Carl on TheCivShow Podcast

[UWHabs]

Thanks for posting! I always appreciate hearing an "insider" talk about things. I work as a game dev in a small company, and definitely curious to see how the "big guys" handle things vs how we do.

 

[InsidiousMage]

So my point is, if I were FXS and get some few Reports about a Bug, but the majority of Players don't seem to bother about that and there is no to little complaint from the Community about that, then why should I put more Resources on fixing it or even decrease resources from other Projects to focus on that?

If you looked at any social media post from Firaxis over the last couple of months there are always people complaining about the state of the game on consoles, even on this forum. If a bunch of people are complaining about something but Firaxis isn't getting bug reports then something in process has gone wrong and they need to be one to fix it, not hope player's jump through a bunch of hoops.

(Some Months between the Packs: enough for the Players to enjoy them and find Bugs and leave Feedback, and for the Devs to collect data, while keeping in touch with the Players)

I can't think of a Game, Software, App...etc that doesn't rely on the Customers/Consumers to collect necessary data to keep things running....Even if FXS had 100 of Play-Testers to just test the game for Bugs (that's a potential 200 000 hrs per year), I'm sure Bugs will still arise after the game has been released.

Player feedback an important part of any game, and I'm probably lucky, but I've never played with so many problems as the NFP and some pretty basic stuff seems to have been missed. Take the Hermetic Order for example. Increasing the adjacency bonus for Ley Lines is good change based on player feedback but changing it so Ley Lines don't appear in snow terrain anymore seems like something that should have been caught in testing but wasn't and way too much of NFP feels like that. "Lack of polish" would describe way too much of the NFP to me and I get that Firaxis can only do so much with a given amount of staff but their release schedule needs to reflect that. I'm fine with Firaxis doing multi-part releases in theory but after NFP I don't trust them to do right so I'm wait and see if they ever do another one.

 

[bbbt]

If you looked at any social media post from Firaxis over the last couple of months there are always people complaining about the state of the game on consoles, even on this forum. If a bunch of people are complaining about something but Firaxis isn't getting bug reports then something in process has gone wrong and they need to be one to fix it, not hope player's jump through a bunch of hoops.

I don't think Firaxis does any of the console versions. I think all of the ports (anything but Windows basically) is Aspyr.

 

[InsidiousMage]

I don't think Firaxis does any of the console versions. I think all of the ports (anything but Windows basically) is Aspyr.

The 2K and Firaxis social media accounts always tell people to send reports using the 2K website, hence my mentioning them instead of Aspyr, who do all of the ports including the Switch and PS4 versions, which I have both played. Still, it's Firaxis's game and its ultimately their, or 2K's, responsibility for the game to play well which includes choosing a good company to port the game. Pre-NFP I would have they did a good job but not anymore as, based on my personal experience, the game has been kind of a disaster on at least PS4 starting with the Gaul/Byzantium pack. Check the Console and Switch sub-forums, people have been complaining about severe crashing and related issues for months at this point.

 

[aieeegrunt]

It does not take tens of thousands of hours of playtesting to identify some of the more basic fails of this UI

Like not displaying the effect changjng a policy card will actually have on yields

Or the trade report screen being non functional

I spotted both of those literally the first time I opened the menu

 

[Galvatron]

Yes, it does. However, why then not to arrange beta testing periods? The game got rather complex, of course 10 people can't deal with all the testing. Some 200-300 dedicated players over a couple of weeks would put in enough hours to locate the majority of the most prominent issues beforehand. Why not to adapt feedback handling teams and taking the feedback in procedures accordingly? Why do we have to wait for months for game crippling bugs to be fixed (2017 Summer patch, remember?), when those could be handled in advance with a little bit more of organization?

Everything you say is will put ands accurate. I entirely endorse these statements.

And I'd add that these hundreds of people will do it for FREE. The issue will not be getting enough people to do this work for them but sorting out the thousands of applicants for the best ones.

It stands to reason this is due to corporate fear of leaks, in all probability.

It's a video game, not a next generation nuclear warhead. There are no 'leaks' to worry about. And leaking weeks before the release date has non effect on sales in industries where it matters. Of which this is not.

Not to mention the fact that Firaxis keeps saying to send them bug reports but the amount of details and whatnot that they want from players is actually burdensome in my opinion

Yeah I comment and send feedback to everyone for every product I use and I was like 'duck that'. Literally the worst feedback apparatus I've seen ever. And like I just said:I send feedback to everyone constantly, so I've seen a lot of them.

 

[MoySauce]

Hey thanks for sharing here! There seems to be a lot of controversy over the AI, and Carl stated that he is not testing the AI or the interactions with the AI. His team tests to make sure, for example, the Feitorias are only placed in enemy territory and only beside bonus resources and never directly next to each other.

The AI is a massive beast, and with so many different difficulty options, there is only so many ways to achieve a feat like that. Yes, I'm sure they understand there are better AI systems, but their business decision was to make this as accessible as possible leading to many different difficulties.

I'd like to extend the question, if you were to make a game with at least 4 different difficulties, what specific changes would you make to the AI to make the AI dumber on easier difficulties, and harder on higher difficulties? I don't mean to challenge anyone specifically, but just something to think about when making a game for both casual users and hardcore users.

 

[InsidiousMage]

I'd like to extend the question, if you were to make a game with at least 4 different difficulties, what specific changes would you make to the AI to make the AI dumber on easier difficulties, and harder on higher difficulties?

Let the AI do things they wouldn't do otherwise. I play on console so I don't have access to the game code but people in the forum have constantly pointed out that the AI is prevented from doing certain things in the different games modes, like being restricted to one hero project every 50 turns and can't use the hero abilities. Let them use tactics or strategies that they normally wouldn't be allowed to, like nuking the player, if I remember correctly. Presumably people on higher difficulties have a deep enough understanding of the game to deal with more complex AI strategies.

 

[MoySauce]

Interesting, I didn't know that was explicitly coded in the AI. Restricting hero projects definitely seems like an intentional decision where AI was probably claiming too many heroes before the user can claim any.

 

[shaglio]

I'd like to extend the question, if you were to make a game with at least 4 different difficulties, what specific changes would you make to the AI to make the AI dumber on easier difficulties, and harder on higher difficulties? I don't mean to challenge anyone specifically, but just something to think about when making a game for both casual users and hardcore users.

The simplest way would be using % modifiers for the various yields (Science, Culture, Production, Food, Faith, etc.). So on the middle difficulty level, the AI gets 100% to those yields. But on the next lowest difficulty, they only get 80%, then 60%, and so on. Likewise, on the next difficulty, they get 120%, then 140%, and so on. It's not the most elegant method, but it's a simplistic method that is already used with Combat Strength (I believe).

A more complex way would be having 2 choices - one optimal, one suboptimal - and the various difficulties would have higher or lower % chance of selecting the optimal choice. I'm not coder, but that seems extremely cumbersome to write and will likely slow down the turn times (especially on older machines).

 

[MoySauce]

The simplest way would be using % modifiers for the various yields (Science, Culture, Production, Food, Faith, etc.). So on the middle difficulty level, the AI gets 100% to those yields. But on the next lowest difficulty, they only get 80%, then 60%, and so on. Likewise, on the next difficulty, they get 120%, then 140%, and so on. It's not the most elegant method, but it's a simplistic method that is already used with Combat Strength (I believe).

This is exactly what Firaxis did.

A more complex way would be having 2 choices - one optimal, one suboptimal - and the various difficulties would have higher or lower % chance of selecting the optimal choice. I'm not coder, but that seems extremely cumbersome to write and will likely slow down the turn times (especially on older machines).

The problem is, there are more than 2 choices being made, and it's subjective as to what is optimal and what is suboptimal. Also, does Norway value something as optimal more than Hungary would value something as optimal? How much do players value civilizations roleplaying their civ? i.e. Norway prioritizing ships and navy vs. India prioritizing Faith and eventually sometimes Nukes. Optimal in-game decision sometimes doesn't mean optimal agenda/roleplay decision. So how do you balance that as well?

All I'm trying to do, is educate that it's a big ask. With many civilization combinations, city-state combinations, natural wonder combinations, tile combinations... Looping in unique units, abilities, districts... a lot of variables to consider.

 

[shaglio]

All I'm trying to do, is educate that it's a big ask. With many civilization combinations, city-state combinations, natural wonder combinations, tile combinations... Looping in unique units, abilities, districts... a lot of variables to consider.

Oh god, I put in all that time and effort and thought, just to find out it was a rhetorical question all along :facepalm:

 

[B0RDERL1NE]

Hey everybody! A few notes.

Basic thrust of his analysis - we have 10 people in QA testing, if all we do is test and take no holidays that's 21,000 hours testing in a year.

For clarification, this is just an example. Our team size changes based on need, but between all our various testing departments, we usually have more than 10 people doing testing! That said, there are more than 50,000 people that play the game, and they obviously play it for more than one hour! I just wanted to plug in some numbers to give an idea of the scale.

...when Carl mentioned Portugal. Small Continents map is a core for balancing.

We test civs on a large variety of maps. We have to! For a civ like Portugal, we test Duel all the way up to Huge, Archipelago to Highlands. But we don't balance their core mechanics around those kinds of maps, we have to focus around a handful of core map types and sizes—Portugal wouldn't be very fun if they were balanced only for Huge or only for Duel. As I mention in the interview, we want people to be able to play games where they can purposefully make things crazy just for fun. Being able to go ham with Portugal on Huge Archipelago is by design! But it's not what they are balanced for.

Carl's either talking about big multiplayer extravaganzas in the office ... or autoplays.

I talk about a lot more testing than that! Multiplayer was only on the tail end of my list of methods we test. It is actually one of the smallest methods. The majority of testing we do is in single player, including full-length games.

As for bug reporting, we gather bug reports of all kinds from a large variety of places. The official reporting system is only one of those sources. This forum is another, but there are more. We do see your reports.

 

[TreVor_ish]

We test civs on a large variety of maps. We have to! For a civ like Portugal, we test Duel all the way up to Huge, Archipelago to Highlands. But we don't balance their core mechanics around those kinds of maps, we have to focus around a handful of core map types and sizes—Portugal wouldn't be very fun if they were balanced only for Huge or only for Duel. As I mention in the interview, we want people to be able to play games where they can purposefully make things crazy just for fun. Being able to go ham with Portugal on Huge Archipelago is by design! But it's not what they are balanced for.

I only commented about balancing because You mentioned Portugal on Small Continents. (as in You were talking about it as a reference) Going ham with Portugal on Huge Archipelago being by design is something I don't think anyone has any issue with, (it's awesome) and I'm sure You guys doing whatever You can to make the game as fun as possible.

My wondering was rather focused on general playing around the map. As Civ VI core change, compared to previous iterations, is infrastructure being placed away from City Center. This creates the feeling of it being crowded when map sizes are smaller than before. (and why we still can't put in production queue all prerequisite buildings with district, or pillaged ones for repairs)

Nonetheless I'm looking forward to April's update and Huge TSL Earth - would love if it could be even a little bigger thou, like at least 128x80. (160x100 would be a dream come true)