Mark Darrah (Former Bioware) on why AAA games cost so much

[Shifty Geezer]

This game was made by a single person a less than a month apparently. So I believe that gaming is just too inflated.

It's a load of bought assets and a super basic gameplay loop. It proves nothing about nothing. All this shows is the entry-level creation, what's possible by indies, is a much higher base standard than yesteryear thanks to the availability of such assets and engines that can piece them together and make them look good with no effort.

So I believe that gaming is just too inflated. While I like the last of us, the story is not that good.

You may not like it, but it's widely well received and sold a lot of units, and done great on TV too where story is everything and gameplay nothing. Objectively, the investment in TLoU's story paid off and you are the outlier.

Whereas the game could be made with unreal assets in a few months

This is an utterly clueless statement. Sure, all these games that take forever to get made can be cobbled together in a few months on Unreal... That's why in the past 3 years we've had over 40 AAA Unreal Engine 5 games that blow everything out the water. Not even one UE5 game has overrun release schedule because you can bang out a world-class title in 6 months or less, it's so cheap and easy to use!!

and be nearly indistinguishable from the end result, minus the cutscenes

So lose the cinematic part of Sony's cinematic story. Or use UE4's automated facial tech from 2010-2013 instead of motion capture and copy/pasted animations.

 

[XboxKING]

To create a realistic looking bus takes a few people a few weeks work. Most people would not even notice nor care if the bus in the last of us is the same bus as in resident evil 3, a bus is a bus. Yet developers will waste essentially tens of thousands of dollars on a single asset..

Now multiply that by thousands of assets and you know why the costs are so much right now.

Also get some local drama/theatre school actors who will probably do as good a job as the more expensive ones. As for the music, basically same thing. Publishers pay for the name and that they can have the big names do a press tour, but the results can easily be achieved with a fraction of the cost. If anything, I’d like to have more varied actors and voice actors

 

[XboxKING]

Sure, all these games that take forever to get made can be cobbled together in a few months on Unreal... That's why in the past 3 years we've had over 40 AAA Unreal Engine 5 games that blow everything out the water. Not even one UE5 game has overrun release schedule because you can bang out a world-class title in 6 months or less, it's so cheap and easy to use!!

Case in point:

This is also made by one or two people.

Imagine what 50 people could do?

 

[Cyan]

another interesting take from this female developer -I mentioned her in the Concord thread, her videos are usually right to the point- on why are we living in a gaming bubble.

 

[techuse]

A lot of the problems and budget inflation developers face are a direct result of their own poor decision making. The game design aspect of game development is not that hard and it's usually fairly obvious what the majority of gamers want. You could rightly say it's quite difficult to come up with a revolutionary idea, but it is not hard to avoid making completely boneheaded decisions.

Titles like TLOU are not at all what most developers are doing so the reasons for the budgetary explosion of specific Sony studios are not at all applicable to the majority of developers.

 

[Dominik D]

Nobody can rationally explain why Spiderman 2 cost $300m to make. When it's really just a minor evolution of Spiderman 1, which was only a fraction of that cost.

An average team size of 250 people over 5 years is around $190kk. Add infrastructure cost to that (50% is a sensible ballpark) and you arrive at a value not that far from $300kk. And why do you need an average of 250 people on a project like this? Because the time required to make content, integrate it in the game, and test it is massive these days. Spidey 2 has completely redone crowds (models and animations) and traffic simulation, all of the cutscenes had to be done from scratch (obviously), there are significant improvements to lighting and materials that required pass on all of the reused content. On top of that all of the characters were at least retouched.

All of those things add up and when you update this many things you need to align results with multiple stakeholders: producers, art directors, designers,... The more things you touch between releases the more people you need to make sure everything fits together. And if you're doing tech and content in parallel, there's going to be some waste, there's no way around it. I don't know where this $300kk comes from but it doesn't sound unreasonable. Whether all of the improvements made were needed/worth it is another issue entirely. But what's there would cost hundreds of millions.

The game design aspect of game development is not that hard and it's usually fairly obvious what the majority of gamers want.

That's a myopic view of someone who's never designed a game. If game design was obvious, all of the games would be pleasantly playable and engaging (at least to the intended audience). But that's clearly not the case and never has been.

 

[pcchen]

If game design was obvious, all of the games would be pleasantly playable and engaging (at least to the intended audience). But that's clearly not the case and never has been.

Yes, unfortunately this applies to practically all products. To use a simpler example, the basic idea of an e-commence website is very simple. It's just showing products and let users to pay for them. However, when you go into the details, it's very different in each case and it's the many tiny details that cost the most of the time and money. As they said, "the devil is in the details."

 

[iroboto]

Everyone should always look at the credits scroll for over an hour when a game is finished. That’s how much the game costs.

 

[techuse]

An average team size of 250 people over 5 years is around $190kk. Add infrastructure cost to that (50% is a sensible ballpark) and you arrive at a value not that far from $300kk. And why do you need an average of 250 people on a project like this? Because the time required to make content, integrate it in the game, and test it is massive these days. Spidey 2 has completely redone crowds (models and animations) and traffic simulation, all of the cutscenes had to be done from scratch (obviously), there are significant improvements to lighting and materials that required pass on all of the reused content. On top of that all of the characters were at least retouched.

All of those things add up and when you update this many things you need to align results with multiple stakeholders: producers, art directors, designers,... The more things you touch between releases the more people you need to make sure everything fits together. And if you're doing tech and content in parallel, there's going to be some waste, there's no way around it. I don't know where this $300kk comes from but it doesn't sound unreasonable. Whether all of the improvements made were needed/worth it is another issue entirely. But what's there would cost hundreds of millions.



That's a myopic view of someone who's never designed a game. If game design was obvious, all of the games would be pleasantly playable and engaging (at least to the intended audience). But that's clearly not the case and never has been.

100 or 150 million of the cost was the license to Disney.

Game execution is hard, but design choices are fairly easy, at least in so far as knowing what not to do. It’s rare, if it happens at all, where a big game comes out and flops and the reasons why aren’t known within an hour of playing the game. Often just from the first gameplay trailer it’s easy to predict a flop. Developers are bricking the easiest, lay up decisions. It’s completely unnecessary too. In a world where there are tons of youtube game “influencers” for any given genre, you have so much potential for completely free consultation and early game testing.

 

[Dominik D]

Game execution is hard, but design choices are fairly easy, at least in so far as knowing what not to do.

No they aren't. The larger the project the more stakeholders you have with competing priorities. This results in a lot of risky decisions that often lead to suboptimal or even catastrophic results. Even when everyone knows that (not) to do in principle, if there is more than one potential direction and people with leverage push for different things, cracks will emerge. Have you ever worked on a game? How big was the team and what was the financing structure?

 

[techuse]

No they aren't. The larger the project the more stakeholders you have with competing priorities. This results in a lot of risky decisions that often lead to suboptimal or even catastrophic results. Even when everyone knows that (not) to do in principle, if there is more than one potential direction and people with leverage push for different things, cracks will emerge. Have you ever worked on a game? How big was the team and what was the financing structure?

You are bringing in personnel and ego/power/feelings issues. While those can result in poor design choices, it doesn't make those choices any less obviously bad. Having worked on a game isn't a requirement to judge game design. The games are made specifically for us.

 

[Dominik D]

While those can result in poor design choices, it doesn't make those choices any less obviously bad.

Like I said before, this is a myopic view from the outside. It's not even universal: things that you'd find obviously correct or wrong may very well be the exact opposite for a significant portion of the audience. There's no "us" that games are made for, you're conflating your personal preferences with non-existing consensus on what's correct design. Bioshock is considered great game by most people I know (and reviews reflect that) but in my mind it has significant design issues that cause Bioshock to bore me to death almost instantly. It would be extremely vain of me to state that these are obvious characteristics of the game and expect everyone to agree. Ask gamers if they want more pixels or prettier pixels, I'm confident there's no obvious answer here.

 

[Shifty Geezer]

You are bringing in personnel and ego/power/feelings issues. While those can result in poor design choices, it doesn't make those choices any less obviously bad. Having worked on a game isn't a requirement to judge game design. The games are made specifically for us.

I think Dominik's earlier point kinda concludes the argument effectively - if game design was easy, surely every game would be good to play? Why are so many games riddled with weaker gameplay?

But I guess your point is more about the cost of game design. It may not be easier, but it doesn't require a team of highly paid engineers several years to make it, so good game design shouldn't add to the cost of a project. And the counterpoint to Dominik's argument there, I guess, is the number of tiny indies with no budget and incredibly game design. Good game design doesn't take a lot of resources and shouldn't add to the cost of a game overall as it's obtained without a budget.

But then I guess the follow up from that is the bigger the project, the more game design starts to cost as iterative changes to facilitate different mechanics can have wide-spreading repercussions. The studio could have worked hard, got so far with a project that looks great, but it feels a little stale. They come up with some changes that should improve the gameplay, and then proceed with complex system rewrites and art changes to make that happen.

That's where a perfectly solid prototype should be used, surely? Start with a small team and work on the game long before you have the art and story etc. Once you know exactly how the game will feel and what your requirements are, go ahead and bake it into the final project. Maybe that's just too hard due to the inherently iterative nature of game development and idea evolution?

 

[Dominik D]

And the counterpoint to Dominik's argument there, I guess, is the number of tiny indies with no budget and incredibly game design. Good game design doesn't take a lot of resources and shouldn't add to the cost of a game overall as it's obtained without a budget.

I think this is fairly well illustrated by Tyroller's "Design Is a Search Algorithm" video on YT. The way gameplay tends to emerge in successful indie games is through ruthless experimentation and tight feedback loops. If you're disciplined and you work alone or in a small team, this means you can test hundreds of gameplay ideas in a short span of time and further evolve only the most salient ones. But this requires a person or a tight group of people with a very broad set of skills, something you typically are not hiring for in a large organization. This is also not how large organization tend to experiment (if they do at all).

There's also survivor bias in place: there are around 50 games released on Steam each day. That's almost 20k games per year (and it's been steadily growing for the past few years). How many of those are incredibly playable, a hundred? And even for those that are successful, sometimes it's a matter of luck that a word about the game spreads (e.g. Vampire Survivors) or it takes a lot of hard work to get the game out (Signalis took 8 years to make).

But then I guess the follow up from that is the bigger the project, the more game design starts to cost as iterative changes to facilitate different mechanics can have wide-spreading repercussions. The studio could have worked hard, got so far with a project that looks great, but it feels a little stale. They come up with some changes that should improve the gameplay, and then proceed with complex system rewrites and art changes to make that happen.

There's another factor that needs to be considered: even if you descope you discovery/preproduction in a large studio to match what indies (sometimes) do, what do you do with all of those people who don't have to be involved at this stage? This is a constant problem for larger studios. If you care about retention and don't want to fire people who are on the bench between projects you incur a cost that has to be accounted for somehow. One thing you can do is include bench cost in your next game cost. But that doesn't solve a real problem of having nothing to do. It's fine when you're between projects for a month or two but pre-production can take a year and your benched people won't be happy having nothing to do.

What tends to happen is you put some of your content folks on DLC/expansion while your design core works on prototyping and tech is split between bugs, QoL for DLC and prototyping for the next game. But in order for large expansion to be financially viable you have to have money. Publishers often won't commit to that cost upfront so studio is probably financing it themselves or gambling for this deal to happen (often publisher won't give DLC a green light before the release and initial reviews of the game; but even then they can hire another cheaper studio to do the DLC). If you're already successful, you've built a financial buffer for this mode of operation, if you were able to negotiate just the right contract with the publisher, executives aren't greedy, and you're not publicly traded (a lot of ifs) then this will work, maybe.

Conversely, if you permanently staff only the core you need to prototype and later lead the project, what do you do when you need to scale for production? Hiring is expensive and error-prone, finding the right co-dev partner isn't easy (although companies are so desperate for work right now it's probably easier than ever) but even when you do, if you need to scale from 20 people to 500 for the AAA production, there are going to be layers of decision-making that aren't the people intimately knowledgeable with the intended design.

You can see how things are aligned just right for Insomniac. They did Miles Morales between Spidey 1 and 2 plus R&C Rift Apart and a number of smaller VR games. They are set up for two AAA titles in parallel (but seemingly with pre-production of one overlaping with production of another) and have sufficiently good BizDev strategy to push a smaller VR game every year or two. You can always pull that "satellite" team into any of your two core titles instead of having them work on a VR game. This shit is hard and many studios seemed to be perfectly aligned with this mode of operation before one or two disasters struck and they vanished, often not without a fight.

 

[techuse]

Like I said before, this is a myopic view from the outside. It's not even universal: things that you'd find obviously correct or wrong may very well be the exact opposite for a significant portion of the audience. There's no "us" that games are made for, you're conflating your personal preferences with non-existing consensus on what's correct design. Bioshock is considered great game by most people I know (and reviews reflect that) but in my mind it has significant design issues that cause Bioshock to bore me to death almost instantly. It would be extremely vain of me to state that these are obvious characteristics of the game and expect everyone to agree. Ask gamers if they want more pixels or prettier pixels, I'm confident there's no obvious answer here.

Would you agree that you are in a small minority who found Bioshock’s game design to be problematic? There tends to be huge amounts of overlap in what gamers enjoy. That is why the live service market is so hard to penetrate. There isn’t this vast array of people who just have wildly different ideas of good and bad game design. Most people tend to fall under a similar umbrella.

Suicide squad is a great example. So many absolutely baffling decisions I cant even fathom how they were decided upon. Why not have the Justice League itself fight a corrupted Superman? That makes so much more sense from a plausibility standpoint. The characters are more interesting and liked. It allows for actual gameplay mechanics that are unique to the characters rather than having to pigeon hole them all into bland and mediocre shooter mechanics with the only differentiation being some traversal differences that end up bring mostly visual. It’s just one of so many easy wins that would have dramatically increased the game’s chances of success.

Helldivers 2 patch history is yet another example of a string of absolutely braindead decisions that defy belief. You could pluck any seasoned player from the community to consult with and end up with a vastly better outcome.

Remember the mess starting with post launch BFV leading into the BF2042 debacle that caused David Sirland to quit Dice? Many easy lay ups just completely bricked due to braindead decision making.

@Shifty Geezer games would still fail because execution is hard. As to why bad game design has become so common, I attribute it primarily to people being unqualified for their roles. Talent in any field is hard to come by. With more and more people developing games, it’s inevitable that a smaller and smaller percentile will be capable of quality output. It’s made even worse in today’s culture as outside forces make the water even murkier.

 

[Dominik D]

There tends to be huge amounts of overlap in what gamers enjoy. (...) Most people tend to fall under a similar umbrella.

This is verifiably not true, not within a single genre, not even within a single franchise. By the time gamers see changes that, say, Destiny brings with expansion N, expansion N+1 is already in production and N+2 is at the final stages of planning. So things may be "obviously" bad for people playing expansion N daily. But 1) things in production already depend on these "bad" decisions, 2) these decisions potentially lead to something important that end user simply doesn't see/understand yet, 3) large enough games (especially service ones) have more than one target group with conflicting "correct" decisions for them. Plus: 4) even if you correctly identify problem, chances are your solution wouldn't work but you don't have enough context to know that.

1) Let's say you're right that certain things are obvious. Unless you can address how production that takes gamers' feedback is supposed to be structured (in terms of work organization but also financing) things being obvious doesn't mean anything in real life.

2) You don't know what you don't know. Sometimes you need to jinx something small to build something greater on top of it. Gamers have little to no insight into the process behind the scenes so obvious conclusions are very often wrong with the benefit of hindsight and more data. Problem is: even if you get all of the background two or five years down the line, how likely are you to recall your bad, "obvious" takes and course-correct? Chances are by that time you've moved on and solidified your previous opinions about the title as facts. We all do that.

Alternatively you could have an extremely public development process where you essentially collaborate with the community on the game. There are very few successful examples of that (BG3, Warframe) and many more that ended up a joke/scam/vapourware (Em-8er, Star Citizen). And even if you're successful this is either built on pre-existing community and cash reserves (BG3) or, essentially, luck (Warframe).

3) This killed many games. The moment your GaaS is successful, you have at least two different groups of customers: hardcore metagamers and newbs. Their needs are drastically different and you can't have either without the other group. This is bound to generate different obvious solutions to the same set of problems.

4) This last point is a massive problem with gaming discourse these days. "Everyone's a designer" was a fun joke 20 years ago ("I've got a brilliant idea: GTA but in space!" - every armchair designer in early aughts) but it's an absolutely toxic undercurrent in gaming circles today. Fans of a franchise will write confidently about obvious fixes a game needs for success or broader appeal. I mean, I get it, it's fun to think you know better than a bunch of devs who's done this stuff before, makes you feel special (and I don't mean - here or elsewhere - you specifically, it's a collective you). "Hire fans" and all that BS. Given how you can make a living talking online confidently about the things you don't know it is a real poison.

Single player games are not exempt from these problems. Building single player AAA experience takes money and time. And time is crucial here: you're designing game two to five years before it hits the shelves so you're essentially guessing what trends will be in the future. This often leads to bad outcomes through two separate paths. One: your financing structure forces you to make concessions about the design based on, say, publisher's likes and current trends. If trends change, you're effed. Two: you're guesstimating what will be hot in 3 years or so based on your experience. This may work out, or not. That's why investing in franchises is so important these days: you basically shape perception of what your IP is or is not and can take safe risks with the changes you're making.

 

[iroboto]

Design is probably the primary reason why games fail. Yet everyone talks about it like they know the answer to make a game great.

It’s statistically impossible that it’s so easy and yet so many games run into design issues that need iteration and testing. There are hundreds of designers working on their own specialized part of the game as well, tweaking the look and feel.

Pre-production is where most games designs are built and tested prior to moving to full production. When they move into full production it may find that their game is no longer innovative and offers the space anything new, and that’s when development hell creeps in because they are now redesigning aspects of the game during full production.

Design is by far the hardest aspect of game building if you are designing for success and not just trying to be another title. If all you want to be is a reskinned copycat; sure design is straight forward. Just lift and go.

If you’re trying to be the game designer for Halo, where the new generation doesn’t care about Master Chief and need an introduction to the series and the old generation won’t let you change the game, good luck.

People talk about catering to your customer base, what happens when your customer base is the one that is in decline, they are the ones that stop playing games, they are the ones that are fatigued by your franchise, and yet they are the ones that buy and criticize your game the most. And you need to capture the young audience, because that’s who you initially captured 20 years prior.

 

[Cappuccino]

This is verifiably not true, not within a single genre, not even within a single franchise.

It's interesting you say this, as to me games are incredibly homogenous now. Shooters for example, they are all descendants of Cod4 and all play the same (save for an outlier or two like Counter-Strike, but that's a unique case where a pre-Cod4 game was preserved by its PC identity, console shooters are all heavily influenced by Cod4). Playing pre-2007 shooters is interesting because they all play so much different from each other, that is mostly lost today.

Third person shooters all play like Uncharted/TLOU now. Third person action games all play similar now (ofc theres unique quirks like the slinging in Spiderman or whatever but the combat is basically the same, string moves together a-la Devil May Cry or Soulsborn).

 

[Dominik D]

It's interesting you say this, as to me games are incredibly homogenous now. Shooters for example, they are all descendants of Cod4 and all play the same

I was responding to a claim that there's a massive overlap of what gamers enjoy saying that that's not the case. So which way is it: virtually all gamers enjoy the same stuff and developers refuse to listen or gamers enjoy the same stuff hence experience is very uniform? I think you have to hash that one out with techuse.

On a serious note I think I addressed the reason for homogeneity (perceived, I don't really agree with this sentiment) above: some stakeholders, typically those with money, tend to gravitate towards "safer" bets and those are typically "stuff that sells right now". I've seen projects where developers disagreed. You won't play those games because in every case development got canceled by the publisher. Back in the day when everything was supposed to be WoW killer or Halo killer these decisions were often up to developers and we could easily blame them for that. But playing field has changed massively in the past 10-ish years. If you see a bland AAA game chances are developers knew how things would end but had little control over it.

 

[see colon]

You are bringing in personnel and ego/power/feelings issues. While those can result in poor design choices, it doesn't make those choices any less obviously bad. Having worked on a game isn't a requirement to judge game design. The games are made specifically for us.

Should we no judge a game if it isn't intended for the "us" we belong to? Looking back to earlier this year, when Hellblade 2 released, I described that game as a 10 out of 10 in a genre that 5 out of 10 people like, mostly in reference to it's varied review scores. If we use that metric, should Hellblade 2 only be reviewed as a 10 because because that's what it is to the "us" that enjoys it, or should the 4s and below be considered even though the game wasn't meant for them? What about a game like Monster High: Skultimate Secret? That's not a game I would think is made for me. Could I possibly grade it's design objectively?

I think I might also take issue with the general statement that "games are made specifically for us". Games are art, and sometimes, often even, art isn't made for the consumption, they are made for the artist.

Design is probably the primary reason why games fail. Yet everyone talks about it like they know the answer to make a game great.

I don't know, a game's ability to find it's audience is often a larger factor. Sometimes there are design issues, but often budgets, marketing, and other external factors could tank a game. Concord was, by most accounts, a fine game. It's presentation, it price, and it's contemporaries in the genre are probably the factors that killed it. It's publishers readiness to make it go away can't be overlooked, either.