In the video game industry, there aren't too many designers whose names are immediately recognizable to gamers, and within the strategy genre that number is even smaller. Chris Taylor, mastermind behind the acclaimed 1997 realtime strategy game Total Annihilation, is one among that number. After ditching RTS for a bit to work on the Dungeon Siege games, Taylor is returning to the large-scale strategic warfare segment in which he built his reputation.

Supreme Commander, developed by Taylor's studio Gas Powered Games, is one of the most anticipated titles in the strategy genre and, by just about all accounts, the most ambitious. The game promises to bring strategy back to realtime strategy, and give players the ability to wage simulated war on an unprecedented scale and with new degrees of realism and control. I recently had the chance to speak with Taylor about some of the game's features, what's wrong with most RTS games these days, why he does what he does, and how he ended up working on epic warfare games with adventure game icon Ron Gilbert.

Shack: What do you see as lacking on today's realtime strategy environment? You're known to be fairly opinionated on this matter, and Total Annihilation is known as being a very unique game within its segment. What's missing that you're trying to provide?

Chris Taylor: Well, it probably goes back to my teen years when I played a lot of Risk. You know, Risk was a board game but it was very much a strategy game, and you were playing on a world map. Strategy, I've come to believe, happens when you play on a large theater of war. It's hard to employ a strategy when you're in a tactical situation or a battle. See, I often say that strategy happens within a war context, and tactics happen within a battle context. RTS games are lacking strategy. You're fighting little skirmishes and little battles, and you really don't have enough time. If you and I sit down to play a game and we're on top of each other in five to six minutes in terms of our units, how could I have hatched some massive plan when I've only had a few minutes? But when you put me into a large theater of war and I've got fifteen or twenty minutes, and I see a scout fly ahead I can say, "Oh, they know my position!" then the options open up and I can start making much more strategic plans.

People worry and they say, "Well, does that mean Supreme Commander is going to be just huge long games on gigantic map?" Well, you have a choice of what size maps you play on. If you want, you can play on small maps and it becomes very tactical again, but you can also have incredibly large options available to you where the maps get bigger and bigger and bigger. You can play in an incredible theater of war and the game goes from being very tactical to very strategic. Then things like information war comes in, and you start thinking, "Well, if I put a base up here, my opponent won't think to look there, and then I have these control routes," and you start to have much more sophisticated plans that way.

Shack: Speaking of those kinds of broad strategies, we haven't heard too much in terms of how campaigns are going work. Are you planning on using the single-player to ease gamers into that kind of gameplay if they maybe aren't accustomed to that kind of RTS?

Chris Taylor: Well, we haven't talked about our campaigns because we have a stretch of time before now and launch and we have a PR campaign that has us releasing details of the game over time. We will be forthcoming at some point about the multiplayer and how the campaign is going to work in the coming months. We have a plan there. You don't want to throw the kimono off completely, it's part of launching a new IP.

Shack: So you mentioned recently that you're shooting for unit caps of about 500 per player. Have there been any difficulties getting this working with the enormous maps on the technical side, or in terms of making sure it's comprehensible to people?

Chris Taylor: The way we're working is that we have a pool of units, so that if you're playing with fewer people, you have more units [per person]. We've toyed with two to four thousand units in a pool. So let's say you and I play head to head and it was four thousand units, we'd get two thousand each. If we played eight player, we'd get five hundred each. What that final unit count is, it's not yet been decided, but in all fairness whatever it is when we launch the game in January it will be different in June. The powers of the machines will go up, it's fully moddable, people can change the number of units to whatever they want with a mod. So if you and I get together and we know a little something about how to mod the game, we can change it to ten thousand units each. There's really nothing about how the game is architected that limits that number. It's an arbitrary limit that will be set at release, and it will probably be within two and four thousand units total.

Shack: What kind of machines are you targeting at launch for the default values?

Chris Taylor: It will be something north of a 2GHz CPU. There will be a fairly wide range of video cards, probably north of 128MB cards. I would say that there will be a suggested system spec of somewhere closer to 3GHz machine with 256MB of video RAM or more. But we have got rendering detail settings and so forth, such that if you've got a lower end machine, you'll be able to play the game, you'll just have to dial it in so that it doesn't look as pretty. The game's CPU is dedicated to simulation of the very comprehensive gameplay. Because of the fidelity of the simulation, that has to be there no matter what. In multiplayer, we use a fully synchronous multiplayer model, which means everything that happens on your opponent's machine happens on your machine in lockstep. So, our choices are generally in visuals since it has no negative impact on gameplay. At the end of the day you'll get the game running on a 1.8GHz or an AMD 1600+, but the ideal experience is going to be had with a state of the art gaming rig.

Shack: And you guys have dual core support in there as well?

Chris Taylor: We do.

Shack: Going back to that gameplay aspect, stuff like fully simulating projectiles obviously has big gameplay ramifications. For some people, the weight of that may not strike them immediately; could you maybe give some hands on examples of how that impacts the game and what could happen using that model?

Chris Taylor: When you have a simulation, you have a byproduct which is fraught with what we call emergent gameplay. If you use a determinant system, a system whereby if a tank turns its turret and fires on another tank the only time it will fire is if the enemy tank is within range and in sight and is guaranteed to hit, in that old non-simulated system, when you see the tank shoot you've already subtracted the shell damage from the health of the opposing unit. Effectively, everything that happens and will happen all in one chunk.

When you simulate, if the tank sees another tank, it leads the tank by a certain amount and fires the cannon, the shell may hit the enemy unit, or maybe it hits a tree, or maybe another unit is driven in front which wasn't anticipated and takes the round instead. Maybe that other unit is stronger, or weaker, or maybe it's wreckage from another battle. It creates this total product of excitement, an emergent feast for the eyes. It's different, it's better, it's more interesting, and everybody knows it in their gut when they see it.

Remember the old physics, before we had physics, where everything would bounce and roll as if it was a sphere, and then one day someone showed us Havok and we all said "Ooh, that's the way it's supposed to be done." There's still a lot to be done there, but we were on the right path. Well, emergent gameplay behavior comes from simulation and we saw that in Battlefield 1942--well, we've had it games for years, it was in Total Annihilation and it goes back before that, but the reality is that nobody quite put this finger on it. Nobody could quite explain it. It was one of those situations where gameplay had evolved and nobody had the terminology to describe what was happening. Now we understand that emergent gameplay comes from simulated systems.

Shack: I've noticed that some people seem to think it's a marketing buzzword, and it can be hard to convince them it's not.

Chris Taylor: It sure sounds like a marketing buzzword, but it's a very real thing, and it's very tangible and it adds value to the game in a big way.

Continue to the next page for Chris Taylor's thoughts on RTS balance and more.