[NFP] Is Civ 6 doomed?

[Socrates99]

I know you are saying that, but adjudicating it to turn times, or "derisive comments about CoDs", and speculating that it was a choice and not a "mistake" (a.k.a. bad coding). My point is that with that simple example, we could disregard turn times, CoDs and choice.

Yeah, those are just speculation. It could be production traps like wonders, expensive districts or builders. Design choice.

They overemphasize the race for wonders or religion to challenge players on getting them. Best way to beat the AI is to not take the bait because if 7 AI are going for one wonder you're gaining an advantage over the 6 ai who lose.

I also hope they settle on a better way to throttle expansion than ever increasing district costs. That kills AI that will build districts it doesn't need. That really slows them down and they dont shift production on the fly like a player would.

The reason I refer to it as a choice is because it was so different in V and even vanilla VI. On release Monty and other warmongers used to build massive armies. They had to have dialed it back for some reason. Turn times on weaker machines is my best bet.

 

[civac]

Actually, one could answer, yes of course it is doomed. Consider instead, is Civ 3 doomed? I doubt if anyone plays it any more, so yes, it was doomed some time ago. Eventually Civ 6 will go the same way as everyone will be playing something else. Maybe Civ 7, we will see. I don't much fancy Old World - the idea that you can move one unit ten hexes if you don't move any other unit is not unknown to me, but I am doubtful it will work in a Civ-type game. Humankind doesn't appeal - I really don't like the idea that the Mayans suddenly turn into Vikings. Maybe I will spend some more time with CK3.

Of course, people are still playing Civ3. Fewer, obviously. I even know of a Civ3 Youtuber.

I've almost hit the 1000 hours played mark with CIV VI and I feel by this point the game doesn't have too much surprises left. It's become a bit too easy to win now, I think... well, on normal mode at least.

By contrast, with CIV IV it's the opposite problem, I find the game too hard, and even on Prince mode I'm not sure I've ever reached the Industrial era. What inevitably happens is that the two biggest Civs will ask me to go to war or stop trading with the other, which will inevitably annoy the other, and then I'm promptly crushed by the stacks of doom. The last game I tried I thought I had a large enough army assembled to repel any attack, only to be crushed by an AI army twice the size, which made me wonder: "How fast are they cranking these units out? Is the AI cheating?" And so on and so forth. So it's like with one game I win too easy and the other I can't win at all (I guess CIV V falls somewhere in-between these two extremes but I find that one a bit bland sometimes).

Civ4 Prince AI has a small production bonus over the human as well as some other advantages but it's fairly even overall. Beating a Prince AI is reasonably easy in Civ4 (as in any Civ game). My guess is that you are simply a much better Civ6 player than a Civ4 player.

(Depending on the terrain, early production is hard to come by in Civ4. To produce anything use slavery liberally in cities with a granary. The granary is vital. Flatland city require midgame tech to reach good or even mediocre production numbers. This is one of Civ4's weaknesses. Since slavery is so important (and powerful/OP) food is the key resource in Civ4, far more important than hammers/cogs.)

 

[Aristos]

On release Monty and other warmongers used to build massive armies.

True, but I remember they stayed massive armies of warriors and catapults until the end game. I even remember the infamous "AI Battle Royale" that they prepared as a showcase of their "advanced AI", with the lead (only?) AI developer there to "explain" (brag?) about said AI, only to end up laughing himself about his own AI fielding massive armies of catapults in the modern era... the chat was full of it, I remember the term GDC for Giant Death Catapult came up... I wish I could find the video of that day.

So, no, I don't think it is choice.

 

[Socrates99]

True, but I remember they stayed massive armies of warriors and catapults until the end game. I even remember the infamous "AI Battle Royale" that they prepared as a showcase of their "advanced AI", with the lead (only?) AI developer there to "explain" (brag?) about said AI, only to end up laughing himself about his own AI fielding massive armies of catapults in the modern era... the chat was full of it, I remember the term GDC for Giant Death Catapult came up... I wish I could find the video of that day.

So, no, I don't think it is choice.

Not upgrading and not building are unrelated. They're still pretty halfassed about upgrading. At some point they dialed back unit spam on purpose.

I know your default is wondering how the devs manage to get dressed in the morning, that's not mine. They tweak things here and there with purpose. I just think there's often side effects to those tweaks that a criminal lack of playtesting lets by. Playtesting is likely the first thing bean counters cut.

Seems to be the model these days. Release something buggy and let consumers pay to playtest.

 

[Aristos]

Not upgrading and not building are unrelated. They're still pretty halfassed about upgrading. At some point they dialed back unit spam on purpose.

I know your default is wondering how the devs manage to get dressed in the morning, that's not mine. They tweak things here and there with purpose. I just think there's often side effects to those tweaks that a criminal lack of playtesting lets by. Playtesting is likely the first thing bean counters cut.

Seems to be the model these days. Release something buggy and let consumers pay to playtest.

Do they get dressed in the morning? (you made me laugh with that comment!)

Yeah, you have a point there; lack of testing and the new model of beta/alpha releasing at full price is killing the industry, or at least its quality. I still think though that at least part of that "tweaking" here and there is just "bad tweaking", and not only tweaking with purpose but without testing. Why do I think that? Because some of those "bugs" or fails are just too obvious (embarrassing?) to only attribute it to lack of proper testing. But yes, proper testing would filter out all of them regardless of their complexity.

 

[UWHabs]

Not upgrading and not building are unrelated. They're still pretty halfassed about upgrading. At some point they dialed back unit spam on purpose.

I know your default is wondering how the devs manage to get dressed in the morning, that's not mine. They tweak things here and there with purpose. I just think there's often side effects to those tweaks that a criminal lack of playtesting lets by. Playtesting is likely the first thing bean counters cut.

Seems to be the model these days. Release something buggy and let consumers pay to playtest.

Their model is heavily revolved around not changing things if they don't have to. I do remember seeing one talk by a civ developer (I forgot who it was), but it was essentially "if you need to change something, double it or halve it." So if that's the mentality, then they certainly don't want to too much editing.

Personally, I'm a tinkerer. I have no problem if something is not quite right to change a bonus from 20% to 18% if that makes it better balanced. Like, by default, if I were a civ developer, I would probably hardly ever have a patch where something didn't change for every civ. But if their mantra is more "don't change things unless if they are far off", that certainly does seem like how they are approaching things. While some civs are better than others, or some game concepts are better balanced than others, could be a reason why you don't often see little tweaks here or there, and some civs have stayed exactly the same through many patches since nothing was so egregiously wrong with them that it's not worth it to them to make those tiny edits.

 

[Civinator]

Consider instead, is Civ 3 doomed? I doubt if anyone plays it any more, so yes, it was doomed some time ago.

Here you are completely wrong. Per example my C3C mod CCM 2.50 in about 1 1/2 years was downloaded at CFC more than 3.200 times.

 

[aieeegrunt]

Not upgrading and not building are unrelated. They're still pretty halfassed about upgrading. At some point they dialed back unit spam on purpose.

I know your default is wondering how the devs manage to get dressed in the morning, that's not mine. They tweak things here and there with purpose. I just think there's often side effects to those tweaks that a criminal lack of playtesting lets by. Playtesting is likely the first thing bean counters cut.

Seems to be the model these days. Release something buggy and let consumers pay to playtest.

The only thing that makes this tolerable is a healthy modding community and full mod access

 

[Linklite]

I'm wondering how much playtesting could feasibly fix. I mean, they made a screw up back in September (if I understood and remember the thread correctly) and players only recently found that the problem. How many hours of playing (testing) did that represent? The actual problem never came up in our discussions until that thread, to my memory. We have players who, individually, put thousands of hours into the game, and how many players are there here? Yet, no one talked about the crazy science focus, except obliquely and seemingly not realising it.

That isn't a criticism, I started to notice a problem, but never full on mentioned it. Noticing bugs like that is hard. Each game takes (say) ten hours. How many people would actually something wrong? You need multiple times that to get the bug noticed. Then the game needs to be replayed to ensure that it really was a bug, and to see what the effects are. Each step requiring multiple people playing 10 hour games. Then it needs to be tracked down, located and fixed. The fact that it took so long for the hundreds of Civ players (CFC and Reddit) to do it tells you that this isn't something that could be reliably left to a few people to do in a few days. And that needs to be done at least once for every update (so twelve times over for NFP).

I'm not trying to excuse Firaxis or the other companies that do this (I'm annoyed that they're charging £30 or £40 to basically playtest the musings of GalCiv4). However, at least with 4X, proper playtesting has to be really expensive and even prohibitive to do it properly. With FPS, it's probably a case of letting people run riot with it for a few hours. It'll soon become apparent if that pistol is OP or weak sauce. 4X games require full games to really see how things play out, and they're not short. As I said though, it passes me off when I'm charged to find their bugs for them, so this is not a defence. I'm just wondering what the alternative is. Playtesting doesn't seem a practical solution.

 

[MrRadar]

Each game takes (say) ten hours. How many people would actually something wrong? You need multiple times that to get the bug noticed.

You don't need to play the game yourself to notice abnormal AI behaviour, I'm sure they have ample means to watch AI only games. I remember Ed Beach telling us in one of the early videos that in the office they had screens on constant autoplay. One standard speed AI only game start to finish takes ~2 hours to observe. And when you observe it yourself and select where to look on the map, a lot of general AI behaviour becomes obvious very soon. You certainly do not need that many hours to notice abnormal behaviour. You only need to pay attention.
And if you read the replies on this reddit thread, there were quite numerous players who felt something was amiss.

 

[DizzKneeLand33]

How many hours of playing (testing) did that represent? The actual problem never came up in our discussions until that thread, to my memory. We have players who, individually, put thousands of hours into the game, and how many players are there here? Yet, no one talked about the crazy science focus, except obliquely and seemingly not realising it.

I for one stopped playing the game (according to Steam) in October of 2020 before playing again in June of this year. So, I wouldn't have seen the issues until now (personally).

 

[PiR]

I'm wondering how much playtesting could feasibly fix. I mean, they made a screw up back in September (if I understood and remember the thread correctly) and players only recently found that the problem. How many hours of playing (testing) did that represent? The actual problem never came up in our discussions until that thread, to my memory. We have players who, individually, put thousands of hours into the game, and how many players are there here? Yet, no one talked about the crazy science focus, except obliquely and seemingly not realising it.

They should have different tools than us. Not only to run the game as @MrRadar pointed, but also to monitor it, via logs, they would have key metrics that should be within certain ranges, etc. Of course it requires time to design those things to check but any big software development needs to invest quite heavily into its testing.

And that needs to be done at least once for every update (so twelve times over for NFP).

Then maybe the frequent delivery is not appropriate for a 4X game, and they should stick with expansion packs and small patches.

 

[Infixo]

@Linklite I found the issues with heavy science focus just by looking at the source code when working on an update to Real Strategy. I was reviewing changes done in each patch and looking for AI related things. That’s it. Took maybe 2 hours tops to go through all 12 patches.
In fact, those issues are easily spotted when you understand how AI works and how you „program” it in the game’s config files. What is not easily spotted are the consequences. Because Civ6 is a complex game and there are hundreds others parameters that influence how AI works and we don’t have any documentation about it. And here all the comments and observations made by others playing many games were helpful in confirming that science focus is a real issue and not just a minor blip.

 

[Aristos]

@Linklite I found the issues with heavy science focus just by looking at the source code

Please tell me you don't have the source code, or else.... My prrrecioous!! Give it to me!

I know what you meant, just wanted to make some fun.

 

[TheMarshmallowBear]

They should have different tools than us. Not only to run the game as @MrRadar pointed, but also to monitor it, via logs, they would have key metrics that should be within certain ranges, etc. Of course it requires time to design those things to check but any big software development needs to invest quite heavily into its testing.


Then maybe the frequent delivery is not appropriate for a 4X game, and they should stick with expansion packs and small patches.

This isn't appropriate for any big game in my opinion. The Sims does it too and it's in my opinion disaster, so many bugs and so many frequent issues.

1 month is not enough to balance, verify and quality assure a game.

The game was "fine" before the NFP and I think the quality of Civ 6 just dropped after it was all wrapped up.

The fact that bugs like the Monopoly Tourism Multiplier bug made it to release is astonishingly bad.

 

[DizzKneeLand33]

The game was "fine" before the NFP and I think the quality of Civ 6 just dropped after it was all wrapped up.

Yes, I play with all of the civs, but with literally *every* option from NFP turned off since they are all just exploitable by the human (or bugged in some cases). So, the value added by NFP is nil for me (again, other than the Civs).

I have purchased every item for this game in hopes that it would be better btw. Just wanted to point out that it was a lot more than 60 USD for people like me. I don't ever want to come off too negative, and I think there are things that are fixable.

A quick question: I played like 10 games in a row where the AI never would ask for a friendship deal, but when I asked they would always accept (at friendly). Did anyone else have this happen? I changed a parameter that said Declare Friendship where the setting was "AiAllowed=False" to True. Now I get asked. And, the AI has friendships with each other now. Don't know if that was a one off or if it was a systemic problem. And, of course, it's hard to tell without the behind the scenes code, only trial and error with a whole bunch of moving parts (as @Infixo said above).

 

[Laurana Kanan]

A quick question: I played like 10 games in a row where the AI never would ask for a friendship deal, but when I asked they would always accept (at friendly). Did anyone else have this happen? I changed a parameter that said Declare Friendship where the setting was "AiAllowed=False" to True. Now I get asked. And, the AI has friendships with each other now. Don't know if that was a one off or if it was a systemic problem.

Looking at my backup files going back before R&F, and all five of the StateTypes with DIPLOACTION_DECLARE_FRIENDSHIP have always had AiAllowed set to "False". There's also this included note with the Allied & Friendly states: "Note: This may look silly here, but it predisposes the AI to accepting a new declaration when the old one expires"

 

[DizzKneeLand33]

Looking at my backup files going back before R&F, and all five of the StateTypes with DIPLOACTION_DECLARE_FRIENDSHIP have always had AiAllowed set to "False". There's also this included note with the Allied & Friendly states: "Note: This may look silly here, but it predisposes the AI to accepting a new declaration when the old one expires"

Right, my October 2019 file has this as well. For some reason, though, the behavior seemed to change after one of the later patches. Perhaps in the dll? Or again, my results were just strange. I'm just glad that it works now.

By the way, through turn like 180 (220 in one instance) the AI wasn't making friends with each other either until I changed this, at least the AI that I could see. I'm just playtesting manually (no tools). After the change there was an AI friendship at like turn 40 (and one with me as well). I also raised all thresholds by 10 with regard to diplomacy, making friendly status actually harder and still got good results.

EDIT: The reason I'm glad this works is that from Civ V I have a house rule that I cannot ask for friendships -- the AI has to ask first. I don't like games to be overly amicable.

 

[PiR]

This isn't appropriate for any big game in my opinion. The Sims does it too and it's in my opinion disaster, so many bugs and so many frequent issues.

1 month is not enough to balance, verify and quality assure a game.

The problem is all companies now want to go "agile" because GAFA and successful new companies have done so. But if a website can have frequent delivery, it's not necessarily transposable to all. Decreasing the time-to-market is not always better...

 

[Laurana Kanan]

Right, my October 2019 file has this as well. For some reason, though, the behavior seemed to change after one of the later patches. Perhaps in the dll? Or again, my results were just strange.

Oct. '19 would've been after the Sept. Update post-GS - I'm talking about before both expansions. Although I don't have my 2016 backups any longer, I looked at the files after all the Deluxe DLCs were released in 2017.

I think your results may have just been strange. I've never encountered the issue you described, unless it was something that changed with the April '21 patch, since I haven't played a game since then.