I’ve seen a couple of articles that argue that real time strategy games (RTS’s) are “dying”. They’re not. They’re waiting. Waiting for the adoption of the hardware to implement the next generation RTS designs.
Hardware drives innovation
It is easy to forget how closely tied video game innovation is to the technology that it runs on top of. We didn’t even have “first person shooters” until Castle Wolfenstein 3D because the computing horsepower wasn’t there. The hardware drove the creation of an entire new genre of game.
Another hot market is “MOBA” (mobile online battle arena). This first got started as an RTS mod called Defense of the Ancients (DOTA). But this market only came into existence once low latency, high speed Internet became mainstream. It’s not as if the concept for MOBA had eluded game designers. Popular MOBAs like League of Legends, DOTA2 and of course the much anticipated Sins of a Dark Age are premised on players having high quality network connections.
We have done everything we can within 2 GB of memory and 2 cores
So what about real-time strategy games? Why has their design stagnated? The answer is that developers have pushed their designs about as far as they can go with the current hardware. We’ve been stuck with a 2GB memory limit for over a decade and limited to 2 or less processors (cores) for longer than that.
Strategy game designers regularly quip about first person shooters and MOBA game hardware requirements “Must be tough having to mange and render 4 units all at the same time…”
Demigod has amazing visuals because it can – as a MOBA, it doesn’t have to have as many units in the world
Gas Powered Games made one of the most ambitious RTSs of all time – Supreme Commander. They also made the first stand-alone MOBA – Demigod. The graphics of Demigod go far beyond what’s in Supreme Commander at a fraction of the budget. Why is that? Because Demigod didn’t have to deal with hundreds of independent units running around at once which reduces memory use.
The gorgeous visuals and detail in Ironclad’s Sins of a Dark Age also mean fewer units in the world in order to be compatible with most PCs
The bane of game developers: No matter how much memory you have installed on your computer, 32 bit games can only “see” 2GB. That’s a limitation put on 32-bit processes in Microsoft Windows.
Ironclad’s Sins of a Solar Empire series pushes RTS’s about as far as they can go on the current hardware
I think most gamers would be shocked at how quickly memory gets consumed today. The icon for your favorite game probably uses more visual memory than many DOS games did.
What would a next-generation RTS be like?
Are you an RTS fan? Imagine what a truly ground breaking RTS would be like. Huge scale? Lots and lots of units? Huge armies battling it out on screen with amazing graphical fidelity? Throw in some innovative features to make managing these units really intuitive and compelling? It’s technically possible to do it, you would just have a hard time making money because most of the market wouldn’t be able to play it yet.
As a publisher, I get to see a lot of really innovative game designs cross my desk. There’s no shortage of innovative RTS designs. They just can’t be made profitable yet – their design requirements revolve around a player with DirectX 11, a 64-bit Windows OS and 4 cores (minimum). They could do their innovative design with much lower visual fidelity and get most of what they want but then it becomes a “budget” title.
After all, can you imagine what would happen if you tried to make some sort of City simulation game today? People would expect to be able to zoom in on individual citizens and have breathtaking visuals but also expect it to take place on some giant map. The designers would have to either have to make the map you play on be really small or cut down the amount of detail or cut down the visual fidelity of the world/buildings (or a combination of all 3). It would be very difficult and expensive to pull off right…
Chris Taylor’s Total Annihilation, arguably still the best RTS ever made
Third Phase RTSs
We’ve been through two phases of RTS’s already. The first phase was the DOS and early Windows era games. Dune, Warcraft 1/2, Dark Reign, Starcraft. Sprite Based.
The first game of the second phase I’d argue was Total Annihilation. Even though it didn’t make use of 3D hardware, it was the first game to deliver real-time rendered units. Think of the second phase as the age of 32-bit, 3D RTSs. Supreme Commander, Warcraft 3, Sins of a Solar Empire, Starcraft 2.
In both phases, the hardware drove the design of the RTS. Nothing highlighted the difference between Phase 1 and Phase 2 more than real-time Strategic Zoom. Strategic Zoom gave players the ability to zoom out and control their armies in a truly strategic way or zoom in and take control of individual groups. If Ender’s Game was, ironically, an RTS. You can assume that players will expect these kinds of features at a minimum going forward.
The third phase games can be broadly described by the technology under them: 64-bit memory, massively multithreaded. And these 3rd phase RTSs will be breathtakingly beautiful to look at, have amazing scope and micro AI (sophisticated rules for units interacting with one another without human involvement) that is astounding.
Publishers aren’t approving these designs not because there isn’t demand. There is. The problem is you can’t make a game that only a fraction of the player base can currently play.
They’re coming…
According to the Steam Hardware Survey of January 2013 over 40% of users now have 4-core (or more) machines. DirectX10/11 is mainstream and almost 70% of users have 64-bit machines. Not good enough. Not yet. But almost.
Uber Entertainment is working on Planetary Annihilation which might be the start of 3rd phase of RTSs
Keep a close eye on the specs of games. Be looking for when games explicitly require a 64-bit PC and DirectX 10 or later as minimums. Once that has become mainstream, it’ll be a race by studios to position themselves.
The RTS game concept is compelling. It works as a single player experience and can be played as a multiplayer game. Gamers want these games. They’ll pay for these games. But they also have to be measurably superior than what’s already out on the market. To do that, the hardware has to catch up. The good news is that it’s almost ready.
Total Annihilation is still the best RTS ever made!
Amazing write up, thanks for taking the time to put it into words.
I remember when you said that the reason you were on PCs rather than consoles was because of the additional memory available to you there.
Later this year we'll probably have new consoles out, and those new consoles are rumored to have between 3.5 and 5 gigs of memory available to the developer along with 8 cores of AMD64 processors... Will that be enough for the next generation of RTS games and leapfrog ahead of the average PC?
Also, in regards to the Steam survey, I guess I'm in the "Other" category on processor count, because my gaming PC is a 3-core.
Good article/post Frogboy - I agree the demand is definitely still out there - it's just pent up at the moment.
Completely agree sir. Anyone remember that 4th dimensional time traveling RTS game Achron? It failed for many reasons but the big one was hardware limitations. Outdated engine, windows x32 system memory limit, no multicore, etc.
The funny thing was despite Achron's severe shortcomings in game design or lackluster visuals, going back to the old school 2nd dimension RTS just didn't feel right. It's weird. After playing in the multiverse you become accustomed and the normal RTS seems strangely inferior and unentertaining. You can't play it anymore like how mathematician geniuses won't or can't do simple math.
Yeah I always assumed that's why Ironclad is doing Dark Age now. There is just no way to innovate much more until the technology catches up to their ambitions.
I have to say I don't agree that hardware is the limiting factor (as I don't agree that RTS are dying either). There is no hardware related issue that made Blizzard make Starcraft 2 a nearly carbon copy of Starcraft 1, while Relic has managed to put original RTS games out there for their whole career (with better or worse luck, but at least exploring original design directions with most of them).
AI for example, it has not evolved really much in the last years, nor the budget in games has grown too much either (as usually most of the programming/time/money budget is spend on graphics/art). What is the biggest AI innovation we have had "recently"? GOAP and behavior trees, and those came from shooters (F.E.A.R. and Halo).
So yeah, better hardware will allow us to have prettier games (like the people of Creative Assembly are doing, Rome 2 looks impressive), and brute force more of the AI (but sadly, overall I don't think it will get much better than it is right now), but I don't see a revolution coming just from the fact of having better hardware honestly.
Kohan > Total Annihilation.
Beaten to the punch by Alstein
Total Annihilation is the second best RTS ever made and was my firm favourite until I found Kohan: Ahriman's Gift.
So Frogboy's point is still a good one
Cult following will tend to do that. Innovation was never going to come from Starcraft no matter how good the technology is, beyond graphics relating improvements and minor usability tweaks.
Games like Sins and Supreme Commander were definitely hitting the technology limit as Frogboy mentions. I know some of the fan matchmaking system automatically put Large Address Ware on it (too bad that doesn't seem to help for Sins). I have no doubt they'd love to make even larger scale games but or add a lot more details to things but just couldn't at the time without destroying performance.
I still think the FE world would be a perfect place for a Kohan-style RTS. It's got many of the things needed to make it work.
Too bad Stardock isn't interested in doing a TBS themselves.
One crazy idea: since Brad is a big fan of GPG and they're in serious trouble, why not pitch the idea to them of doing an RTS in the Elemental world if the Wildman kickstarter fails?
Well, I'm not getting a four-core. By the time I can afford one they'll be obsolete.
This is an area where I disagree somewhat with Frogboy. Sure, having more hardware available is a nice thing. You can do great things with it. Things that would have required more thought if you had less resources and might have even been impossible.
But just labeling things as impossible without more hardware seems like giving up. I guess viewed from a game company angle it's financially best to do so, and looking at the industry on a macro-level it's wise to do so. But from the perspective of an individual game, I wouldn't give up so easily. By this I mean that I admire greatly the pioneers who have invented great ways of providing more with less. For example the Galaxy generation of Elite, hundreds of distinct star systems are created by a few kilobytes of code. You didn't see Braben and Bell just giving up and saying "we need to wait for the next generation of computers." Or more recently the Total War games by Creative Assembly. Thousands of detailed, individual units fighting real time with amazing graphics. There are some very cool innovations there.
it does depend on what you mean by rts, doesn't it.
if you are aping being a so called REMF that does the strategic stuff, then you would think the ranks below would know how to think and act on its own initiative. hey.. you are getting bombed (air or artil).. scatter and do something about the source without being told to do so. you got that supergun. hey.. use it regularly instead of getting destroyed whilst using a pea shooter
obviously. it would get to a point where you have to decide whether you are playing the game or the game is playing you if the ai is that good.
gfx? lots of units? seems to be missing the point a bit (for me). no good to have brilliant gfx and a ton of unit if they all just rely on you to tell them what to do.
or put another way.. has anyone tried to make something that's low on visuals but high on ai... within current limitations? it probably won't sell but hey.
there's always better hardware to be wanted, with better gfx and numbers hitting whatever the latest limitation is and still have the same old idiot units sitting there.
The problem is that only graphical innovations sell. If you switched out the graphical power of SC2 for a game with the diplomacy of CK2 but with a stat system more in line with modern RPG complexity as opposed to the crappy 5 stat 1999 DnD crap CK2 uses, and the city building power of an Impressions game, and stuck that on a WCB3 or Kohan base, with better AI and formations/squads/regiments/legions as icing it wouldn't sell because the graphics would have to be Kohan or WBC3 level to account for all the awesome features. Or god forbid Achron level. Achron was pure genius but it didn't sell because the graphics were terrible. Well marketing budget also, but the most common complaint even among fans is the graphics. They even fixed a lot of the pathfiding issue.
I mean come on, how can people dislike a game where you can blow all your chronoenergy just to time travel a Gargantuan back 5-10 minutes and steal all the resource boxes at your enemies main base while the whole time they thought they were kicking your ass, causing their entire economy to implode?
Graphics. The bane of awesome video games since, well, forever.
I think Starcraft Brood War is the best RTS ever made.
Total War doesn't have tons of units.
The typical battle in Shogun 2 is less than 60 units. This holds AI wise, where each regiment is a cohesive individual that is ordered and reacts en mass, and graphics wise, where each regiment is comprised of identical copies.
There are functions where the system usage is more detailed, such as the choreography of the individuals during combat, but utilization is massively lower than it would be were the individuals actual individuals. The hottest processor out would bleed rectally before your units even engaged.
Total War can also get away with significantly higher resource consumption because it has a rather weighty load time before every combat. Were it a standard RTS like Starcraft, the graphical fidelity of the system would be a small fraction of what it is. Terrain.pack is nearly 4GB just on it's own.
So, guys, if hardware isn't the limitation, why aren't you making the next gen RTS? Obviously you're in the business of making games or you would be ignorant of the actual challenges while the OP is deeply involved in them daily. So get to work.
I agree with most of what you are saying, but it works both ways. How many folks here upgraded their hardware just so they could play Crysis at the highest settings? No one had the computer power to play that game when it first came out. Newegg owes the Crysis team some royalties for all the business generated by that game.
Someone who is willing to take a risk and release a game that requires a 64 bit OS, machine and 4-core processor could very well push the community to upgrade. But it would have to be a ground breaking game.
Hardware may drive game development, but IMO, games drive hardware adoption.
I've never been a fan of real time strategy. There was a time when there seemed to be tons and tons of these RTS games coming out and I was begging for a TBS.
Crytek isn't that successful either. They're well managed and turn a profit, but the difference between an absurdly cutting edge engine and being more intelligent about resource utilization is exponential. Just from retail, Half-life 2 outsells it 4-1, and Valve wont even tell us how loaded they get from Steam sales.
One is an institution, the other a benchmark. Benchmarks aren't the target audience.
When Half-Life 1 was released it was a benchmark. When Half-life 2 was released it was a benchmark as well.
Asking Stardock to make an RTS is like asking a pizza place to make Chinese. It requires a different skillset.
Stardock has their hands full making TBS right now.
So are we supposed to take this article as an assurance, that SoaSE 2 will eventually appear, but only when the hardware would allow for it (or in other words when majority of players will own quadcore, 64bit DX11 systems)?
Anyway, StarCraft 2 should not be counted as a part of that second generation, as it its basically the same game as the first one. It does not try to modify the basic gameplay with elements from other genres, it is small scale and lacks the strategic zoom altogether (on purpose) and despite being 3D, the animation of units is so bad, it looks like sprite based game in motion (in order to provide "instant" control over your units, therefore again on purpose).
Have you played Anno 2070? Big map, great graphics, great detail. Very hard on most computers though.
There are many great features available to you once you register, including:
Sign in or Create Account