Civ VII Post-mortem: Crafting a redemption arc

[stealth_nsk]

So if we take that pivot table, basically the overall team grew by about 75% between versions. Of the big teams, the dev team grew by 21%, the art team by 41% (and the audio team by 22%) the 3 teams who were under that mark, and UI, QA, and a lot of the production and management teams grew by more than that amount.

What that means? Who knows.

I think it's hard to assess, because team responsibilities shift. For example, there are some new teams and their responsibilities were distributed among other teams in Civ6. So, each new team is kind of hidden growth.

 

[Gorbles]

Thanks a lot! One minor correction - development started way earlier - Firaxis shown playable prototype to 2K managers in 2019. I'd guess this prototype used Civ6 engine (it's pretty common for early prototypes to be made on previous or common engines) and work on Civ7 engine started later after the general requirements for the engine were fleshed out.

Firaxis (like most studios with an in-house engine) seem to build on previous iterations instead of starting new versions from scratch. The engine for VII will have started as the version that still exists in VI (and to VI, from V, in turn).

Chunks of it will probably have been rewritten (game engines tend to be a lot more modular and less monolithic than folks presume), but mapping the line of which game's engine it is is . . . pretty difficult. Very Ship of Theseus-like.

 

[aieeegrunt]

Firaxis (like most studios with an in-house engine) seem to build on previous iterations instead of starting new versions from scratch. The engine for VII will have started as the version that still exists in VI (and to VI, from V, in turn).

Chunks of it will probably have been rewritten (game engines tend to be a lot more modular and less monolithic than folks presume), but mapping the line of which game's engine it is is . . . pretty difficult. Very Ship of Theseus-like.

I would not be surprised if there are still significant chunks of the Halo Blam engine (used for Halos 1 -5) in fhe supposedly new Slipstream (used for Halo Infinite).

 

[stealth_nsk]

Firaxis (like most studios with an in-house engine) seem to build on previous iterations instead of starting new versions from scratch. The engine for VII will have started as the version that still exists in VI (and to VI, from V, in turn).

Chunks of it will probably have been rewritten (game engines tend to be a lot more modular and less monolithic than folks presume), but mapping the line of which game's engine it is is . . . pretty difficult. Very Ship of Theseus-like.

Yeah, I played with some internals for Total War and Elder Scrolls series and they have a lot of legacy inside. For Total War they even kept all properties from previous games, so you're looking at data for ancient Rome units and see properties for samurai things or gunpowder.

I didn't look inside Civ engine, but as a sad owner of HDD I often see how resources are loaded in real time and I think Civ 5, Civ6 and Civ7 do it pretty differently. I surely expect a lot of code to be shared between those version, but I think the engines in Civ are more different than those in TW or ES.

 

[Kenshiro70]

So if we take that pivot table, basically the overall team grew by about 75% between versions. Of the big teams, the dev team grew by 21%, the art team by 41% (and the audio team by 22%) the 3 teams who were under that mark, and UI, QA, and a lot of the production and management teams grew by more than that amount.

What that means? Who knows.

While the management team grew by a large percentage, it was a small absolute number, and if you look at the positions, they are the standard ones that a company needs as it gets above 100 employees (e.g. Chief Operating Officer).

For the others, it's best to look at them in context of both the game and the outsourced vendors. For better or worse, the game is designed to create an evergreen pipeline of DLC to sell. The art and audio teams grew as a result. The outsourced teams also art-heavy as well. That pipeline requires people to coordinate bringing all of those assets into a package and getting them ready to release; which accounts for the increase in producers.

Unfortunately, we've now seen the flip side of that, when the sales can't sustain the staffing. I checked LinkedIn profiles of 27 of the 70, and 13 were from the Art Team (48%), while Production accounted for 15%. That highlights the real danger to Firaxis - hollowing out. It's cheaper to outsource the artwork than maintain full time positions.

Tracking that back to flawed design principles, it seems like someone made an explicit decision that Civilizations are very content-heavy (building models, animations, etc.) That decision had real impact on the staffing levels and composition. If only the civs were as fun to play as they are beautiful.

But those are obvious issues. I prefer to look for the dogs that don't bark - things that are missing. For example, there's no one at Firaxis explicitly responsible for Localization (L8N). It's all handled by outsroucers and 2K people, and from what I've hear it's a mess - not just poor translations, but fixed translations being overwritten by later patches (= regressions). That's actually a bigger deal than it fist appears, as localization strings are held n XML files, just like the other game design items are, which causes concern about data integrity. But that's a subject for another day.

I'm not surprised at all by the QA growth, as Firaaxis is finding out the hard way that even if you used standards-based technology, exch platform (consolte) implements it just differently enough that you can't skimp on the testing. Unfortunately the QA didn't anticipate this and build out a testing automation plan, and now they are hopelessly behind. We're all paying the price for that (e.g., the Modern Age yield explosion).

What's more surprising is that Engineering didn't grow all that much. Given Sid's background as an engineer, Firaxis has tended to be a little more development-heavy. I think I know the reason - the developers outsmarted themselves by relying too much on metadata for the leader and civ abilities. There's a section on this in the Google Doc that I may promote into the thread:

---

Metadata: Amazing, but too much of a good thing​

---

The term “Metadata” is short-hand for a coding approach called Declarative development. Traditional coding combines both logic and values together. For example, a Library might have the code “On turn start, add 5 to Research”. Declarative pulls the value 5 out into an XML file. So if you want to change the yield of a Library, you don’t have to recompile code; you can simply edit the text file.​

​

This approach is enormously powerful, because it allows you to create and test new buildings quickly, as long as they follow the basic template. As a bonus, it's also easier to pass console patch certification - no code changes means it's much easier to pass security tests.​

​

Note: If you want to see the game metadata, most of it can be found in the game directory in “\Base\modules\base-standard\data”. Knock yourself out.​

​

Of course, if you want to change the template, that requires engineering involvement. For example, you'd like to add a condition, "if next to 'River' terrain, add 2 bonus science" You'll need to change the code and go back and make sure both that the new building types work and that the existing building definitions don't break.​

​

Fun real-world fact: Remember the massive Crowdstrike/Windows meltdown in July 2024? If you read through the analysis of what happened, all twenty pages of fluff and misdirection, you'll find that the root cause was that they improperly tested a new parameter they added.

​

This is where you can get too much of a good thing. If you could make five new buildings by following the template versus adding a new attribute to the template, which would you do? Did I mention that you've got to get four civs delivered this month for the next DLC? So you can see how over time there starts to be a bias towards using the existing templates.​

​

The problem is that from the player's perspective, what a designer would call "template", the players call "cookie cutter". Humans are very good at pattern recognition, and if you overuse the existing cookie cutters, they'll notice. It doesn't matter if this cookie has sprinkles on it and that one has green food coloring if the basic recipe is the same.​

​

That's essentially what's happened with the Leaders and Civs. They've got too many "+1 Food on improvement adjacency" abilities instead of unique abilities. Open up the "leaders-gameeffects.xml" file and read through it, and you’ll see that I mean. In fact, if you look at the 1.1.0 and 1.1.1 patch notes, the majority of items were tuning of parameters rather than code changes.​

​

The end result is that Leaders feel like cheap knock-offs of one another. It's hurting replayability because they don't feel that different to play.​

​

The only way out of this trap is that Firaxis will have to ratchet back on the template usage and enforce a ratio of 2 template-based abilities allowed for every new unique code-based ability.​

---

That over-reliance on metadata is one of the more subtle traps and reasons that designers and developers can fall into.

 

[Verified_Confection_Being]

Hadn't heard that about localisation issues being reintroduced after being fixed. That smells a lot like pipeline issues ie. Outsourced company merging into master, and internal development team not syncing with master before continuing development.

Everything I hear about where things are going wrong with this game feels like it's in practice and fundamentals of software development. You can get a company filled with the best individuals at their jobs in the world but if your processes to get them working together are flawed you're still gonna end up with a mess

 

[UWHabs]

While the management team grew by a large percentage, it was a small absolute number, and if you look at the positions, they are the standard ones that a company needs as it gets above 100 employees (e.g. Chief Operating Officer).

For the others, it's best to look at them in context of both the game and the outsourced vendors. For better or worse, the game is designed to create an evergreen pipeline of DLC to sell. The art and audio teams grew as a result. The outsourced teams also art-heavy as well. That pipeline requires people to coordinate bringing all of those assets into a package and getting them ready to release; which accounts for the increase in producers.

Unfortunately, we've now seen the flip side of that, when the sales can't sustain the staffing. I checked LinkedIn profiles of 27 of the 70, and 13 were from the Art Team (48%), while Production only dropped 15%. That highlights the real danger to Firaxis - hollowing out. It's cheaper to outsource the artwork than maintain full time positions.

Tracking that back to flawed design principles, it seems like someone made an explicit decision that Civilizations are very content-heavy (building models, animations, etc.) That decision had real impact on the staffing levels and composition. If only the civs were as fun to play as they are beautiful.

But those are obvious issues. I prefer to look for the dogs that don't bark - things that are missing. For example, there's no one at Firaxis explicitly responsible for Localization (L8N). It's all handled by outsroucers and 2K people, and from what I've hear it's a mess - not just poor translations, but fixed translations being overwritten by later patches (= regressions). That's actually a bigger deal than it fist appears, as localization strings are held n XML files, just like the other game design items are, which causes concern about data integrity. But that's a subject for another day.

I'm not surprised at all by the QA growth, as Firaaxis is finding out the hard way that even if you used standards-based technology, exch platform (consolte) implements it just differently enough that you can't skimp on the testing. Unfortunately the QA didn't anticipate this and build out a testing automation plan, and now they are hopelessly behind. We're all paying the price for that (e.g., the Modern Age yield explosion).

What's more surprising is that Engineering didn't grow all that much. Given Sid's background as an engineer, Firaxis has tended to be a little more development-heavy. I think I know the reason - the developers outsmarted themselves by relying too much on metadata for the leader and civ abilities. There's a section on this in the Google Doc that I may promote into the thread:

---

Metadata: Amazing, but too much of a good thing​

---

The term “Metadata” is short-hand for a coding approach called Declarative development. Traditional coding combines both logic and values together. For example, a Library might have the code “On turn start, add 5 to Research”. Declarative pulls the value 5 out into an XML file. So if you want to change the yield of a Library, you don’t have to recompile code; you can simply edit the text file.​

​

This approach is enormously powerful, because it allows you to create and test new buildings quickly, as long as they follow the basic template. As a bonus, it's also easier to pass console patch certification - no code changes means it's much easier to pass security tests.​

​

Note: If you want to see the game metadata, most of it can be found in the game directory in “\Base\modules\base-standard\data”. Knock yourself out.​

​

Of course, if you want to change the template, that requires engineering involvment. For example, you'd like to add a condition, "if next to 'River' terrain, add 2 bonus science" You'll need to change the code and go back and make sure both that the new building types work and that the existing building definitions don't break.​

​

​

​

This is where you can get too much of a good thing. If you could make five new buildings by following the template versus adding a new attribute to the template, which would you do? Did I mention that you've got to get four civs delivered this month for the next DLC? So you can see how over time there starts to be a bias towards using the existing templates.​

​

The problem is that from the player's perspective, what a designer would call "template", the players call "cookie cutter". Humans are very good at pattern recognition, and if you overuse the existing cookie cutters, they'll notice. It doesn't matter if this cookie has sprinkles on it and that one has green food coloring if the basic recipe is the same.​

​

That's essentially what's happened with the Leaders and Civs. They've got too many "+1 Food on improvement adjacency" abilities instead of unique abilities. Open up the "leaders-gameeffects.xml" file and read through it, and you’ll see that I mean. In fact, if you look at the 1.1.0 and 1.1.1 patch notes, the majority of items were tuning of parameters rather than code changes.​

​

The end result is that Leaders feel like cheap knock-offs of one another. It's hurting replayability because they don't feel that different to play.​

​

The only way out of this trap is that Firaxis will have to ratchet back on the template usage and enforce a ratio of 2 template-based abilities allowed for every new unique code-based ability.​

---

That over-reliance on metadata is one of the more subtle traps and reasons that designers and developers can fall into.

Relying on the metadata/modular design yeah it can be a crutch, but at the same time, it's also a way to shift work from developers to others. It means a dev just needs to set up the foundation, but someone else who is not a coder can edit tech costs, building yields, etc... and be able to give some basic updates or bonuses. If you want to change a +50% to a +25% you can do it easily. And obviously it's a godsend for modding and modularity.

The downside as you mention is that often the bonuses tend to be same-same. +1 to blah. And you also can suffer a little bit because the AI now has to react to whatever randomness is in the files. If you change a library from 2 science to 3 as its base yield, you need to make sure the AI will appropriately react to that change.

Although frankly, the civ series has been this way for multiple versions, so the fact that it works this way again is not a surprise. I do think every civ has some unique abilities at least, so it's not like every civ just gets a bonus to one building type and that's it. I do think the leaders probably could use a little more of a refresh for some, but I don't think they do a bad job at making sure there's some unique abilities around.

 

[Kenshiro70]

Hadn't heard that about localisation issues being reintroduced after being fixed. That smells a lot like pipeline issues ie. Outsourced company merging into master, and internal development team not syncing with master before continuing development.

Look at one of the Localization XML files. There's not a modified date one each string. That means they're synching at the file level. With a distributed team. Ruh Roh.

If they're sending the file to multiple translators (e.g., 2k subsidiaries) and then trying to merge them back together, there's a huge chance of regression because the file modified date is not reliable for use in conflict resolution.

I'm a little stunned because this stuff is pretty basic engineering, both adding per-tag modification dates and anticipating out-of-order data load scenarios. Even worse, there are tools that will do all this for you. It's a classic build-vs-buy mistake.

It's not just a L8N issue. It's potentially a data integrity issue. The other XML files with the game data have the same weakness and can overwrite changes when more than one person works on them at the same time. The only difference is that it happens more frequently with the L8N files. The only way they'd avoid that is if they the auto-generate the game data out of a spreadsheet (thus my theory on spreadsheet driven design).

Everything I hear about where things are going wrong with this game feels like it's in practice and fundamentals of software development. You can get a company filled with the best individuals at their jobs in the world but if your processes to get them working together are flawed you're still gonna end up with a mess

You got it in one. That goes double when they are working with distributed teams for QA and Localization, where there's a 24-hour turnaround on questions.

If you go back to the first six months of patch notes and strip out everything that's a stat change (i.e., metadata), you'll see their real rate of progress. It is not pretty.

 

[Kenshiro70]

Relying on the metadata/modular design yeah it can be a crutch, but at the same time, it's also a way to shift work from developers to others. It means a dev just needs to set up the foundation, but someone else who is not a coder can edit tech costs, building yields, etc... and be able to give some basic updates or bonuses. If you want to change a +50% to a +25% you can do it easily. And obviously it's a godsend for modding and modularity.

The downside as you mention is that often the bonuses tend to be same-same. +1 to blah. And you also can suffer a little bit because the AI now has to react to whatever randomness is in the files. If you change a library from 2 science to 3 as its base yield, you need to make sure the AI will appropriately react to that change.

Although frankly, the civ series has been this way for multiple versions, so the fact that it works this way again is not a surprise. I do think every civ has some unique abilities at least, so it's not like every civ just gets a bonus to one building type and that's it. I do think the leaders probably could use a little more of a refresh for some, but I don't think they do a bad job at making sure there's some unique abilities around.

Just to be clear, I love metadata. To me the biggest issue is that QA didn't properly leverage it to automate the testing. Then the designers could have specced out the tests right alongside their designs. That would have stopped the Base Yield bug and the Modern Age Yield explosion in its tracks.

 

[Kenshiro70]

Firaxis (like most studios with an in-house engine) seem to build on previous iterations instead of starting new versions from scratch. The engine for VII will have started as the version that still exists in VI (and to VI, from V, in turn).

Chunks of it will probably have been rewritten (game engines tend to be a lot more modular and less monolithic than folks presume), but mapping the line of which game's engine it is is . . . pretty difficult. Very Ship of Theseus-like.

This might be the one exception to the rule, as they had to make it platform-agnostic (i.e., run on both consoles and PCs). Depends on if they were able to convert/reuse old code or had to rewrite it. If they had to rewrite it, dropping some of the features we were irritated about (e.g., renaming cities) suddenly makes a lot more sense, particularly if the deadlines got moved up on them. If they were able to convert/reuse old code, removing those same features would take more effort than keeping them in (because regressions).

Impossible to know for sure unless without inside knowledge.

 

[Gorbles]

This might be the one exception to the rule, as they had to make it platform-agnostic (i.e., run on both consoles and PCs). Depends on if they were able to convert/reuse old code or had to rewrite it. If they had to rewrite it, dropping some of the features we were irritated about (e.g., renaming cities) suddenly makes a lot more sense, particularly if the deadlines got moved up on them. If they were able to convert/reuse old code, removing those same features would take more effort than keeping them in (because regressions).

Impossible to know for sure unless without inside knowledge.

I don't know much about games development pipelines, so I'm welcome to corrections, but as I understand it the pipeline to build something for a specific pipeline tends to be (broadly) independent of the code.

I work with (among other things) React Native, which compiles down to native code (broadly-speaking, Java / Kotlin on Android, Objective C / Swift on iOS). Any time I'm using React Native's webview, the build step (which though I configure, I do not control myself) makes sure the resulting package / application uses the platform-specific components (e.g. WKWebView on iOS).

There may be some development paradigms to account for to ensure something works the same way regardless of the platform you run it on (we have web, Android and iOS platform switches for this exact reason), but there's no reason to assume a massive rewrite happened. Impossible to know for sure, but at the same time, no point supposing said massive rewrite either then!

 

[Krajzen]

I am kinda grim about the prospects to save civ7.

There have been many miraculous rebirths in the recent video game history, but I don't recall one happening when the main source of the game's criticism is its core premise.

Hell knows what you are supposed to do with this as a developer.

 

[SerbianYugoThief]

Tracking that back to flawed design principles, it seems like someone made an explicit decision that Civilizations are very content-heavy (building models, animations, etc.)

This is a bit of a tangent but I really wish they'd stop increasing the workload for leaders and civs. The game genuinely does not need to look any "prettier" than V and as nice as the leaders looked in V, it's clearly a bigger bottleneck than it's worth; both in terms of the workload, and in terms of who it allows them to add. Cut out the full voice acting, or stick people back into portraits, or at the very least just dial the graphical fidelity back. I would rather a mechanically coherent but visually lacking game (not in the sense that it's ugly, but in the sense there's room for improvement) like 4 over a mechanically lacking but "pretty to look at with the UI off" game like 7.

 

[GeneralZift]

This is a bit of a tangent but I really wish they'd stop increasing the workload for leaders and civs. The game genuinely does not need to look any "prettier" than V and as nice as the leaders looked in V, it's clearly a bigger bottleneck than it's worth; both in terms of the workload, and in terms of who it allows them to add. Cut out the full voice acting, or stick people back into portraits, or at the very least just dial the graphical fidelity back. I would rather a mechanically coherent but visually lacking game (not in the sense that it's ugly, but in the sense there's room for improvement) like 4 over a mechanically lacking but "pretty to look at with the UI off" game like 7.

I'll be quite frank, the only reason I played Civ as long as I did and gave it a chance was because I really enjoyed Civ5's exquisite throne room leader setup.

I played on to see how each leader's personality varied, and I was really intrigued to see how each leader had an entirely different aesthetic. You see Rashid in his Palace with his Peacocks, then you meet Montezuma in his Temple, surrounded by chanting, and you see Kamehameha on a literal beach.
Each leader oozed personality and was a whole atmosphere on their own.

Also the real voice acting is all the fun. I personally enjoy hearing my own languages used by real leaders of my own nations from history. I also enjoy listening to foreign languages and trying to pick apart familiar words. Plus I enjoy hearing ancient languages and their pronunciations.

The authentic art, music and sound is part of the gameplay which makes Civilization so great and which makes many other 4X / grand strategy into Spreadsheet simulators.

This gives the game like half of the interesting flavour. I'm sorry, but this was the peak for leaders, and it gradually got more lazy as they released more games.
The idea it could get even more lazy would completely turn me off the series.

Firaxis and TakeTwo are not broke. And they don't need to reinvent the wheel when they have 25/30 years of proven experience and built up knowledge of how a 4X game works.
So why should they compromise on the art so they can deliver a mechanically sound game? Doesn't even make sense given the designers and the artists are two completely different teams.

 

[Gorbles]

This is a bit of a tangent but I really wish they'd stop increasing the workload for leaders and civs. The game genuinely does not need to look any "prettier" than V and as nice as the leaders looked in V, it's clearly a bigger bottleneck than it's worth; both in terms of the workload, and in terms of who it allows them to add. Cut out the full voice acting, or stick people back into portraits, or at the very least just dial the graphical fidelity back. I would rather a mechanically coherent but visually lacking game (not in the sense that it's ugly, but in the sense there's room for improvement) like 4 over a mechanically lacking but "pretty to look at with the UI off" game like 7.

Your opinion would not create a popular game in 2025. Civ IV came out 20 years ago. Nowadays, graphics* matter as much as everything else being compelling does.

*a competent, well-executed artistic vision. It doesn't have to be photorealistic. But at the same time a lot of complaints about VI were because of the more stylised art direction.

 

[UWHabs]

Look at one of the Localization XML files. There's not a modified date one each string. That means they're synching at the file level. With a distributed team. Ruh Roh.

If they're sending the file to multiple translators (e.g., 2k subsidiaries) and then trying to merge them back together, there's a huge chance of regression because the file modified date is not reliable for use in conflict resolution.

I'm a little stunned because this stuff is pretty basic engineering, both adding per-tag modification dates and anticipating out-of-order data load scenarios. Even worse, there are tools that will do all this for you. It's a classic build-vs-buy mistake.

It's not just a L8N issue. It's potentially a data integrity issue. The other XML files with the game data have the same weakness and can overwrite changes when more than one person works on them at the same time. The only difference is that it happens more frequently with the L8N files. The only way they'd avoid that is if they the auto-generate the game data out of a spreadsheet (thus my theory on spreadsheet driven design).



You got it in one. That goes double when they are working with distributed teams for QA and Localization, where there's a 24-hour turnaround on questions.

If you go back to the first six months of patch notes and strip out everything that's a stat change (i.e., metadata), you'll see their real rate of progress. It is not pretty.

I don't know what they're using for localization, but I would expect in most circumstances, you farm out one language to one team, and so they would deliver the file back to you in whole. You're not generally going to send out half the Portugese lines to one firm, and the other half to another. You might work with multiple localization groups, but usually you'd have one that does all the Portugese, another than does all the German, etc..

Plus, just because files that are packaged in the game don't have other tags on them, it doesn't mean that internally there's not some other process going on. Especially since you need to coordinate and update some tags for each, odds are that internally they have a database with all the translations that they can update individually, and then as part of the process they spit out all the loc files to be included in the build.

As for the rest of it, it's really hard to say. So far in 7 I think they've been a bit more reactive than in some of the 6 cycles. When they had the monthly patches in 6, more often than not, if something broke in one patch, it couldn't get fixed until 2 months later, presumably because the delay and turnaround of identify-fix-test didn't allow that in quick enough time. We've seen them turn around patch updates for some platforms or specific issues a few days or a week after the main patch, so at least those parts of the system they have improved on.

Just to be clear, I love metadata. To me the biggest issue is that QA didn't properly leverage it to automate the testing. Then the designers could have specced out the tests right alongside their designs. That would have stopped the Base Yield bug and the Modern Age Yield explosion in its tracks.

I highly doubt something like civ is something that you could ever truly automate testing on. It's a fluid game, things change a lot. It's not like financial calculations where you know what you need to get out of it and need to make sure that doesn't change with another fix. Some things you may be able to scope out, but only at a very high level.

 

[aieeegrunt]

Your opinion would not create a popular game in 2025. Civ IV came out 20 years ago. Nowadays, graphics* matter as much as everything else being compelling does.

*a competent, well-executed artistic vision. It doesn't have to be photorealistic. But at the same time a lot of complaints about VI were because of the more stylised art direction.

One thing I will give 6 is having the best art direction in the series. It had great utility, looked good, and will age really well.

I will be able to play 6 ten years from now, where as 4 aged so awfully I can’t even look at the screen

 

[notNamed]

Also the real voice acting is all the fun. I personally enjoy hearing my own languages used by real leaders of my own nations from history. I also enjoy listening to foreign languages and trying to pick apart familiar words. Plus I enjoy hearing ancient languages and their pronunciations.

The authentic art, music and sound is part of the gameplay which makes Civilization so great and which makes many other 4X / grand strategy into Spreadsheet simulators.

This gives the game like half of the interesting flavour. I'm sorry, but this was the peak for leaders, and it gradually got more lazy as they released more games.
The idea it could get even more lazy would completely turn me off the series.

Details come with an associated cost. You cite ancient languages and I'll tell you that you're confusing Civ 5 with 6 and 7 as Civ 5 did not yet do ancient languages for older leaders. They only had Attic Greek, Latin and Akkadian (for the two Akkadian leaders). Ramses speaks Arabic. Boudicca speaks Welsh.

Money is not an issue, but you can't just hire 3000 Indians to do AAA graphics and be done with it. Even if you can afford 3000 Indian wages for however long the game is in production. You need to give out directions, you need to judge the output, you need to coordinate stuff, you need to make corrections or adjustments.
And so finally time becomes the scarce resource. AAA graphics have a shelf life. If your game takes 10 years to make, you'll need to scrap all the old assets and remake them later because they're no longer up to the modern standards. But before you're done with those, the ones that were made after them are also starting to show their age and this repeats ad infinitum.

In the end, unless you already own a perfect team of a 10 000 artisans of their craft who all speak the same language and understand each other better than their own wives, you will have to make compromises and set priorities.
And that's why Civ 5 Ramses has a very pretty background and grandiose speeches, but those speeches are in Arabic and you're building English Knights, Castles and Trebuchets.
Whereas Civ 7 Hatshepsut is with no background or grandeour, but she does actually speak Egyptian, her houses and palaces look Egyptian instead of European and her troops likewise are ancient Egyptian soldiers rather than some English Knights here and Roman Legionaries there.

I won't say it's perfect, but you can't say that one of these is the magnum opus of historical immersion and the other a Paradox MS Paint simulator.
Especially not in a art terms when you're comparing a game with each civilisation sporting unique music, units and buildings to one where basically the only thing identifying it is the leader and everything else is to some degree generic.

 

[GeneralZift]

Details come with an associated cost. You cite ancient languages and I'll tell you that you're confusing Civ 5 with 6 and 7 as Civ 5 did not yet do ancient languages for older leaders. They only had Attic Greek, Latin and Akkadian (for the two Akkadian leaders). Ramses speaks Arabic. Boudicca speaks Welsh.

Money is not an issue, but you can't just hire 3000 Indians to do AAA graphics and be done with it. Even if you can afford 3000 Indian wages for however long the game is in production. You need to give out directions, you need to judge the output, you need to coordinate stuff, you need to make corrections or adjustments.
And so finally time becomes the scarce resource. AAA graphics have a shelf life. If your game takes 10 years to make, you'll need to scrap all the old assets and remake them later because they're no longer up to the modern standards. But before you're done with those, the ones that were made after them are also starting to show their age and this repeats ad infinitum.

In the end, unless you already own a perfect team of a 10 000 artisans of their craft who all speak the same language and understand each other better than their own wives, you will have to make compromises and set priorities.
And that's why Civ 5 Ramses has a very pretty background and grandiose speeches, but those speeches are in Arabic and you're building English Knights, Castles and Trebuchets.
Whereas Civ 7 Hatshepsut is with no background or grandeour, but she does actually speak Egyptian, her houses and palaces look Egyptian instead of European and her troops likewise are ancient Egyptian soldiers rather than some English Knights here and Roman Legionaries there.

I won't say it's perfect, but you can't say that one of these is the magnum opus of historical immersion and the other a Paradox MS Paint simulator.
Especially not in a art terms when you're comparing a game with each civilisation sporting unique music, units and buildings to one where basically the only thing identifying it is the leader and everything else is to some degree generic.

I'm not comparing Civ7 and Civ5, I think the improvements are good but I want the old stuff too. Keep improving it rather than improving some parts and downgrading other parts.

The suggestion was to throw even that away and have them be literal portraits with no VA no art no nothing. What's the point? I might as well play against a cardboard box.

And about the cost. So what? It's an AAA title that costs £70 - £80. And then there's DLC on top. They can afford it.

It's not even that tough to be frank. You view the leader from one static perspective, so you don't need to model the entire throne room from each direction.
Throw in a decent throne chair model, and some wallpaper or paintings, some drapes, or maybe a garden in the background, it's not rocket science.

 

[Core Imposter]

One thing I will give 6 is having the best art direction in the series. It had great utility, looked good, and will age really well.

I will be able to play 6 ten years from now, where as 4 aged so awfully I can’t even look at the screen

You can stop dissing Civ IV. You probably think Picasso is art.