Activision Boss Predicts Rising Game Development Costs

[cidbahamut]

It seems like development costs going up would be pretty self-evident.

To be fair, he said this while talking to investors (who may not understand the industry all that well). He wasn't really trying to present this as some kind of great revelation, just saying "here's how things work, and they're probably not gonna change".

I get that. I'm just sort of sitting here scratching my head trying to figure out why this is noteworthy enough to warrant an article.

 

[josemlopes]

But arent the bigger improvements of most game engines now the fact that they are a lot easier and faster to use meaning that it costs less while achieving better results?

 

[Ryan Hughes]

Yves Guillmont of Ubisoft sai they would rise to $50-60 million USD for next-gen, meaning that either a) price points will have to go up to $65-$70 USD, or b) the triple-A industry will go bankrupt.

At the 2005 Game Developers Conference, Factor 5 (makers of Rogue Squadron) Predicted that PS3 dev costs would rise from $3-6 million USD for gamecube, then $15-17 million for their title, LAIR. Instead, LAIR cost them about $25 million, and dev costs for triple-A games are now averaging $30-50 million.

Of course dev costs will rise, not because of technology, but because people like Guillmont and Kotick are idiots. The issue is whether or not they are grossly underestimating how much costs will rise.

 

[deathbydeath]

No no no no no. You are doing it fucking wrong. You have the technology, you have the potential labor force, so make games for fucking less. Rising costs are going to screw you over, as if you needed the help. Lower costs means cheaper games, which in turn means more and happier customers. It will mean less risk for you, it will reduce piracy, hell it'll probably cure cancer, just lower your goddamn development costs.

 

[Covarr]

If you look at the games everyone was talking about this year, The Walking Dead, Journey, FTL, Katawa Shoujo, XCom, (To The Moon, Super Hexagon, Thomas Was Alone), I reckon small scale games have attracted more attention than the big budget ones as far as gaming media goes.

If prices rise, maybe all but the very best games will be forced to stop competing for the top and we'll begin to see the niche games finally expanding at last

I wouldn't necessarily call the smaler-scale games niche, as the word typically suggests a more limited appeal rather than limited scale or budget. Games like Train Simulator 2013 and Cabela's Big Game Hunter, which aren't necessarily bad games, but by design target a much more specific and less general audience than, for example, Battlefield 3. Journey and The Walking Dead were certainly quite different from other games on the market, but I think they still had a good deal of mainstream appeal.

The other way around definitely is true, though; niche pretty much need to be smaller in scale and budget. Developers know going in that some of these games will likely never pass ten thousand sales, let alone millions (New Super Mario Bros. U says hi, Mr. Kotick! That game had no trouble succeeding!), and they budget and scale their games accordingly. This leads to an unfortunate situation of niche games frequently looking like shovelware, because a similarly small amount of resources put into them, even though they're often developed by people far more passionate than whoever Ubisoft puts behind Imagine gamez.

P.S. Thanks

 

[Grabehn]

So basically... "We gon' spend moh money on your gamez so we haz to sell them for moar" Yeahh... nop, fuck you ugly guy.

 

[jon_sf]

Better hardware means that games can have better graphics, higher polygon counts, richer textures, better sound, etc.

But those don't spring out of thin air. To create more detailed models, you need modelers spending more time creating and animating them. In the original Legend of Zelda, Link was [link url="http://www.nesmaps.com/maps/Zelda/sprites/ZeldaSprites.html"]15x16 pixels big[/link]. In AAA games these days, the main character will be comprised of 30,000+ polygons. They are animated doing increasingly complicated things, which takes a lot of animators, and you've got huge numbers of high resolution textures to be created by visual artists.

Then you've got all the audio. Mass Effect 3 had [link url="http://en.wikipedia.org/wiki/Mass_Effect_3#Audio"]40,000 lines of spoken dialog[/link] and hours of orchestral music. The scoring, recording, editing, re-recording, all takes a lot of time, people-power, and money. And that's not even touching all the programming that goes into the graphics & physics engine, AI, game mechanics, play testing, multiplayer development and balancing. The list goes on and on.

If the next generation of consoles comes out and the games look just like the old generation, people will be up in arms and say "why the hell did I buy this new console if everything looks & feels the same?" To make things better takes more effort. There's no magic "make game more awesome" chip that goes into the next gen of hardware.

So a company spends millions and millions of dollars to develop a game, they need to make sure they make it back. So they do a big marketing campaign with billboards, posters, TV commercials, etc. If spending N dollars on marketing & advertising didn't earn them more than N dollars in revenue, they wouldn't do it. Think of your favorite game that didn't do well commercially, and is a "hidden gem." Chances are it's a hidden gem 'cause somebody botched the marketing.


Big companies like Activision will continue to make big games that cost a fortune, and hopefully earn them a fortune. Just like the big movie houses create blockbuster movies like the Transformers movies, which were critically panned but have made [link url="http://en.wikipedia.org/wiki/Transformers_(film_series)#Box_office_performance"]over $2.6 billion worldwide[/link]. Or like John Carter, which Disney apparently [link url="http://en.wikipedia.org/wiki/John_Carter_(film)#Box_office"]lost about $100 million[/link] on. Win big, or lose big.

 

[Covarr]

Better hardware means that games can have better graphics, higher polygon counts, richer textures, better sound, etc.

But those don't spring out of thin air. To create more detailed models, you need modelers spending more time creating and animating them. In the original Legend of Zelda, Link was [link url="http://www.nesmaps.com/maps/Zelda/sprites/ZeldaSprites.html"]15x16 pixels big[/link]. In AAA games these days, the main character will be comprised of 30,000+ polygons. They are animated doing increasingly complicated things, which takes a lot of animators, and you've got huge numbers of high resolution textures to be created by visual artists.

Hi-poly models and hi-res textures don't necessarily take more money to make. In fact, one of the biggest timewasters in the industry today is in optimizing the work of overzealous artists to use FEWER polies and LOWER-res textures. Not to mention, as often as not, they need better models and textures anyway for pre-rendered cutscenes. Voice acting and orchestral music? That was an increase in costs when we started using them, but it's a continued cost for studios that already do that. The potential cost of sound design has largely hit a ceiling with the PS3 (since its biggest requirement is storage space).

I think the biggest technical cause for higher development costs is probably going to be in threading for these monster processors with so many cores. The vast majority of PC software uses a maximum of two cores, even if it would benefit from more, simply because it's so time consuming (and therefore expensive) to write adequate sync code for threading to work correctly and actually give a performance boost. But with modern architectures valuing more cores over faster cores, this problem is only going to become more prevalent.

That's not to say art or sound are going to get any cheaper. I'm sure as devs manage to squeak more performance out of the next generation of consoles, they will need to make more detailed assets. But since those already tend to be made in higher quality masters than are actually used in the game, I sincerely doubt it's going to make much of a difference.

P.S. Thanks

 

[Bad Jim]

But arent the bigger improvements of most game engines now the fact that they are a lot easier and faster to use meaning that it costs less while achieving better results?

That is true, but there is a bigger factor. Better hardware allows more detail which costs more money to produce. The 360/PS3 both have 512MB RAM, and I think the next generation will have 2GB RAM. That's four times the amount of polygons, four times bigger textures, and in general four times the amount of work required for producing art assets. Competition favours prettier games, so AAA developers who don't increase their art budgets will be overshadowed by those who do.

Of course there is a limit to how much gamers will pay and some evidence that publishers are quite close to that limit. Orwellian DRM and absurd DLC practices seem like desperate attempts to squeeze money out of gamers who aren't really willing to pay what the games cost to make. Just as with Hollywood extravaganzas, AAA games need exponentially increasing budgets to sustain the wow factor and will inevitably become financially impossible, the novelty of high detail will wear off, players will be less interested, and suits will turn to projects with more reasonable budgets.

 

[DarkSoldier84]

In other news: ice is cold, the sun is hot, water is wet, and space is really, really big.

Amazing work there, Creskin.

Dagnabbit, I was going to make that joke!

 

[FoolKiller]

Here's an idea... avoid having superstars like Kobe Bryant show up in commercials. I like basketball, and I like video games but they are two different things (NBA 2Kx excluded), but I don't want to see Kobe run around with a gun, I would rather see gameplay footage or something else.

 

[Saulkar]

Better hardware means that games can have better graphics, higher polygon counts, richer textures, better sound, etc.

But those don't spring out of thin air. To create more detailed models, you need modelers spending more time creating and animating them. In the original Legend of Zelda, Link was [link url="http://www.nesmaps.com/maps/Zelda/sprites/ZeldaSprites.html"]15x16 pixels big[/link]. In AAA games these days, the main character will be comprised of 30,000+ polygons. They are animated doing increasingly complicated things, which takes a lot of animators, and you've got huge numbers of high resolution textures to be created by visual artists.

Hi-poly models and hi-res textures don't necessarily take more money to make. In fact, one of the biggest timewasters in the industry today is in optimizing the work of overzealous artists to use FEWER polies and LOWER-res textures. Not to mention, as often as not, they need better models and textures anyway for pre-rendered cutscenes. Voice acting and orchestral music? That was an increase in costs when we started using them, but it's a continued cost for studios that already do that. The potential cost of sound design has largely hit a ceiling with the PS3 (since its biggest requirement is storage space).

I think the biggest technical cause for higher development costs is probably going to be in threading for these monster processors with so many cores. The vast majority of PC software uses a maximum of two cores, even if it would benefit from more, simply because it's so time consuming (and therefore expensive) to write adequate sync code for threading to work correctly and actually give a performance boost. But with modern architectures valuing more cores over faster cores, this problem is only going to become more prevalent.

That's not to say art or sound are going to get any cheaper. I'm sure as devs manage to squeak more performance out of the next generation of consoles, they will need to make more detailed assets. But since those already tend to be made in higher quality masters than are actually used in the game, I sincerely doubt it's going to make much of a difference.

P.S. Thanks

Those words ring pretty true. It is a significant time sink optimising 3D models for game engines where as for cinema you basically just need to create a low poly model in 3DS Max, Maya, or Soft Image, sculpt 10 million polies worth of detail in Zbrush or Mudbox then paint it, retopologise it in Topogun, then finally export the hi-res version as a displacement and texture map to apply to the new model exported out of Topogun. There done.

In game making this is compounded by having to hyper optimise the model several times over to get it to function within the game engine, especially after triangulation which is far from a perfected art. More powerful computers and game engines can get over these kinds of things, here's hoping that future engines can smooth out the process.

 

[SL33TBL1ND]

I'm fairly sure that eventually companies are going to realise that this photo-realism race is not sustainable for their profits. Just sayin'.

 

[Strazdas]

But arent the bigger improvements of most game engines now the fact that they are a lot easier and faster to use meaning that it costs less while achieving better results?

That is true, but there is a bigger factor. Better hardware allows more detail which costs more money to produce. The 360/PS3 both have 512MB RAM, and I think the next generation will have 2GB RAM. That's four times the amount of polygons, four times bigger textures, and in general four times the amount of work required for producing art assets. Competition favours prettier games, so AAA developers who don't increase their art budgets will be overshadowed by those who do.

Of course there is a limit to how much gamers will pay and some evidence that publishers are quite close to that limit. Orwellian DRM and absurd DLC practices seem like desperate attempts to squeeze money out of gamers who aren't really willing to pay what the games cost to make. Just as with Hollywood extravaganzas, AAA games need exponentially increasing budgets to sustain the wow factor and will inevitably become financially impossible, the novelty of high detail will wear off, players will be less interested, and suits will turn to projects with more reasonable budgets.

not competely true. for consoles, they create high resolution textures and then down-sample it to run on consoles. With new consoles being faster (and with current PC releases sometimes) donwsampling wont be needed, since console can already run it. automatically - higher resolution textures by actually cutting the costs since no downsampling is needed.
Of course i agree that there will be other csts added, but there are ways to cut costs too with having a better console. many of them
as for ram, i really doubt they will go with 2 GB, mre likely 4 GB. beucase with 2GB they would be using a medium-to-low PC from 2010 for a 2013 console release? that would be like releasing Xbox380 in 2009, its obsolete on arrival.

 

[Happiness Assassin]

Well, yeah. That is quite obvious. As technology progresses forward, it becomes more expensive and difficult to program a game. Back in the 80s a company could shit out a game in a couple of weeks. Now a game that doesn't spend years in development just can't compete in today's modern market.

 

[Cecilthedarkknight_234]

I should really care what some one from Activision says for what legit reason??? I can not even remember the last game they released that didn't involve map pack dlc/zombies.

 

[jon_sf]

I think the biggest technical cause for higher development costs is probably going to be in threading for these monster processors with so many cores. The vast majority of PC software uses a maximum of two cores, even if it would benefit from more, simply because it's so time consuming (and therefore expensive) to write adequate sync code for threading to work correctly and actually give a performance boost. But with modern architectures valuing more cores over faster cores, this problem is only going to become more prevalent.

Interesting point. I wonder if there's been a move to use more functional languages in game design, that are better suited to run in parallel, and can be more easily optimized. I know that graphics rendering has been something that has always been well-suited to parallelization. (Thus why companies like Pixar have massive render farms working in parallel on scenes.)

I'd imagine this is an area where it will make a difference how the APIs for the next gen consoles are designed. If developers are given tools to easily take advantage of the parallel processing and spin off worker threads, it will make a big difference in development and debugging. (Oh, the joys of debugging threading issues.)

It would be interesting to see a project plan for a major AAA game, and find out what the "long poles" are in development. I know that a lot has to happen in parallel for a game to be created, but I don't have a good sense of the total hours spent on modeling & animation vs. coding the AI logic, as 2 random examples.

Thanks for your thoughts, Covarr.

 

[Stavros Dimou]

they spend more on development because THEY GET MORE.

If anything else,they shouldn't ask for a price raise to get even more than the already more they get.

 

[Ympulse]

$5 says that this announcement was only made so that there is less of an explosion of hate when next-gen games from Activision retail at $100 a shot.

Because, you know, solid gold toilets.