Article : Only 30% of games even break even!

[ShootMyMonkey]

Yeah, to me, 30% initially sounded extremely high, but then I've never worked on casual or web-portal games or games for mobile platforms and such. I would imagine that when you put all of that together, then the 30% sounds reasonable. I think if we confined only to the current-gen consoles, then the percentage is probably in the single-digits, especially considering that far less than 30% of the console games on the market are even *known* to the majority of the public.

And even for those you have heard of, people are not often going to buy that which isn't front and center on the shelf. If a game, not matter how miraculous a monument to design it is, gets shoved aside from the prime shelf space to be replaced by Halo 12, its story is pretty much over at that point.

 

[eastmen]

Nice of you to ignore the part where I actually said what the point was. It doesn't matter how they're making money when the question is what sort of costs exist (existence of a cost has nothing to do with how you subsume it). Nor does it matter that Gears is turning a profit when you're talking about the industry as a whole being largely full of failures, simply because Gears is ONE game.

My point is , If the engine itself is a money maker and has paid for itself and continues to pay for itself why would you lump its cost into gears cost. That doesn't make any sense. Perhaps for say Metal Gear Solid 4 if the engine was specificly designed for the game and wont be liscensed out and upgraded for other games for the company that can be lump into costs. But with Unreal 3 engine you have Gears of war , Gears 2 , Unreal 3 and then about 20-30 liscensed games that I can think of. So why lump the cost of that into the gears budget. Unreal 3 engine should have had its own budget.

That's all well and good .....

Thanks for all the good info , I'm learning alot from you and others posting. However I still have more questions or ideas and so I'm going to annoy you some more. Can't you also move over animations and other thigns for the characters. For example , In assains creed they have all these crowd character models and animations and obviously they are themed for a certian time period. But now if the same company is making an rpg set in that time period can't they adapt that crowd tech and perhaps add to the models to vary it. Can't the same be said about alot of other things. I understand costs will go up over each generation but devs should try and find ways to save as much as they can and reduce the costs over the generation. Of course your not going to ever get back down to the old generation , however if a last gen game was 5m a next game gen is 15m at the start of the generation should they be working to get the costs down to 10m .

I'm not sure I follow the question. What do you mean by "reducing" it for there? What does the original Xbox have to do with that? What happens when the old Xbox is dropped altogether? What happens when you need to support something that the Xbox couldn't do at all?

Sorry reuse it. I'm asking about the xbox because didn't they have a tomb raider earlier this gen that also apeared on past consoles . I believe it was legends . What was the uncompressed size of legends on the xbox 360 and then the xbox ? You pointed out that the last tomb raider game was 400gigs of content , but what were they on the xbox 1 is what i want to know.

In a word, YES.
This is one of those examples I mentioned where part of the process was completely ripping out all legacy code and pipeline. And no, it's not running on the same engine at all anymore.

I can understand that since previously they used an older engine from last gen for the games. But going foward do they need to redesign this current engine and content again this generation ?

 

[Shifty Geezer]

My point is , If the engine itself is a money maker and has paid for itself and continues to pay for itself why would you lump its cost into gears cost.

Because we're talking about multiplatform development for all developers, which includes a lot of people writing engines that they aren't then going to sell to other folk. If in order to create a game, you write an engine for it produce the content and final game, the cost to create that game in total is the combined costs, right? If you then reuse that engine to make another game, the cost of the new game is much reduced, but that doesn't change your outlay on the first game. You can consider averaged costs across titles, and later on see a ROI on engine development considered independently of specific titles, but the long and the short of it is, to get Gears on the shelves, Epic had to have so many people working for so many years to create the game including the engine to drive it, which all cost - and that is how much it would cost anyone to recreate that game.

 

[eastmen]

Because we're talking about multiplatform development for all developers, which includes a lot of people writing engines that they aren't then going to sell to other folk. If in order to create a game, you write an engine for it produce the content and final game, the cost to create that game in total is the combined costs, right? If you then reuse that engine to make another game, the cost of the new game is much reduced, but that doesn't change your outlay on the first game. You can consider averaged costs across titles, and later on see a ROI on engine development considered independently of specific titles, but the long and the short of it is, to get Gears on the shelves, Epic had to have so many people working for so many years to create the game including the engine to drive it, which all cost - and that is how much it would cost anyone to recreate that game.

The point is they can liscense the engine for much less. I'm sure EA can liscense the unreal 3 engine or rage or what have you across a bunch of their games and spend alot less than developing an engine for each of those games.

 

[Carl B]

Eastmen you have a very engine-centric view of game development; that's understandable, because that's what a lot of us find 'sexy' in game development, but I think you're over-emphasizing it. In any event, EA developing their own engines is not something that's going to be on a per-game basis, and can be just as efficient (or moreso) than simply licensing UE3. Licensing UE3 doesn't mean that no work needs be done on the engine incidentally.

For Gears itself, the engine would have been an expense of sorts as well, because it was an internal project, developed by Epic, and engine development was occurring in tandem with Gears development (Gears was the first major title to ship UE3). Epic does license the engine out at significant gains, but that income is recorded outside of per-title development costs/profits, as it should be, since its development exists in a different context than just that of a single title as well.

Not sure if you've seen this thread either or the pros/cons of internal/external engine development, but might be of interest to you:

http://forum.beyond3d.com/showthread.php?t=47138

 

[eastmen]

1) I'm not saying it has to be ue3 engine , i was just using it as an example , even still ea can create 1 engine and use it over the majority of their games , ue3 is beign used in fps , 3rd person shooters , rpgs , racers and other things. Sure you will have to spend some money adapting it , but the cost along with liscensing or with ea making their own for the majority of their games would work out cheaper

2) Yes ue3 engine cost money to make , no one is saying it was free , however it paid for itself by being used in multiple games and beign liscensed out to multiple games , lumping the cost of unreal 3 engine into gears of war isn't a fair thing to do. We can add a liscensing fee of how much it would cost to liscense unreal 3 engine and add it to gears , but still gears was 10m for a tripple a quality game , even if it cost 10m to liscense unreal 3 engine thats still only 20m for tripple a quality gaming (and i doubt unreal engine 3 costs that much to liscense , most likely 1m or a % of each copie sold) IF a game of gear's quality can be made content wise for 10m why can't other games , esp those of lesser quality.

 

[Carl B]

Even still ea can create 1 engine and use it over the majority of their games...

What makes you think EA is wasting development resources on many redundant engine designs? Of course they leverage in-house engine design across all titles that will benefit from specific arches. Worth mentioning that EA is a licensee of UE3 on the side, on top of their own internal efforts.

2) Yes ue3 engine cost money to make , no one is saying it was free , however it paid for itself by being used in multiple games and beign liscensed out to multiple games , lumping the cost of unreal 3 engine into gears of war isn't a fair thing to do.

The $10 million spent on Gears probably has very little to do with Unreal 3 development, which would have been expensed elsewhere within Epic. Any of that $10 million going to engine development within the project would probably have originated from customization/tweaking rather than cost of direct engine development. Indeed I think I remember that some of the advances/changes made by the Gears team were later adopted into the larger UE3 offering. would be

We can add a liscensing fee of how much it would cost to liscense unreal 3 engine and add it to gears , but still gears was 10m for a tripple a quality game , even if it cost 10m to liscense unreal 3 engine thats still only 20m for tripple a quality gaming (and i doubt unreal engine 3 costs that much to liscense , most likely 1m or a % of each copie sold) IF a game of gear's quality can be made content wise for 10m why can't other games , esp those of lesser quality.

Eastmen what in the world is your continued emphasis on what a AAA title costs to develop? No one is arguing how much AAA titles cost to develop, what this entire discussion is about is how much it costs to develop games that are not AAA... just games themselves, in isolation of their perceived quality by the public. Now your original stance here was that this entire thread premise could not be true; that the majority of games must be profitable. You have had many developers speak up themselves in this thread supportive of a situation many here take as a given (hit-driven business), so submit to reality: many games lose money!

You're taking the position that because Gears of War was a $10M project, that by extension any team with $10-20M lying around should be able to make a game that sells millions of copies at retail. Does that honestly make sense to you as a position? Because it seems completely disconnected from the core concern here.

 

[eastmen]

What makes you think EA is wasting development resources on many redundant engine designs? Of course they leverage in-house engine design across all titles that will benefit from specific arches. Worth mentioning that EA is a licensee of UE3 on the side, on top of their own internal efforts.

I don't know about that , they have an engine for spore , one for skate , one for that racing game. They may use the same engine in the sequals but they have many diffrent engines running around at ea. Madden has its own engine also.

The $10 million spent on Gears probably has very little to do with Unreal 3 development, which would have been expensed elsewhere within Epic. Any of that $10 million going to engine development within the project would probably have originated from customization/tweaking rather than cost of direct engine development. Indeed I think I remember that some of the advances/changes made by the Gears team were later adopted into the larger UE3 offering. would be

Yes , because gears and unreal 3 were seprate projects at epic.

Eastmen what in the world is your continued emphasis on what a AAA title costs to develop? No one is arguing how much AAA titles cost to develop, what this entire discussion is about is how much it costs to develop games that are not AAA... just games themselves, in isolation of their perceived quality by the public. Now your original stance here was that this entire thread premise could not be true; that the majority of games must be profitable. You have had many developers speak up themselves in this thread supportive of a situation many here take as a given (hit-driven business), so submit to reality: many games lose money!

My point has allways been that you can make a tripple a title for 10m excluding engine costs . You should be able to make crap games for 10m also. You don't need huge budgets to make games. That is my point. If a crap game costs 30m perhaps they have more problems than just a crap game.

I also never said it can't be true that this many titles fail to make money. I simply have said that smarter developing can reduce the cost of games and that there are other reasons why these games are no making money and its not just game sales or costs of the game that are causing these things.

 

[obonicus]

What makes you think EA is wasting development resources on many redundant engine designs? Of course they leverage in-house engine design across all titles that will benefit from specific arches. Worth mentioning that EA is a licensee of UE3 on the side, on top of their own internal efforts.

Just to be devil's advocate, but aren't they? Criterion is still using their now-internal version of Renderware for Burnout, right? And DICE just created the Frostbite engine, no? (And then for some reason used UE3 for Mirror's Edge. We need scorecards to keep track of this.) Dead Space seems to be running on its own engine, too. At least Pandemic had been working on the Mercs 2 engine before EA bought them, AFAIK. With EA's city-state approach, do we even know how much tech-sharing goes on?

Actually, come to think of it, it does seem like there's a lot of waste. Maybe not more waste than typical software development, though, where reinventing the wheel is often par for the course.

 

[Carl B]

Just to be devil's advocate, but aren't they? Criterion is still using their now-internal version of Renderware for Burnout, right? And DICE just created the Frostbite engine, no? (And then for some reason used UE3 for Mirror's Edge. We need scorecards to keep track of this.) Dead Space seems to be running on its own engine, too. At least Pandemic had been working on the Mercs 2 engine before EA bought them, AFAIK. With EA's city-state approach, do we even know how much tech-sharing goes on?

Actually, come to think of it, it does seem like there's a lot of waste. Maybe not more waste than typical software development, though, where reinventing the wheel is often par for the course.

Having EA as the hallmark example may have opened up a different set of questions here, because I was speaking to EA as shorthand for the collective of development houses and you are referring to EA the over-arching publishing house. Yeah there are a lot of different engine designs at use within EA's individual studios, but those individual studios make (hopefully) intelligent choices in terms of their own engine development/licensing requirements.

Now, in terms of EA imposing an umbrella engine design across these houses as a cost-savings measure... sure on paper there would certainly be that. But I think if our end goal is profitability, that might actually get us farther from it. It benefits these teams that they keep an in-house (beyond EA in-house) development/technical expertise, and in many cases the existing engine platforms are pre-existing in any event. Real waste would be tossing that genre-specific work out to start from scratch with an EA 'omni' approach.

 

[Carl B]

My point has allways been that you can make a tripple a title for 10m excluding engine costs.

Why are you even excluding engine costs though? You should be able to make a AAA title for 10M inclusive of engine costs. But the industry isn't a limitless pool of top developers either; these teams are composed of individuals, and the individuals differ in their strengths, like in any industry. Not every airline is SouthWest airlines, know what I mean?

You should be able to make crap games for 10m also. You don't need huge budgets to make games. That is my point. If a crap game costs 30m perhaps they have more problems than just a crap game.

I also never said it can't be true that this many titles fail to make money. I simply have said that smarter developing can reduce the cost of games and that there are other reasons why these games are no making money and its not just game sales or costs of the game that are causing these things.

The point is that 10M is not a small budget in-and-of itself; it is still large. It is above the average for last gen. I feel we're going in circles! Going back to the airline analogy, it's rising fuel costs. Success is still success, but failure inflicts a much heavier price than last gen.

NavNuc's point earlier in the thread (what drew me in) was that a game might actually be better than a recognized "AAA" title, but even said quality - the final would-be statement in it all - does not ensure success. It's that the costs involved in development increase the pain inflicted on both quality and also-ran projects that produces a harsh environment.

 

[Shifty Geezer]

Actually, come to think of it, it does seem like there's a lot of waste. Maybe not more waste than typical software development, though, where reinventing the wheel is often par for the course.

I think reinventing the wheel is inevitable for a long time to come, simply because not all wheels fit all vehicles, and not all software solutions fit all problems, especially among widely varying titles. You also find that a wheel that worked for your last title can be vastly improved upon for the new one, but that requires a new hub/transmission/chassis to integrate into the new title.

We have to be clear how much 'waste' is real waste, and how much is part of improving systems. I know there's a lot of code waste, but you may be surprised how often ditching some code and starting from scratch is quicker and easier than reworking an existing system! I have myself written probably 3 different UI engines for various small PC developments, and that's after considering several off-the-shelf libraries, because firstly it was easier for me to write the engine then learn someone else's idea of how it should work (especially given absent/God-awful documentation ), and then because the shortcuts I used to make a quick yet functional engine for one application created a non-scaling system that wouldn't work for other titles. And yet if I had spent the time creating a single, perfect, truly scalable custom universal UI engine, it would have taken too long, been overkill for the first project, and not seen returns on investment until long after the initial project. It's kinda hard to budget for returns like, five years in advances, especially on the rapidly moving technology we have!

As long as developers are trying to squeeze juice of of these finite systems, system are going to need constant reworking. If we could happily bear the processing overhead of a lot of high-end systems, this wouldn't be the case. The ideal solution is a single Virtual Universe engine with perfect rendering and physics, truly natural motion, all the models you could ever want, where the developer only needs to set the parameters, create the levels, and write the stories. We are so very, very far from that though!

 

[ShootMyMonkey]

My point is , If the engine itself is a money maker and has paid for itself and continues to pay for itself why would you lump its cost into gears cost. That doesn't make any sense. Perhaps for say Metal Gear Solid 4 if the engine was specificly designed for the game and wont be liscensed out and upgraded for other games for the company that can be lump into costs. But with Unreal 3 engine you have Gears of war , Gears 2 , Unreal 3 and then about 20-30 liscensed games that I can think of. So why lump the cost of that into the gears budget. Unreal 3 engine should have had its own budget.

The funny thing about game development studios who also license out their engines is that the vast majority of their work on the engine is specific to their own games and they do a lot of stuff that is often useless or even counterproductive to their licensees. For the most part, when a UE3 or a Lithtech or a CryEngine licensee needs a fix for something, the fix never comes unless the company themselves ran into the same problem. Other times, there will be adjustments and changes to things you were using, but the middleware provider needed to change it for their own purposes. The reality is that the upkeep and updates that occurred to UE3 over the course of Gears' development cycle was likely 100% specific to Gears.

A funny pattern I've heard about from several people who are working with UE3 was that throughout the dev cycle of Gears, the support for PS3 was sad and pitiful beyond words (and many things didn't function at all). UT3 comes along, and that not being a platform-exclusive, all of a sudden the PS3 support shoots way forward out of necessity and then 360 lags behind... Now the 360-exclusive Gears 2 is the focus, and once again, and PS3 is left alone, and there is zero support provided for it no matter how much the licensees ask for it. I don't mean to suggest that Epic is a failure as a middleware company -- simply that they're still a game company first.

Thanks for all the good info , I'm learning alot from you and others posting. However I still have more questions or ideas and so I'm going to annoy you some more. Can't you also move over animations and other thigns for the characters. For example , In assains creed they have all these crowd character models and animations and obviously they are themed for a certian time period. But now if the same company is making an rpg set in that time period can't they adapt that crowd tech and perhaps add to the models to vary it.

As long as the bone riggings are identical, you can share animations just fine, and that happens quite a bit as well. Crowd characters and random stuff are commonly handled this way. Things like base animations tend to be very context-free and localized in their impact... it's when larger contexts and more dependencies for data start coming into the picture that things stop being shareable.

Sorry reuse it. I'm asking about the xbox because didn't they have a tomb raider earlier this gen that also apeared on past consoles . I believe it was legends . What was the uncompressed size of legends on the xbox 360 and then the xbox ? You pointed out that the last tomb raider game was 400gigs of content , but what were they on the xbox 1 is what i want to know.

The 200 GB of content was in reference to the current TR due to come out in November. Legend was before my time, so I have no idea how big it was... that said, the leftover assets from Legend that were used as a "startup" baseline content set when Underworld was less than 6 GB, but that was with zero actual content-filled levels (all of that stuff had been removed since we didn't need it as a starting point). It was all testing units and things -- everything was block mesh. Sure, most of these testing units had script and triggers and all sorts of things set up within them, but it's generally one thing at a time in order to isolate features.

I can understand that since previously they used an older engine from last gen for the games. But going foward do they need to redesign this current engine and content again this generation ?

Depends on to just what extent you're talking about... Is this engine going to be torn down again like the last within this generation? nope... is there going to be some stuff ripped out and replaced or renewed? You bet there will. It's not like it's in a finished state or anything. And it's also in the nature of any tech development that once it's used, you realize what sucks about it and you end up thinking about how to improve it. And of course, in terms of game-specific features and the sort of content and tweaking they demand, that's often a blank slate to start.

It's often easy to look past it when you're writing about it in short form, but there's really a pretty astonishing amount of stuff here, because it goes so incredibly low-level. In the old days, people missed things like a grimace on a characters face when they walked within 177 inches of a burning tree, and so developers wouldn't really bother either with such a feature let alone impart control to designers over that sort of thing... nowadays, artists and designers have a lot more control, and that means a lot more stuff to adjust and tweak and a lot more emergent features and techniques come up in the process. And to add to the work, they not only *get* to think about things at this level, but they also *have to* think about what could happen if they overlook something at this level.

 

[obonicus]

I think reinventing the wheel is inevitable for a long time to come, simply because not all wheels fit all vehicles, and not all software solutions fit all problems, especially among widely varying titles. You also find that a wheel that worked for your last title can be vastly improved upon for the new one, but that requires a new hub/transmission/chassis to integrate into the new title.

True enough, and that's why Software Engineering is an active field of study, despite it being something of a bad word. But I'm not sure how much of that isn't actually a consequence of the way software is built; it's still very much artisan-work, with very little actual process involved. There's very little concept of 'good enough' in software development, unlike, to extend your metaphor, the car industry: most cars don't have the man-hours or the components of a Lamborghini, and yet they still run pretty well. You want Lamborghini quality, then you unfortunately have to pay for a Lamborghini.

Maybe it's a matter of tools, though. Last-gen we had renderware, this-gen it looks like it's UE3. We have a kajillion other middleware solutions, and I doubt most of them are truly robust in any real sense of the word. There certainly are no widely-available training programs for use of these tools, and, from the looks of it, they don't really offer anything like a service contract.

EA developing a half-dozen engines for games is pretty wasteful, no matter what. I understand why it happens, but it doesn't stop being wasteful. In fact, I'm sure EA knows it: there were reports early on in the gen that they tried to use Renderware and it didn't live up to expectations. Would they be better served by an internal 'engine' team that took in teams' requirements and pumped out a continually-evolving engine (would such a thing even be possible in the time-frame required?)? I have no idea, but it certainly can't look good to have to pay multiple times for someone(s) to, say, 'figure out the PS3'.

 

[SG79]

I think he was referring to this article. Here's the image in question:

Problem is with this chart, or any comparison that includes the three hardware manufacturers, it's "too" accurate and includes all company/sector businesses. It doesn't reflect software profits or less accurately at all. Not that I'm saying that you're implying that.

Nintendo for example includes all their investments (stocks in other companies like Bandai etc), as well as Pokemon among other things. MS's includes their PC games as well. That leaves Sony, which is as pure as console figures get but that big red stems from PS3 hardware and associated costs.

 

[Carl B]

I think the charts value stems more from ignoring Nintendo, MS, and Sony rather than considering them. You'll note no one has referenced any of those three really, because they're recognized to exist outside of the context of the topic. It's still a nice chart to have.

 

[dobwal]

I don't know about that , they have an engine for spore , one for skate , one for that racing game. They may use the same engine in the sequals but they have many diffrent engines running around at ea. Madden has its own engine also.

Yes , because gears and unreal 3 were seprate projects at epic.

My point has allways been that you can make a tripple a title for 10m excluding engine costs . You should be able to make crap games for 10m also. You don't need huge budgets to make games. That is my point. If a crap game costs 30m perhaps they have more problems than just a crap game.

I also never said it can't be true that this many titles fail to make money. I simply have said that smarter developing can reduce the cost of games and that there are other reasons why these games are no making money and its not just game sales or costs of the game that are causing these things.

You think Gears runs the same generic engine that everyone who has a U3E license gets from Epic. Nope, Gears runs a customized U3E engine and just about everyone that uses a U3E engine has a custom build to suit their game. SK took the general U3E engine thats wasn't anywhere near functional and fleshed out its functionality and features and renamed it.

EA may have many different engines but I bet many are just derivatives from a few common sources.

 

[ERP]

Here's a datapoint, I think most large companies have cost themselves more money trying to unify technologies and enforce sharing, than it would have cost to have independant teams maintain there own "engines".

People underestimate the cost of sharing, and the effect it has on the technology.

For example I know a large company where pretty much every title uses the same sound library, because of the divergent requirements accross the 100 or so SKU's a year, the library is probably 10x larger than any one team needs, is significantly less efficient than it should be for any given 1/10 of the functionality and still has significant bugs, in infrequently used code paths.

For Another example look at the size of the Havok library.

Now you extend that to a whole game engine and you have a mess usually with little or no local support. Most large companies ty to do it (usually on some cyclic schedule) then back away from it.

In a large company IME it's much better to provide ways for people to see what tech is available, give them access to it, provide contact information for the owners, and let teams adopt what works for their problem. If your hiring people who rewrite everything all the time without reason resulting in late/ovre budget products fire them.

But this is somewhat off topic, the cost of developing a game at a large company (like EA) and at an independant developer (like Epic) are radically different and very little of that has to do with the cost of engine development.
Independant developers are just a lot more efficient as much because of need to be as anything.
Big companies can have large teams, so they do, they build infastructure to deal with large teams (leading to more overhead) and they tend to have deadlines that are expensive to move, leading to more people if the project starts to flounder. Big companies will spend 2 or 3x what a small company would spend to develop the same product regardless of technology concerns.
The argument for internal vs external development has always been control vs cost and the industry has been pretty cyclic in the preffered development paradigm.

 

[obonicus]

Here's a datapoint, I think most large companies have cost themselves more money trying to unify technologies and enforce sharing, than it would have cost to have independant teams maintain there own "engines".

People underestimate the cost of sharing, and the effect it has on the technology.

For example I know a large company where pretty much every title uses the same sound library, because of the divergent requirements accross the 100 or so SKU's a year, the library is probably 10x larger than any one team needs, is significantly less efficient than it should be for any given 1/10 of the functionality and still has significant bugs, in infrequently used code paths.

For Another example look at the size of the Havok library.

But is this the impact of sharing, or the impact of not having an actual enforced architecture to your library? I'm certain you can correct me, but most of what I hear about the game industry is that development is extremely ad hoc. There are counterexamples, naturally, like Insomniac, who actually appear to have a more robust architecture, but then again they've been releasing a game a year.

Again, the only advantage I see to each studio rolling their own engine is if that it may be easier to meet a deadline. Otherwise, essentially, we're dealing with Not-Invented-Here syndrome. I mean, if all generic engines were necessarily going to become overbloated, then every single engine licensee would fail before they've even begun. Hell, if I'm being alarmist, it means that there'd never be a set of standard industry tools of any real use, which means you wouldn't be able to rely on finding resources with the necessary skills outside your own team -- there'd be no such thing as a required skillset, other than the extremely broad ('graphics programming', 'C++', 'AI'). It doesn't seem sustainable. [Though as I write this I find it does sound a lot like the games industry from where I'm sitting, haha, excluding maybe the unsustainable part.]

 

[Laa-Yosh]

I'll eat a hat if the dev costs for Gears are truly 10 million. That was obiovus marketing talk from Rein to boost their licensing business; as far as I remember it did not include not just the engine development, but none of the art outsourcing costs either.

COD2 was $14 million with far less detailed art content, and a pretty much inferior engine that's been built upon a previous one. GOW is at least as much, considering it wasn't multiplatform and had less gameplay, both single and multi, to develop and test.


And anyone can tell you that licensing UE3 isn't gonna help costs at all, it just speeds up the early stages by not having to start from scratch.