KZ2 and game budgeting in general *spin-off

[Laa-Yosh]

As for the reactions to the opening post, I find it kinda hard to believe that a game would make $43 per copy, seeing as it's average retail price seems to be around $55. But it may be possible, in which case I have to correct my own statement to KZ2 making about 50% of its budget back with that 1 million copies sold.

Mind you, I haven't considered it being a first party release, I've only approached it from the comments made by Capcom and others earlier this generation that an average AA game needs to sell at least a million copies to break even. I've then simply multiplied that 1 million by how much more KZ2 must have cost compared to the average AA game.


The other approach is to see how previous first party big names have fared:
- Lair was an obvious flop and simply drove Factor5 bankrupt. Nothing left to discuss here.
- Heavenly Sword has sold more than a million copies and still we didn't get a second game, so it has probably not made a significant profit.

So, if KZ2's budget is two times as big, it'll need to sell ~2.5 million to have the same return of investment.

 

[Laa-Yosh]

I'm no programer, but here's an ironic question to all who think that tech is easy to share between completely different engines... Why do UE3 projects need programmers at all, if there already is a complete and functioning engine with all the tools? It'd only take level designers and artists to build a new game, right?

 

[JardeL]

50 million budget ?.. 200 people ?.. Some of you guys really like to speculate when it comes to Sony ...

Let me give you some info ;

According to a 2006 article in the Dutch newspaper de Volksrant, Killzone 2 is the most expensive entertainment project in the Netherlands' history, with a budget of around €16 million--approximately $21 million at today's exchange rates.

Yes , 2006 article but it doesnt say " 16 million euros spent so far " ... So 50 million seems pure bullshit to me ... Even if it is so , Sony got valuable tech , for all of its first-party devs' use , by KZ2 development ... I said it before , KZ2 is not just a game , it is like an R&D project for PS3 development tools ...

About KZ2 dev team , from GDC 09 GG slides ;

Team:
- Peak 140 in Amsterdam, 50 at other Sony studios
- Average around 120

I think that 50 people are TEAM ICE members ... And developing tech for PS3 is their job [ ND tech team ] , helping GG is not an extra job , am i right ?..

 

[tha_con]

I'm no programer, but here's an ironic question to all who think that tech is easy to share between completely different engines... Why do UE3 projects need programmers at all, if there already is a complete and functioning engine with all the tools? It'd only take level designers and artists to build a new game, right?

I understand you're taking a jab at my post, but it's completely misguided and extreme.

Techniques and technology can be shared across engines, everything is always applicable. You can take something someone else learned, adapt it, and create something of your own. The work involved doing it THIS way, is considerably less than creating tech completely from scratch.

Or would you disagree?

 

[Laa-Yosh]

Yes, I disagree. We're not developing anything as advanced as a game renderer or physics engine and still we usually don't even try to adopt to existing code. We have our own shaders, project management system, lots of Maya plugins, and a gazillion of scripts, that are usually developed from scratch.

We do use ideas from others, but it's almost always better to treat it as an idea and not as source code or blueprint. And even in that rare case it's still more usual to write it from scratch so that you can understand what's going on and why.

And we're still a relatively small studio with only two fulltime developers and 3-4 other guys doing serious coding work. I can only imagine how much more complicated it can get with 30-50 developers, and an entire game...

 

[Fafalada]

Lair: go read Factor5's GDC presentation on next gen production, they've been very keen to idnicate that Lair was a $20million+ project.

I wouldn't pay much attention to what is indicated in these kind of presentations for 1st gen projects.
Eg., last gen, TwoTowers had a considerably bigger budget then ReturnOfTheKing, despite the team size and scale being much bigger on the latter.
F5 presentation had a lot of emphasis about cost sharing between projects (though I suppose things didn't work out that way in the end).

EA does the same thing, they have massive teams but the same people will work on multiple projects over the course of a year, but they still get credited on all games resulting in team sizes that appear astronomically large.

While there are certainly some shared teams - you are just as overly optimistic here as you accuse others of being regarding Sony tech sharing.

 

[tha_con]

Yes, I disagree. We're not developing anything as advanced as a game renderer or physics engine and still we usually don't even try to adopt to existing code. We have our own shaders, project management system, lots of Maya plugins, and a gazillion of scripts, that are usually developed from scratch.

We do use ideas from others, but it's almost always better to treat it as an idea and not as source code or blueprint. And even in that rare case it's still more usual to write it from scratch so that you can understand what's going on and why.

And we're still a relatively small studio with only two fulltime developers and 3-4 other guys doing serious coding work. I can only imagine how much more complicated it can get with 30-50 developers, and an entire game...

So essentially what you're saying is your opinion is formulated based on your own experiences and assumptions?

Because essentially you've only said that you've created your own technology, and you can't adapt it to other technology (or the other way around). What you haven't said is what could be possible if you didn't have to build your own shaders, scripts, plugins, etc, from scratch. If you could start with a jump off point, and modify, or build your engine to use mutliple tools.

In fact, you've really said nothing aside from your own experiences of building your own tech from scratch in house w/no outside help...which is completely alien to the situations of Naughty Dog, Evolution Studios, Guerrilla Games, Insomniac Games, Sony Cambridge, etc. who all share technology and experiences to better build their own games and adapt it to their technology.

You're right though, in that you can online imagine what it would be like with more developers with more ideas, more talent, and more view points. I'd assume it would be considerably easier to learn, develop, and implement.

Why is everything always "black or white" with folks? I think some of you would make fantastic journo's.

 

[obonicus]

As for the reactions to the opening post, I find it kinda hard to believe that a game would make $43 per copy, seeing as it's average retail price seems to be around $55. But it may be possible, in which case I have to correct my own statement to KZ2 making about 50% of its budget back with that 1 million copies sold.

Mind you, I haven't considered it being a first party release, I've only approached it from the comments made by Capcom and others earlier this generation that an average AA game needs to sell at least a million copies to break even. I've then simply multiplied that 1 million by how much more KZ2 must have cost compared to the average AA game.

But those 'has to sell x to break even' numbers are generally wrong. We've heard that Wii games need to sell ridiculous amounts to break even and that doesn't even make any sense, not unless Nintendo licensing fees are equally ridiculous!

The other approach is to see how previous first party big names have fared:
- Lair was an obvious flop and simply drove Factor5 bankrupt. Nothing left to discuss here.

This is a reflection of what joker was saying, which makes perfect sense. Sony may have lost money, we don't know how much, but they cut funding to Factor 5 and let it sink.

- Heavenly Sword has sold more than a million copies and still we didn't get a second game, so it has probably not made a significant profit.

The problem with this one is that there's like 3 or 4 different stories. I've heard stories that range from Sony offering HS2 to Ninja Theory and them refusing it to Sony saying that no, there will be no sequel. Sony has no IP shortage and it's not like HS is such a strong IP that farming it out to different studios would make that much sense, if Ninja Theory wasn't interested in the sequel.

 

[Laa-Yosh]

The similarity in all of these situations is that tech does not exist in a vacuum. You have to make sure you can work with it as well. It has to fit the individual requirements of your own project(s), it has to be transparent, easily modifiable, and so on. Doesn't matter if it's the way your shader outputs render layers to the compositing app or the character animation engine, it is still a common software engineering issue.

One of our greatest troublemakers here is a software called Zbrush, which is developed by little more then a single guy - and thus you don't have a clue about what's going on under the hood, there's no definitive documentation, no explanations. You have to experiment with it, make assumptions, try to learn from unexplainable errors and crashes and so on. It's been the pioneering app for today's normal/displacement mapping workflow, but it really is a b**** to work with.

And with more developers, it gets exponentially more complicated.

Do you have any software development related experience, by the way? Just because Joker here has a few shipped current gen games under his belt, and some relationship with the Sony dev enviroment, so I wonder what your reasons are to challenge his opinion.

 

[joker454]

While there are certainly some shared teams - you are just as overly optimistic here as you accuse others of being regarding Sony tech sharing.

I heard it from some friends that work at EA LA. It's a bit of an assembly line situation, where someone works on game X for a few months, then gets moved to game Y for a few months, etc. The entire company doesn't operate that way, but it does happen, which accounts for some of their massive credit lists. Likewise, if say an engine programmer works for one month on a engine feature, and that engine gets used on three games, then that programmer will get listed in the credits of all of those games.

 

[Laa-Yosh]

it's not like HS is such a strong IP that farming it out to different studios would make that much sense

My point exactly - if it would have made a lot of money compared to the investment, then Sony would certainly try to capitalize on its success.

Gears of War scored plenty of sales, part 2 as well, and now we can be sure there'll be a third one too. Same goes for other new IPs like Lost Planet, Dead Rising, Assassin's Creed - as well as Uncharted, Motorstorm and Resistance. Some of them haven't sold much more than 2 million copies but as their budget was probably considerably smaller than HS, sequels follow(ed).

 

[chris100]

I think the signifinace of technology interaction between game studios are exaggerated.Every game project has its intrinstic coding scheme because their gerne/design goal...and programmer talent are different.

Surely the idea could be resued,but it just like we could learn something new from the talented developers who shared their new tech on GDC lectures.You have new idea and new shading algorithm,but the essential implemetation for your project are still at blank stage for you and your teammate.

Then it needs lots of time and hard work.

 

[patsu]

The dream is that all of this tech is usable by everyone with little work. The reality is that it's not.

I doubt that's the dream. Not everyone can share an FPS approach.

Makes no difference what pr firms will say in public, there is still little reusable tech out there, and the stuff that is shareable represents a minuscule amount of code in the grand scheme of things, and still requires piles of additional work to make it even remotely functional for coders, artists and tool people. Both of the PS3 exclusive studios I'm talking with are using none of the tech people here love to mention, not even Edge. One place can't afford the time needed to make some of it work, the other simply can't make it work in their existing code/tool chain. Neither one is even considering a deferred renderer. If it's so easy to share all this glorious tech, then why are these two Sony funded studios not using any of it?

It is not necessarily about the amount of shareable code. It's about overcoming obstacles. Without sharing a single line of code, nAo's HDR design in Heavenly Sword already influenced even the XNA games. The shadow implementation, culling, etc. entails how to structure and use PS3 more effectively in specific domain areas.

I don't think anyone claimed that it's easy to share. Just possible to share and some studios have used the techs, even cross platform developers. The fact that others don't doesn't take away the value. The free PhyreEngine uses deferred rendering too (so PSN games may benefit as well). You can't use studios that don't use the shared technologies to counter the claim that the tech are shareable. As long as multiple parties use them, they are shareable.

The clues are real simple, once someone comes up with something easy to use and share, you will instantly see it everywhere. If you don't see it everywhere, then it's because it was deemed either not worth the time/money, or it was simply not possible to use. Everyone wants to save money and we will all jump at the chance to reuse code of value. It's simply no where near as plug and play as it is claimed to be.

Check out Sony's technology sharing slides. I think you'll find some examples there.

There is no GG 'shadow tech', they use cascade shadow maps like everyone else. What GG did was a small tweak on their usage to eliminate shimmering, albeit with another side effect. This is what Uncharted swiped from them. Ideas are shared, but tech for the most part is all created in house and customized to each game. Insomniac does things somewhat differently than Naughty Dog and GG. Likewise, GG and Naughty Dog have different approaches to their differed renders. In the end, each company has their own engines that are incompatible with each other.

This is true. Like I said, some developers have deployed using Edge and PhyreEngine for real. Some reuse the concept and mathematics and internalize them in their own frameworks. In fact, some may confirm independent approaches (Someone mentioned that they have something similar to Mike Acton's SPU Shaders). They are all valuable because they help to overcome specific problems.

So that means they started KZ2 in late 2006? That would imply that they totally guesstimated for the E3 2005 demo, then started started work on the game ~1.5 years later. It's hard to believe they waited that long to start considering PS3 third parties started earlier than that, but if that's what they say then I guess we'll go with it.

They probably had some prelim work going. Even the Flower team had to throw away some early work.

EDIT:

I'm no programer, but here's an ironic question to all who think that tech is easy to share between completely different engines... Why do UE3 projects need programmers at all, if there already is a complete and functioning engine with all the tools? It'd only take level designers and artists to build a new game, right?

Because sharing technology does not mean SkyNet writing your program for you.

 

[Shifty Geezer]

So essentially what you're saying is your opinion is formulated based on your own experiences and assumptions?

People throw the word 'assumption' around when someone presents an idea without hard evidence, but I think oftentimes that's unfair, and these 'assumptions' are more 'educated guesses based on reasonable empirical evidence.'

Certainly there's some return for Sony in technological IP that they can put to use elsewhere, and perhaps, best case scenario, the investment in KZ2 led directly to PhyreEngine and thus it's development can be considered part of KZ2's investment. More likely KZ2 has paid some dividends in best-practice, like any experience would. I'm not sure any economist would try to measure that as added value from investment though!

As for tools and engines derived from this tech, normally these don't save developers money, but enable them to make better use of the hardware. eg. Anyone using PhyreEngine for cheap title production would likely have spent the same amount of money to write a weaker performing engine if it didn't exist. Plus, as most development cost is in asset creation, marketing, and general non-coding things, any savings from an instant, perfect engine wouldn't amount to a lot in the grand scheme of things anyway!

 

[Shifty Geezer]

This is true. Like I said, some developers have deployed using Edge and PhyreEngine for real. Some reuse the concept and mathematics and internalize them in their own frameworks. They are all valuable because they help to overcome specific problems.

In this thread we're talking fiscal value though. That experience has led to ideas is a given. How is this quantified as added value in KZ2 though? How much of the investment cost can be considered claimed back in RnD for use elsewhere? "Yes, KZ2 cost $50 million to make, but it's produce $20 million worth of RnD" I don't think there's a vallid argument here for added value even though it exists.

 

[patsu]

It would depend on how much time it cost the developers to solve their problem, or find an alternative. And how important it is for the tech guys to achieve the vision of the art guys. Like Laa Yosh mentioned, it's project by project.

There are multiple form of sharing here: from ideas to KZ2 presentations, to say, ported code in Edge and PhyreEngine. Even the notion that a deferred rendering engine + forward renderer can be built within reasonable time to good effects may be invaluable to someone trying to play with lighting. Heck, a light-weight, free one is available right now. Sebbbi's post seems to indicate that there are merits to deferred rendering this gen. So may be now, joker454 has more ammo to convince someone that deferred rendering is viable (if it solves his team problem, but not only for rhetoric technology sharing sake).

 

[chris100]

patsu:

First,the usage of PhyreEngine is rare at some point.It even does not need deferred rendering/SSAO featured to be popular for 3rd party developer.Why it lacks so much love? It is free!

Moreover,are SSAO/deferred rendering essential for every game? Feature X could be cut in one project because 1) cost 2) result 3) performance are not optimal. Like joker said,engine feature is not a plug and play process.

New idea/feature is precious. But when you talk about the development cost,it mostly come from implementation/certification/optimization on coding and asset creation(modeling/art)

And small tips for you
Back-face culling is not a new idea from Naughty dog. I once heared this from my computer graphics book in college.

 

[patsu]

patsu:

First,the usage of PhyreEngine is rare at some point.It does not to have defferend rendering/SSAO featured now to be popular for 3rd party developer.Why it lacks so much love?

It's targeted for indie developers. May not meet the full needs of large developers.
PhyreEngine 2.4 press release is here: http://elmundotech.wordpress.com/2009/03/23/scea-phyreengine-240/
You can see a list of developers using it (Not sure why its usage was rare at some point, but it's being used now).

Moreover,are SSAO/deferred rendering is essential for every game? Feature X could be cutted in one project because 1) cost 2) result 3) performance are not optimal. Like joker said,engine feature is not a plug and play process.

No one said it is. All I mentioned is it's project by project. And only if it solves specific problems for them.

And small tips for you
Back-face culling is not a new idea from Naughty dog. I once heared this from my computer graphics book in college.

No one says culling is new.

The culling example in the early days is more a statement on how to distribute work to the SPUs. And that the SPUs are indeed fast enough to mess with the rendering pipeline. It may be valuable to help convince people about their early development approach (I remember there were lot's of confusion and insecurity over SPU usage in those days). SPU is a new beast even though culling is not new.

 

[Laa-Yosh]

In this thread we're talking fiscal value though. That experience has led to ideas is a given. How is this quantified as added value in KZ2 though?

Very true. Using nao32 may save a bit on R&D effort, will almost certainly guarantee better visual quality - but the code has to be written anyway, so almost the same amount of money has to be spent. From a fiscal point of view this does not add up to much, even if better percieved image quality can lead to better sales.

 

[Shifty Geezer]

There are multiple form of sharing here: from ideas to KZ2 presentations, to say, ported code in Edge and PhyreEngine.

Sure, it happens. I've lifted algorithms off the 'net whcih has saved me writing my own. But what dollar figure can be placed on these savings? In the context of multi-million dollar game developments, what is the payback for Sony and developers from Edge etc., principly in regard to what they've got out of KZ2 or other first party games rather than specific developer-tools projects.