Question Regarding Map Changes, Provinces, and Wasteland

[Noble Steel]

It is my understanding that it has been the policy of Paradox not to add new provinces into EU4 mainly over issues of tech debt and performance issues. I was wondering if any devs have written anywhere on the reach of this methodology. For example, I would love a South America/Inca Rework. If we can't add new provinces, is it possible to relocate, or shrink/expand, provinces into a new layout whilst keeping the actual number the same? I especially wonder if this is true for wastelands, as I feel the Andes could use some strategic wastelands there to make it feel more mountainous/defensive, and less traversable in general. What is the design philosophy here, and the extents/limitations of this bottleneck? Are wastelands coded like provinces, or could we open/change/stretch the central interior layouts of wastelands in Australia/Amazonia for instance? Appreciate any insights.

 

[grotaclas]

I just did a quick search and the following two comments were the most important ones which I found

We are not planning at changing the map more in EU4's lifecycle. I am tired of people's savegames being broken.

So, regarding the map, we're not confident that updating/changing the map won't further affect the game's state, so we will not be doing more changes in this update, either. Believe me if I say that we'd like another round of updates/improvements, but as I just said, we have to be careful about what we rework/implement at this point of the game's life cycle.

When we are ready to implement more map changes again, we will be happy to announce them here as early as possible, but to be crystal clear even if repeating myself: There won't be map changes in 1.34 update.

I vaguely remember that I saw another developer comment below a dev diary which explicitly said that they won't change the shape of provinces either, but I can't find that comment right now, so maybe I just misremembered that. I think one reason for not doing slight map changes which have no performance impact is that it opens a can of worms and the playerbase would demand more changes as a result. And the developers might want to avoid discussions which would result from suggestions about the removal of provinces in one region to add them in another region(in addition to the save game issues which it would cause).

 

[Noble Steel]

@Pavía Una pregunta: I remember somewhere a dev posted something about wastelands being similar to provinces but they were open to being changed, do you remember where this was talked about? Appreciate any insights on the design philosophy and limitations in general about this. I know much of the community wants to know the extent of changes we can see on that topic. Gracias por todo!

 

[Less2]

It's a shame. The patch that got a lot of outcry for performance issues did add a ton of provinces, but it was actually the DLC mechanics that were badly optimized and they were fixed over the course of the next few patches. The game runs just fine now, apparently as well as ever from what I've seen. And there's definitely parts of the map that could use more provinces.

 

[grotaclas]

The patch that got a lot of outcry for performance issues did add a ton of provinces, but it was actually the DLC mechanics that were badly optimized and they were fixed over the course of the next few patches.

Do you have any evidence for this? I think the new provinces and tags were the main culprit for the slowdown and the game became faster again, because various parts of it were better optimized than before. So if you would use the 1.30 map with the game version 1.34, it would be much faster than it was in 1.30.

 

[Borgratz]

I remember Johan posting something like "no more map changes, it breaks save games and players absolutely hate it when we do that".
Can't search for it right now.

 

[Less2]

Do you have any evidence for this? I think the new provinces and tags were the main culprit for the slowdown and the game became faster again, because various parts of it were better optimized than before. So if you would use the 1.30 map with the game version 1.34, it would be much faster than it was in 1.30.

IIRC it was something to do with the new native american mechanics (maybe speculation on my part tho). The actual number of provinces added in Leviathan was just a few % increase but performance was down like 30-50%, so performance was way, way worse than could be expected from just the provinces and came from some kind of coding mishaps. There's also the fact that plenty of mods like M&T or Voltaire's Nightmare had significantly more provinces pre-Leviathan than vanilla Leviathan and still ran as well or better as long as you avoided updating to the Leviathan patch. I know there was plenty of discussion by modders about how they weren't updating to the newest patch because it tanked their performance to unplayable levels.

EDIT: Here's some quick commentary I managed to dig up from the M&T discord by the devs of that mod. Pretty sure these guys are some of the most knowledgable about how to eke performance out of EU4 and they said it was Leviathan changes rather than the map update. Obviously the M&T map is entirely its own thing and wouldn't be affected by vanilla map or tag changes.

 

[grotaclas]

The actual number of provinces added in Leviathan was just a few % increase but performance was down like 30-50%, so performance was way, way worse than could be expected from just the provinces and came from some kind of coding mishaps.

First of all, the Leviathan DLC didn't add any provinces. The 1.31.0 patch (codename Majapahit) did that. If you only count land provinces, the number increased from 3137 to 3272 which is 4.3%. The number of tags increased from 850 to 952 which was a 12% increase(I think the percentage increase in the number of tags which exist in 1444 was even higher, but I don't have a number for that). But both adding provinces and adding tags increases the performance in non-linear ways. As an unoptimized example, if all tags would check with all other tags if they want to do diplomacy with them, the number of checks would increase from 850*849=721650 to 952*951=905352 which is a 25% increase. Other kinds of algorithms can have an even worse increase in the number of computations. Finding the fastest path for an army might use such algorithms. They have to be optimized to even work with the number of eu4 provinces, but for example the many new empty provinces in north america can cause problems here, because an optimization which ignores provinces without mil access doesn't help in that case.

here's also the fact that plenty of mods like M&T or Voltaire's Nightmare had significantly more provinces pre-Leviathan than vanilla Leviathan and still ran as well or better as long as you avoided updating to the Leviathan patch. I know there was plenty of discussion by modders about how they weren't updating to the newest patch because it tanked their performance to unplayable levels.

EDIT: Here's some quick commentary I managed to dig up from the M&T discord by the devs of that mod. Pretty sure these guys are some of the most knowledgable about how to eke performance out of EU4 and they said it was Leviathan changes rather than the map update. Obviously the M&T map is entirely its own thing and wouldn't be affected by vanilla map or tag changes.

Thank you for this information. I didn't know that they ran better before 1.31 than vanilla 1.31 did. Especially about M&T I have heard that you need a very fast computer to even run it and Voltaire's nightmare also felt much slower than even vanilla 1.31 when I tried it(but I don't know if I ever tried it before 1.31 and I never really played it, so I my experience might just have been skewed by something).
But the quote from the developers explicitly mention 1.31 changes as the culprit and not DLC mechanics from the leviathan DLC. If it were just Leviathan DLC mechanics, they could have disabled the problematic ones in their mod. If the problem were the native mechanics from the conquest of Paradise DLC, disabling might have been more difficult if the mod relied on migration which has existed before 1.31.

In any case, without extensive performance measurements we can't really say how big the impact of the number of provinces/tags is on performance and we don't know which amount of performance impact the player base would accept to get more provinces in neglected parts of the map. If there ever will be an eu5, it will have a better engine which might be able to handle much more provinces. Or it could move more computations to the state level like in vic3 (but hopefully allow peace deals which take only parts of a state to satisfy the part of the playerbase which likes historical borders).

 

[brifbates]

EDIT: Here's some quick commentary I managed to dig up from the M&T discord by the devs of that mod. Pretty sure these guys are some of the most knowledgable about how to eke performance out of EU4 and they said it was Leviathan changes rather than the map update. Obviously the M&T map is entirely its own thing and wouldn't be affected by vanilla map or tag changes.

View attachment 953201

The quote you post says absolutely nothing about the source of the performance degradation. All it says is they aren't bothering to attempt to port the mod over beyond 1.30 because the game performance is crap...

 

[Less2]

But the quote from the developers explicitly mention 1.31 changes as the culprit and not DLC mechanics from the leviathan DLC. If it were just Leviathan DLC mechanics, they could have disabled the problematic ones in their mod.

Fair enough, I think most people just consider 1.31 and Leviathan to be synonyms. Also, its entirely possible that DLC mechanics are still being processed in the background even if they aren't being used. This is why I think natives were blamed because they had a lot of new province mechanics with their territory and stuff, whatever was being processed if it scaled poorly with # of provinces could be the issue.

I didn't know that they ran better before 1.31 than vanilla 1.31 did. Especially about M&T I have heard that you need a very fast computer to even run it and Voltaire's nightmare also felt much slower than even vanilla 1.31 when I tried it(but I don't know if I ever tried it before 1.31 and I never really played it, so I my experience might just have been skewed by something).

To be clear, "better" is a bit subjective. The performance degredation from 1.31 and the slowdown from mods were very differently feeling things. 1.31 was very stuttery while these mods were overall a bit slower than 1.30 vanilla but still fairly steady going.

f you only count land provinces, the number increased from 3137 to 3272 which is 4.3%. The number of tags increased from 850 to 952 which was a 12% increase(I think the percentage increase in the number of tags which exist in 1444 was even higher, but I don't have a number for that). But both adding provinces and adding tags increases the performance in non-linear ways. As an unoptimized example, if all tags would check with all other tags if they want to do diplomacy with them, the number of checks would increase from 850*849=721650 to 952*951=905352 which is a 25% increase.

I understand the mathematics, but then you go to a mod with vastly more provinces and it doesn't scale nearly this poorly. IIRC Voltaire's Nightmare has 3000 provinces in the HRE alone.

 

[kyrios00]

It is my understanding that it has been the policy of Paradox not to add new provinces into EU4 mainly over issues of tech debt and performance issues. I was wondering if any devs have written anywhere on the reach of this methodology. For example, I would love a South America/Inca Rework. If we can't add new provinces, is it possible to relocate, or shrink/expand, provinces into a new layout whilst keeping the actual number the same? I especially wonder if this is true for wastelands, as I feel the Andes could use some strategic wastelands there to make it feel more mountainous/defensive, and less traversable in general. What is the design philosophy here, and the extents/limitations of this bottleneck? Are wastelands coded like provinces, or could we open/change/stretch the central interior layouts of wastelands in Australia/Amazonia for instance? Appreciate any insights.

mate , just do what most of us are doing and get the Infinite timeline mod from the workshop. The map is way better, a lot of provinces, and the countries you want to play actually look like the countries that exist, not a hunchback version of them. I m not gonna say anything bad about the ones that drew the initial countries but dayum the need glasses.... some countries have no resemblance in shape or form like the real ones... Also, that mod gives you more than a handful of years to play and waaay more eras .... base game is nothing without that mod... regardin the corrupted saves... seeing how we play that mod because it s better than base, they get corrupted anyways so yeah... but it s nice that i get to conquer all euro-asia and africa, instead of barely managing something in whatever hundreds of years the initial time is...

 

[Blindbohemian]

It's a shame. The patch that got a lot of outcry for performance issues did add a ton of provinces, but it was actually the DLC mechanics that were badly optimized and they were fixed over the course of the next few patches. The game runs just fine now, apparently as well as ever from what I've seen. And there's definitely parts of the map that could use more provinces.

It also added tons of provinces where they were completely unnecessary. Fiji has three provinces. That’s more than Cyprus. Navarra. It’s on part with Peninsular Greece. Australia could be just one big wasteland and barely affect the game, but it’s got like 30 provinces. I can’t believe they let that be the end of EUIV map updates instead of just rolling back that terrible patch with an apology.

 

[brifbates]

It also added tons of provinces where they were completely unnecessary. Fiji has three provinces. That’s more than Cyprus. Navarra. It’s on part with Peninsular Greece. Australia could be just one big wasteland and barely affect the game, but it’s got like 30 provinces.

Closer to 50, actually, it's utterly ridiculous. Almost as ridiculous as it being colonized in the early 1600s as it is in most campaigns...

 

[Blindbohemian]

Closer to 50, actually, it's utterly ridiculous. Almost as ridiculous as it being colonized in the early 1600s as it is in most campaigns...

Seriously. It would be more historical for the whole continent to be wasteland.

(More historical than the current situation, of course, being a very low bar).

 

[Rudolf Nechvile II]

It also added tons of provinces where they were completely unnecessary. Fiji has three provinces. That’s more than Cyprus. Navarra. It’s on part with Peninsular Greece. Australia could be just one big wasteland and barely affect the game, but it’s got like 30 provinces. I can’t believe they let that be the end of EUIV map updates instead of just rolling back that terrible patch with an apology.

I think this is the most bitter thing about it. A lot of areas got insanely unnecessary amount of provinces despite a lack of historical importance, which I think is the greatest source of resentment for most people; although, I have no doubt that originally they intended to give the entire map such a level of detail. I don't think most would mind as much if the total number of provinces stayed the same, but they reshaped and redistributed some provinces, because I don't think we really need a highly detailed Australia, nor a 3 province Fiji or 4 province Hawaii.

I may be wrong, but I also think that the argument of save games breaking has always just kind of been a bit of a nonsense reason, because save games might still break with every major update, regardless. We still get this warning with every major patch:

Please note that 1.35 saves may not be compatible with 1.36! See here for information on how to roll back versions or prevent auto updates.

EDIT: I also would like to add that I think mountain wastelands are criminally underused and some areas, especially in Europe could use some, like Anatolia and Spain, or as OP mentioned, the Andes.

 

[klingonadmiral]

And yet they changed the map with 1.36

 

[Oporto]

And yet they changed the map with 1.36

Only in regards of tags present on it - not in terms of provinces' amount.

 

[klingonadmiral]

Only in regards of tags present on it - not in terms of provinces' amount.

Ardabil was shifted from Shirvan area (Caucasia Region) to Tabriz area (Persia region).

 

[Oporto]

Ardabil was shifted from Shirvan area (Caucasia Region) to Tabriz area (Persia region).

It's more in term of reshuffling of provinces between the areas, they did it in "Domination" too - with "Asturias" province for example.

But wait... I don't recall seeing Ardabil in "Shirvan" area before. Maybe it was actually done years ago?

 

[Oporto]

Nevermind, it's seems they really did it in 1.36. Here's a funny thing: In wiki it's still in the Caucaus region:

But when you actually click on the image to zoom in, it looks like this:

https://eu4.paradoxwikis.com/Middle_East_regions

@grotaclas I conjure and summon thee!