enable components of an overall AI design. Tailored aggressiveness (and other behavior trends) also change difficulty levels.
 
I wouldn't jump to conclusions on this. 21% of people just use 9-0 pure infantry, the starting template for many nations. To me that says many players find division design uninteresting/unimportant and ignore the system. It works out because 18w is at least decent in most terrains, but I don't think it's some great achievement that there's no meta division on top. I'd be interested to see the division type breakdown by difficulty in single player. SP vs MP would be an interesting comparison as well.
^This

The problem was never the prevalence of meta width.

The problem was and still is the unnecessary and steep over-penalty against exceeding combat width. All you need is for exceeding combat width to not be an advantage. (ie. 10% over width? match it with 10% penalty) And that does not require penalizing the whole front. Even for historical immersion, a slight penalty does the job. (ie. 10% over width? slightly outmatch it with 11% penalty.) Right now, the steep penalty results incentive for ultra-micro play to fit combat width closely and more importantly avoid that additional division that exceeds combat width -- that is... terrible...

The old meta width of 25 wasn't even the main reason of ahistorical divisions -- though it contributed. That culprit goes to armor weight. All in all, quite a bit misguided effort in my book.
 
  • 1
Reactions:
This is what I don't like about the super hardcore players. They think the game is only for them and any info about it should concentrate on their likes/dislikes.
And this professed knowledge about some "super hardcore players" sub-group comes from not-a "super hardcore players"?

Sounds like inherent logical contradiction to me.