Age of Empires III Q&A - Technology Overview

After seeing the stunning first screenshots of Age of Empires III, we had to know about the technology behind the game, so we caught up with lead programmer Dave Pottinger.


Simply by virtue of its announcement, Age of Empires III instantly became one of the most anticipated PC games of 2005. It's no secret that the first Age of Empires games are two of the most successful and popular real-time strategy games ever made. However, Age of Empires II is almost 6 years old at this point, an eternity in computing terms. While developer Ensemble Studios released the successful Age of Mythology in 2002, fans still yearned for a return to the Age of Empires series. And with Age of Empires III, they'll get it, as the first screenshots of the game to be released so far are nothing less than astounding. For the details on the impressive technology behind Age of Empires III, we caught up with Dave Pottinger, the lead programmer of the game and the director of technology at Ensemble Studios.

Age of Empires III will take place in the New World. It will also have plenty of new, high-end graphical features.
Age of Empires III will take place in the New World. It will also have plenty of new, high-end graphical features.

GameSpot: What were the design goals behind the new engine, besides "it should look really good"? What kind of gameplay did you envision the engine to support? Are there specific gameplay features, like weather effects, environmental deformation, modeled physics for units getting knocked back by cannons, and so on, that the team found itself building tech for into the new engine?

Dave Pottinger: Simple. We told the team to make the best-looking PC game ever. A simple statement that was anything but simple to execute. When we started Age 3, we knew that we had some really cool gameplay ideas to try out. How do you not have that after doing real-time strategy games for 10 years? But, we wanted to recapture some of the initial "holy crap" visual appeal that we got when we showed Age 1 for the first time. Going high-end with the graphics was an easy way to do that. It hadn't been done before in our genre, we had a lot of guys who were crazy excited to do it, and it was hugely bold. So bold, in fact, that we've had a lot of internal discussions about whether it was smart to shoot that high. In the end, the fans will tell us whether we've made the right choice with the graphical direction. We're certainly happy with the current results.

Gameplay-wise, yes, the graphics do drive the gameplay, and vice versa. We try to leverage opportunities in each all of the time. A good example would be cutting down trees. We put physics in for a bunch of things like building destructions and rag-doll effects. But one afternoon a programmer grabbed me and showed me how cool it looked when we put physics on the trees as they fell. It was a perfect extension. Throw in a tiny little camera shake as the tree hits, and the effect is great. It extends the other way, too. Since the trees are physics-enabled, an errant cannonball can hit them and knock them down. Stuff like that is just cool.

GS: When did the team begin working on the game, approximately? Does this represent the evolutionary growth of the existing Age of Mythology engine, or did you end up more or less starting from scratch? Are there any specific lessons the team has taken from previous games, like Age of Mythology or the original Age of Empires 2D series, that are being incorporated into the new game's development?

DP: We started with about seven people right after Age of Mythology shipped. We're up to considerably more than that right now. Apart from a couple of prototype teams, most of our studio is working on Age 3. We started with the Age of Mythology engine. That was nice because we had built-in fallbacks. Age of Mythology was a tough development cycle because we rewrote everything. We were adding core functions to the Age of Mythology engine right up until ship. (Who really wants to work on localization code early?) It was liberating to start with that stuff behind us.

By now, though, so much of the Age of Mythology engine has been replaced that the code base feels very different. We've certainly tried to carry all of the knowledge we've collected over the years into the new code. If there was one thing that we learned from Age of Mythology, it was to really finish features every time we do them. Age 3 has been a smoother development because we've taken that extra bit of time with the various development tasks to try to keep things sane. It helps a lot, particularly since Ensemble is a massively iterative game company. We throw stuff out all of the time. If 25 percent of the gameplay code that I've written for Age 3 is still there when we ship, I'll be very surprised. So, you can imagine how tempting it is just to slap something quick in the code and move on. We've found that just creates too many problems down the road.

And, yes, we worked on localization early this time.

GS: The game really looks good, even this far out, so we'll go ahead and ask: What are some of the really cool technical features that are being built in the game? Pixel shaders? Vertex shaders? Environmental bump-mapping?

DP: Yes, yes, and yes. The single biggest graphical feature is probably the move to high dynamic range lighting. That allows us to have scenes that look much more natural because, in layman's terms, we have more dials to make the lights brighter. That results in a world where you can really feel the sunshine and the heat. Lighting is a subtle beast. It's one of those things that doesn't necessarily jump out at you. If it's good, it just feels right. The best compliment someone can pay to our game's lighting is to say, "Wow. That looks real."

The level of detail is almost astounding. Look carefully--those are turtles sunning themselves on those logs.
The level of detail is almost astounding. Look carefully--those are turtles sunning themselves on those logs.

GS: What kind of system requirements are you aiming at, both in terms of minimum requirements and optimum requirements? Will you need any specific kind of hardware, such as a hardware transform-and-lighting-capable video card? How will the game scale down for older machines? Any plans to build special support for current and/or next-generation hardware features?

DP: Our games have always sold a bazillion copies, so we definitely focus on hitting a wide hardware market. Our games sell for years, so we have to be smart about what we support, both in terms of what we ship on day one to what's going to be coming out in the future.

Age 3 will support fixed-function cards, all the way up to the newer shader model 3.0 cards. Obviously, the better your machine, the nicer your game will look. I would easily expect that the game has enough high-end graphical features that most folks won't be able to crank it all the way up right away. The hardware still has a bit of catching up to do.

Age of High-End Technology

GS: Has the multiplayer code been rewritten from scratch? Will the engine allow for new kinds of multiplayer modes? How many players do you envision it being able to support? Will you find games in the server browser, or do you envision using a matchmaking service of some sort?

This snowy landscape doesn't just look cold; it feels cold.
This snowy landscape doesn't just look cold; it feels cold.

DP: We have indeed rewritten all of the multiplayer code, both client and server. We built Ensemble Studios Online (ESO) for Age of Mythology. That was a huge task for us to undertake; we've certainly carried those lessons forward into ESO 2. At this point, the multiplayer development is going very well. One of the bigger tasks left for the ESO team is to find a better name for ESO 2 than, well, ESO 2.

We're not talking about specific multiplayer features yet, though we do have some awesome things to roll out there in the future. I can, however, talk about our philosophy change for ESO 2. ESO 1 was built with the goal of drawing in the casual gamer. We actually removed buttons from the main ESO 1 page to make it less scary. In hindsight, that perhaps wasn't the best decision. We've totally changed gears for ESO 2. It's built for the folks that use it, the hardcore gamers. We've gotten excellent feedback from the testing we've done on it so far, so hopefully that will prove to be a better direction.

GS: In the screenshots that have been released thus far, we've seen some incredible battles. How many units will the game be able to render onscreen at once? Will there be some artificial cap, such as population limits, that will kick in before you even get to the limit that the engine can support?

DP: The goal for Age 3 was to slap 300 to 400 units on the screen at one time, which is a big step up from Age of Mythology. That's a pretty big number, so we've spent a ton of time optimizing the game all over the place to make that happen. Yes, there is still a population cap. Our goal, of course, is to have as few folks as possible hit the pop cap during normal gameplay. It's there to prevent the edge cases from creating a bad gameplay experience.

GS: We've touched on the visuals, but is the technology team building in any kind of new artificial intelligence, either as campaign AI in the single-player game or as general unit-based AI to dictate formations, morale, and default behaviors? Will the AI "cheat" by getting unfair advantages like seeing through fog of war, building faster, or starting with more resources, or will it play with the same rules that players have to?

DP: Our AI players have always played a fair game. It's one of the big things we take pride in with the Age franchise, actually. On the uberextreme mode, they've gotten extra resources in the past, but we've always been up front about that. The goal for Age 3 is to get the AI to be competitive with the hardcore players in a one-versus-one game. That's a hard task considering the complexity of RTS games. We've had two to three people working on the player AI for the last couple of years, in part to get it to play better. Those folks have spent the rest of their time getting the player AI to play a more fun game. We'll talk more about some specific new features for the player AI down the road, but that's about all I can say right now.

The entity AI (units and groups) has gotten a lot of work, as well. We've spent a lot of time optimizing it so that we can put more units on the battlefield. We've also redone a big chunk of it to do our new combat system.

GS: Sound is also an important feature in games. Will Age of Empires III feature any kind of advanced sound technology? Will each unit have unique sounds? In gigantic battles, will you be able to hear the clash of armies? Are there any plans to build in a dynamic soundtrack that changes music on the fly depending on whether you're at peace or at war?

DP: Yes, yes, yes, and yes. We've actually spent quite a bit of time on the sound. We've been adding more features to our dynamic music system, tying that in to an improved concept of battles and doing cool things like filtered 3D positioning for things like cannon blasts. It's very cool to hear the "thump" of cannons in the distance as you pan across the gameworld.

GS: How much did the fan community (and its many enthusiasts who use editing tools to create mods, new maps, and other custom content) come into play when the team sat down to design and continued to create the technology that powers the game? Will the game be mod-friendly, and will people be able to create unique mods and other content?

DP: It really depends on the part of the game, honestly. We love what the fan community has done with our previous games, so we try to give them easy hooks and support for doing things where it's possible. In the past, we've crammed so much into our schedule, though, that we haven't had time to even do simple documentation tasks. We're hoping to fix that this time around. That said, our game does have some larger hurdles to overcome with modding, given things like its multiplayer model (a peer-to-peer synchronous game has to stay in lockstep data sync, so modded files can frequently cause phantom bug reports if data is changed on only one machine) and online cheating (several things that we left in Age of Mythology to make it easier to mod turned out to make it way too easy to cheat in multiplayer).

GS: Finally, what are some of the cool features or visuals that we can expect in the game? Anything in particular that stands out so far?

The home-city graphics are simply stunning to behold.
The home-city graphics are simply stunning to behold.

DP: At this point, you can fire up the game and it pretty much demos itself, so there's a lot of nifty stuff to choose from. If I were to pick two things that stand out, though, I'd pick the home-city graphics and the physics. I'm a programmer, so I'm obviously picking techy stuff, but I think both of those are also great examples of everything coming together across the project.

The home city is this massive graphical display of coolness with ambient occlusion lighting and superdetailed buildings, all built to support one of the most innovative features in recent RTS memory. The graphics of the home city are cool, in part because it's our number one gameplay feature and everyone sees the crap out of it.

Physics are just fun. It's a combo of a highly technical piece of math, art done just right so that it breaks apart properly, and a game design that gets lots of units on the field at once so they can be blown to hell. We've done a ton of demos for Age 3. Once people stop gawking at the overall graphical impact of the game, physics is the money shot that closes the deal.

GS: Thank you, Dave.

The products discussed here were independently chosen by our editors. GameSpot may get a share of the revenue if you buy anything featured on our site.

Got a news tip or want to contact us directly? Email

Join the conversation
There are 2 comments about this story