Dennis Shirk acknowledges fan disappointment and Civ5 development problems

Civ5 producer Dennis Shirk wrote a post-mortem about Civ5's development process in the March 2011 issue of Game Developer magazine. The magazine isn't available for free, but the German platform 4players.de has posted a summary.

I'm a bit uncomfortable with linking to a paid publication, but re-translating parts of a freely available German summary (with proper linking to the source) shouldn't pose any problems, I think. I'll double-check with the mods to be sure. I definitely hope that it's okay to post the following info since it contains a couple of things that were new (at least for me) and that I found rather interesting. And since there has been a lot of speculation here about Civ5's development process, as well as about Dennis Shirk's role and impact (he was the producer and the source of the now-infamous "Civ5 is for the hardcore" and "Civ5 is a big sloppy kiss/love letter to the fan community" quotes, among others), there's probably some interest in the statements of a Firaxis official.

So, according to 4players.de, Shirk has the following to say:

========================================

(1) The separation of gameplay development and technology development worked very well. Firaxis new that the new engine would be usable 18 months before release at the earliest and planned accordingly, the switch to the new engine (when it was ready) worked with no problems.

(2) Thanks to the support of industry partners like nVidia or AMD, the new technology could be tested better than before.

(3) Communication between development groups (e.g. between programmers and artists) worked well because of well-thought distribution of people to their work rooms.

(4) Civ4 was the most well-rounded game in the franchise at its time, so Firaxis set itself ambitious goals for Civ5, which entailed some big and somewhat risky changes.

(5) The biggest change was the 1upt mechanic. Shirk believes that this made Civ more interesting to play, but implementing it was difficult and time-intensive, especially developing the respective AI. Firaxis underestimated the effort required for this change.

(6) Firaxis had to concentrate on the new systems in Civ5 to make sure that the game's core worked. This entailed neglecting or even cutting well-working features of the series, which disappointed the hardcore fans. However, the developers (says Shirk) crafted a basis that allows to improve/reintroduce the lacking features in the features; this is the plan for the following months.

(7) The Beta-testing of the Frankenstein team (for people who don't know about that: that's a team of external beta-testers from the community, which was assembled for Civ4, and which was also used for Civ5) worked well as long as the developers worked with the old (Civ4) game engine. However, when development switched to the new (Civ5) engine, the game's DRM was not yet decided upon, nor integrated. Firaxis couldn't (or didn't want to?) give the out test samples of the new engine without DRM, which meant that external beta-testing stopped for two months, until the DRM (Steam) had been implemented. This means that the developers worked for two months without external feedback.

(8) Firaxis entered full production with several large construction sites. The multiplayer team was understaffed for a long time, although changes in gameplay required new network code. Thanks to a lot of effort the multiplayer component was still implemented, but it lacked many functions that the players were used to. Firaxis plans to work on further improvements in this area.

(9) The game's combat system was tested mainly in the early phases of the game, partly due to the fact that test builds weren't savegame compatible and testers therefore often restarted their games before they reached the industrial age. In future development processes, Firaxis wants to put more effort into making savegames compatible between test builds.

(10) The layoffs of Firaxis members during the final weeks of beta-testing were detrimental to the moral and productivity of the team.

========================================

Please keep in mind that these are re-translations from a German summary of an English article, so there's a margin for error here. Actually I'd be glad if someone with access to the respective article could confirm or deny the above summary.

What do you think about the new info? Personally, I appreciate that an official source at Firaxis has finally at least acknowledged the disappointment among the hardcore fanbase. It's also interesting to have confirmation about the somewhat problematic beta-testing process. If DRM decisions lead to a two-month standstill of external testing during a critical phase of the game's development, then that could explain some of the glaringly obvious issues of the game. The Franky members on the forum have hinted at some rough times during development, but no one revealed (and wouldn't have been allowed to either) info about a two-month standstill. I've seen some Franky team members being criticized for the state that the game was released in ... that's a bit pointless anyway (the beta-testers have no power in the decision whether to release the game), but even more so if they didn't even have anything to test for months.

I don't buy everything Shirk claims though. Claiming that the switch between engines worked very well due to thoughtful planning seems to contradict with blaming the engine switch (and the unfinished DRM) for the lack of testing. I also disagree with the positive outlook in (6), I don't currently see the basis for reintroducing the feature missed by the hardcore fanbase ... that would require a major rewrite of the game imho. But we'll see. What do you think?

I've got nothing else to say about these issues other than what's already been written (as i understood these little hints at the time) in this thread.

Argh. I actually used forum search _and_ Google search to check whether this info was here already (I couldn't quite believe that it hadn't been mentioned here already), but I searched for "Shirk", and for some reason no one in the whole thread mentioned the author of the article even once.

Thanks for the article. Some food for thought.

I think we basically knew why the bloom came off the rose and this more or less confirms it.

A mixture of arrogance, incompetence by Firaxis and greed (by 2K Games) led to disastrous results.

Although the whiny posts are getting kind of old, there's no doubt in my mind that this could have been a really great game if it was allowed to sit in development another year.

Hmm..look on the bright side, one good expansion and CiV is set to become a great game.

The price of being stubborn.
Implementing tactical fight in a strategic game needs a special window for battles to occur, otherwise it is going to be clumsy and dissatisfying as it is.
Good example for this is Port Royale games, where fight is in a completely new window, where you maneuver your ships as you wish independently from the strategic game space.
Sid Meyer tried this in latest latest Pirates!, but I remember how clumsy it was. The units moved with such obvious difficulty and slowness, I think programming was very bad and usage of animated 3D sprites catastrophic, anyway a very bad solution.
They should go for simple and effective solutions, not for complicated layer upon layer of graphic which exhausts the resources of a platform PC.
The biggest problem of Civ IV and V was and still is memory allocation because of huge and unnecessary usage of graphic. Many people would like to play on big maps, but before you reach Renaissance, the game starts to stumble and finally crash.
These two problems should be addressed properly, or there will be no improvement to this game.
This is my two pence on what I would change in the first place.

No Hailas you couldn't be more wrong.

Battles are one of few things in this game that are better than Civ 4. AI needs to be improved and a few other mentioned things, but battles can stay because they are good as they are. It's one if the reasons why I couldn't migrate back to Stack-of-doomed Civ4.

You cannot resolve this because units are integral part of strategic game, and there is no way you can treat whole map as a battlefield. They haven't been able to develop a simplest chess playing engine this way, as we all know. Units don't even interrelate with tile properties. I haven't yet seen an archer tendentiously using hills to dominate the battlefield. Instead, they will more likely jump into the water and commit suicide.
Programmers were not doing anything in tactical combat programming. It is a fact. The approach is sterile and so are the results (expectations).

1UPT is a question of scale as much as anything. Again, trying to mix 'minor' tactics with a strategic game is a fool's errand. Maybe 'grand tactics' or operations would work. All you would have to do is allow partial stacking and BOOM....you have a fantastic system on your hands, and one that no longer allows arches to fire over the Straits of Gibraltar, etc.

(1) The separation of gameplay development and technology development worked very well. Firaxis new that the new engine would be usable 18 months before release at the earliest and planned accordingly, the switch to the new engine (when it was ready) worked with no problems.

Click to expand...

Shirk is a comedian? It looks like it. "worked with no problems" better be a mis-translation, because if it isn't it's a lie.

(2) Thanks to the support of industry partners like nVidia or AMD, the new technology could be tested better than before.

Click to expand...

Which is obviously why the game crashed a lot on release. Admittedly, AMD and nVidia can't do much about how poorly written the code for this game is. No matter how good the graphics card is, it doesn't generate extra cycles for the processor.

(3) Communication between development groups (e.g. between programmers and artists) worked well because of well-thought distribution of people to their work rooms.

Click to expand...

Apparently they don't like to communicate about the User Interface because it takes several times the number of actions to get anything done as before.

(4) Civ4 was the most well-rounded game in the franchise at its time, so Firaxis set itself ambitious goals for Civ5, which entailed some big and somewhat risky changes.

Click to expand...

Risky, yes. And to have those risks pay off meant properly implementing those changes. What came out was a game that to this day does not have all of its features working properly.

(5) The biggest change was the 1upt mechanic. Shirk believes that this made Civ more interesting to play, but implementing it was difficult and time-intensive, especially developing the respective AI. Firaxis underestimated the effort required for this change.

Click to expand...

You don't say...

(6) Firaxis had to concentrate on the new systems in Civ5 to make sure that the game's core worked. This entailed neglecting or even cutting well-working features of the series, which disappointed the hardcore fans. However, the developers (says Shirk) crafted a basis that allows to improve/reintroduce the lacking features in the features; this is the plan for the following months.

Click to expand...

User interface is a core function.

Multiplayer is a core function.

Processing cycles between turns are a core function.

Having the game work as intended (IE no or few bugs) is a core function.

This game is still lacking in core functions. When will they be addressed? Over half a year since release is a long time to go without "core functions".

(8) Firaxis entered full production with several large construction sites. The multiplayer team was understaffed for a long time, although changes in gameplay required new network code. Thanks to a lot of effort the multiplayer component was still implemented, but it lacked many functions that the players were used to. Firaxis plans to work on further improvements in this area.

Click to expand...

"Functions" alluded to in this quote:

- Having more than a couple players
- Meaningful turn timers
- Balance
- Not sending pointless information over the net
- A guard against double-moving
- Being able to move after end-turn
- Being able to even change certain settings after hosting a game
- Being able to end turn without resorting to letting the turn timer run down or force-ending turn because the game keeps prompting production in cities that have already been given orders

In other words, the "further improvement" to MP will be to deliver a working multiplayer into a game that advertises its existence. Take from that what you will.

(10) The layoffs of Firaxis members during the final weeks of beta-testing were detrimental to the moral and productivity of the team.

Click to expand...

Those layoffs sent us a message and that message is powerful. Look at *who* was laid off and make your own conclusions.

Battles are one of few things in this game that are better than Civ 4. AI needs to be improved and a few other mentioned things, but battles can stay because they are good as they are. It's one if the reasons why I couldn't migrate back to Stack-of-doomed Civ4.

Click to expand...

The problem is the amount of hexes. Units get into logjams and given the uber city powerz offensive warfare reaches joke levels that one can only circumvent because the AI is bad.

Call to Power had rudimentary tactical screen, and even that was a simple solution to stacks of doom, not to mention crowding the place with 1upt, so my workers or settlers can't get where they need to get.
The whole community has to struggle with this stubborn dictatorship of an inadequate mind which has taken the place of once ingenious creator of paradigmatic game(s) which were flexible and allowed creative approach to gameplay.

Just go to show that 2K Games really was unfair toward Firaxis to finish way too early before any realistic deadline, so that Take 2 can satisfy their investors.

Psyringe said:

I don't buy everything Shirk claims though. Claiming that the switch between engines worked very well due to thoughtful planning seems to contradict with blaming the engine switch (and the unfinished DRM) for the lack of testing. I also disagree with the positive outlook in (6), I don't currently see the basis for reintroducing the feature missed by the hardcore fanbase ... that would require a major rewrite of the game imho. But we'll see. What do you think?

Click to expand...

Interesting stuff. I wouldn't really buy it completely, either. I wouldn't really expect anyone too. He's still got a product, company and reputation to defend. Acknowledgement is the first sign (or something like that), so it's good to see this, and it's probably unreasonable to expect complete honesty, but it would be nice if the comments were a little more incisive.

TheMeInTeam said:

User interface is a core function.

Multiplayer is a core function.

Processing cycles between turns are a core function.

Having the game work as intended (IE no or few bugs) is a core function.

This game is still lacking in core functions. When will they be addressed? Over half a year since release is a long time to go without "core functions".

Click to expand...

To be fair, I doubt those are the type of 'core functions' that he is alluding to. I would take it to mean bare basics that make the game function, not necessary extras like UI and smooth out bugs that make it possible to play well.

Thoughtful Thug said:

Just go to show that 2K Games really was unfair toward Firaxis to finish way too early before any realistic deadline, so that Take 2 can satisfy their investors.

Click to expand...

That was my opinion since the release. Now perhaps others see it as true as well.

jjkrause84 said:

1UPT is a question of scale as much as anything. Again, trying to mix 'minor' tactics with a strategic game is a fool's errand. Maybe 'grand tactics' or operations would work. All you would have to do is allow partial stacking and BOOM....you have a fantastic system on your hands, and one that no longer allows arches to fire over the Straits of Gibraltar, etc.

Click to expand...

Well to be honest Civ4 wasn't much better in that respect. Paradox games are 100 times more realistic and even they could never even compare to that (they lack so much to be even considered a basic simulation game) And all civ games are and have always been a laughing stock when compared to realism.

And don't tell me Civ 4 didn't have units that took years and years of game time to travel from city to city and uncountable number of other unrealistic gameplay features you could think of.

I don't understand how can you criticize stuff in Civ5 that are incredibly simillar to Civ4 yet somehow manage to forget the AI or the diplomacy or other more important features.

Hailas said:

The price of being stubborn.
Implementing tactical fight in a strategic game needs a special window for battles to occur, otherwise it is going to be clumsy and dissatisfying as it is.
Good example for this is Port Royale games, where fight is in a completely new window, where you maneuver your ships as you wish independently from the strategic game space.
Sid Meyer tried this in latest latest Pirates!, but I remember how clumsy it was. The units moved with such obvious difficulty and slowness, I think programming was very bad and usage of animated 3D sprites catastrophic, anyway a very bad solution.
They should go for simple and effective solutions, not for complicated layer upon layer of graphic which exhausts the resources of a platform PC.
The biggest problem of Civ IV and V was and still is memory allocation because of huge and unnecessary usage of graphic. Many people would like to play on big maps, but before you reach Renaissance, the game starts to stumble and finally crash.
These two problems should be addressed properly, or there will be no improvement to this game.
This is my two pence on what I would change in the first place.

Click to expand...

Very good summation. I agree with you.

Trying to lure in the noobs and casuals with awe inspiring graphics have lead to many problems. Huge maps are barely functional if at all.

Props for Port Royale as well. I may have to give that game a try again.

Medopu said:

Well to be honest Civ4 wasn't much better in that respect. Paradox games are 100 times more realistic and even they could never even compare to that (they lack so much to be even considered a basic simulation game) And all civ games are and have always been a laughing stock when compared to realism.

And don't tell me Civ 4 didn't have units that took years and years of game time to travel from city to city and uncountable number of other unrealistic gameplay features you could think of.

I don't understand how can you criticize stuff in Civ5 that are incredibly simillar to Civ4 yet somehow manage to forget the AI or the diplomacy or other more important features.

Click to expand...

Oh, obviously every Civ has had issues like this. I just think that increasing the number of units available would make thigns feel more grand and give a better impression of having a large army, that's all.

TheMeInTeam said:

Shirk is a comedian? It looks like it. "worked with no problems" better be a mis-translation, because if it isn't it's a lie.

Click to expand...

He meant worked with no problems in a technical sense, which it did. The only reason there were any issues was because Firaxis decided to prioritize DRM over making a quality product.

And don't tell me Civ 4 didn't have units that took years and years of game time to travel from city to city and uncountable number of other unrealistic gameplay features you could think of.

I don't understand how can you criticize stuff in Civ5 that are incredibly simillar to Civ4 yet somehow manage to forget the AI or the diplomacy or other more important features

Click to expand...

.

That's not even close to similar. For one, years are purely cosmetic in civ. The DO NOT IMPACT THE GAMEPLAY. You can (and should) not think about them with respect to gameplay. On the other hand, map scale very much affects gameplay.

1upt debate (which has plenty of other threads) aside, I think this fits with everything we've known or been able to infer.

The one wish I have is that the reason for the layoffs would be made public. It certainly hurt them with multiplayer, since they had to send out hiring notices post game, indicating that they need more programmers. I remember reading a rumor that they literally had no multiplayer programmers immediately upon release. I've speculated the reason was simply budget problems. Firaxis isn't a large studio, so they need to keep lean and competitive, especially since they only have essentially one viable franchise. But I talked to a friend who knows someone at Firaxis and his belief was they simply wanted to divide the bonus money based on sales among a smaller group of people (which isn't necessarily inconsistent, either way, the underlying goal is to cut down on wage expenses).

Aside from that, deadlines (which were necessary either way) seems to have been the biggest problem. I definitely see a difficult to surmount problem being pushed aside with a "that'll have to do," which leads to unfortunate flaws in the game. I also think the failure to release the new engine to the Frankenstein test group is inexcusable. They already rely heavily on them, they should have trusted them more with their code.