Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 25, 2014
arrowPress Releases
October 25, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

Opinion: The Evolution of the Modern RTS
Opinion: The Evolution of the Modern RTS
April 21, 2008 | By James Lantz

April 21, 2008 | By James Lantz
More: Console/PC

[Despite commercial and critical success, THQ's RTSes, Dawn of War and Company of Heroes, haven't reached Starcraft's level of competitive play. Designer and game student James Lantz investigates whether modern RTS innovations made the genre unsuitable for high-level play.]

After the huge success of Starcraft and the large success of Warcraft 3, Blizzard stepped off the RTS stage and let THQ nudge their way into the spotlight with Dawn of War and Company of Heroes.

But, despite their commercial and critical successes, neither franchise could hold a flame to Starcraft’s ability to produce and maintain competitive play at a high level.

Right now, Blizzard is probably asking the same question we’re about to investigate: what made Starcraft a huge competitive success while Dawn of War and Company of Heroes have a comparatively piddling competitive fan base? And are Dawn of War and Company of Heroes really an example of where the RTS genre is headed?

The first, and largely underestimated, strength of Starcraft is its visual clarity. In Starcraft, you can glance at the minimap and understand almost immediately where your forces are concentrated, or look at the screen and quickly understand the flow of the battle. In Company of Heroes, however, the visual clarity gets lost amongst all the majestic effects that make the game so beautiful.

At first, this seems like a trivial difference. However, visual clarity is crucial when making the split second decisions that can steal victory from looming defeat.

There are easily half a dozen times in any given Company of Heroes game where you must make a quick decision, usually about retreating, amidst huge clouds of smoke and severed limbs. Looking back on the replay, you’ll often have made the wrong decision when, if the circumstances were clear, the decision would have been much easier.

I have absolutley no idea what's going on here. It's pretty, though. Starcraft’s second strength is the precise level of control it gives the player. Many casual RTS players see the gradual curtailment of micromanagement as a boon to the genre. However, the more complex AI comes at the cost of precise control at a higher level of play.

As Starcraft evolved, competitive players tried to balance micromanagement with overall strategy and unit production, allowing them to hone a skill as well as push strategic innovations.

In Company of Heroes, however, the unit AI is complex and difficult to control manually. When you try to run a unit back, or micromanage it around to flank, you’ll often find it moving slowly because it insists on running from cover to cover. In buildings, even anti-armor units will often choose the wrong windows to fire from and get slaughtered by circling vehicles, which they could easily have killed if the player could simply tell them where to fire their rockets.

It’s also rather popular to praise Company of Heroes and Dawn of War for removing the micromanagement involved in resource gathering. In Starcraft, however, one of the most interesting choices that defines the course of a game is when to expand and when not to expand, a choice that Company of Heroes and Dawn of War simply remove.

With the same stroke, Company of Heroes and Dawn of War also remove another layer of micromanagement (worker micromanagement plays a large role in Starcraft) and, consequently, another layer of skill.

Some might say, “Isn’t that just purely physical skill?” Well, yes, but the balance between physical skill and strategic skill is part of any RTS – just as the balance between how quickly you can target someone’s head and tactical knowledge is part of any FPS. If there were no micromanagement in real time strategy games then they’d just be turn based strategy games, and we already have a genre for those.

But Company of Heroes’ largest weakness is its randomness. Starcraft has very little randomness, and so the same ambush in the same place will almost always kill the same amount of units with the same amount of shots. In Company of Heroes, sometimes it takes two rockets to take out a tank, sometimes the first four miss and it gets off unscathed. Sometimes a grenade takes out an entire MG squad, sometimes it doesn’t kill anyone.

To most competitive players, this is unacceptable. If the entire course of the game can be changed by a bad roll of the dice, there is no point in learning the subtleties of the game that competitive players use to get slight edges over each other.

In Starcraft, whether or not your dragoons are positioned in such a way that they get the optimal number of shots off is largely inconsequential in a game between low level players, but it’s crucial at the pro level.

However, almost everything in Company of Heroes has a random number generator attached to it. This lends it a sense of realism and tension (you never know what’s going to happen!) but severely limits high level play.

This game is so pretty. Among random number generators and confusing visual effects, the twin evolutions Company of Heroes and Dawn of War have proved another interesting, subtle point: there is such thing as AI that is too intelligent.

The more intelligent AI is, the less control the player has: when your units scatter and take cover at the sight of an artillery blast, they’ll scatter in unpredictable directions, and sometimes in ways you didn’t mean for them to go at all (like into a tank).

Moreover, the less control the player has over individual units, the less player skill factors into a result. When an AI is dumb and predictable, the player knows exactly what will happen in any given situation and can use this knowledge to pull off difficult and impressive stunts.

However, when the AI becomes unpredictable and intelligent, the player loses that level of precise control over the game, making it a frustrating, slippery and often unintuitive mess, ironically the very thing that intelligent AI was supposed to safeguard against.

So where does THQ go from here? Both Company of Heroes and Dawn of War are innovative, intelligent RTSes. Yet, as both games have evolved, it’s become clear that the elements that make them so cinematic – randomness, visual confusion, lack of precise control – are the same elements that make them unsuitable for high-level play.

The basic mechanics of reinforcement and capturing strategic points are interesting, but ultimately these RTSes need to turn their focus from cinema to gameplay if they want to become competitively successful.

In the long run, Company of Heroes and Dawn of War don’t support high level play in the same way that Starcraft does, even though they are often hailed as the pioneers in the RTS genre and true examples of a “modern RTS.”

Both Company of Heroes and Dawn of War are brilliant cinematic experiences and excellent single player games, but it’s going to take innovation in a completely different direction to compete with Starcraft’s competitive multiplayer juggernaut.

Related Jobs

Giant Sparrow
Giant Sparrow — Playa Vista, California, United States

Lead Artist
Digital Extremes
Digital Extremes — LONDON, Ontario, Canada

Digital Extremes
Digital Extremes — London, Ontario, Canada

Generalist Programmers
Petroglyph Games
Petroglyph Games — Las Vegas, Nevada, United States

Unity Engineer


Andrew Dobbs
profile image
Great article, James. I think more and more developers are recognizing the need for visual clarity with the success of games like TF2. Unfortunately, it seems like many developers believe randomness is unequivocally "good" without taking into account how it can affect the multiplayer experience. As far as precision of control, you can actually see the difference between Warcraft III and StarCraft, which has a lot to do with StarCraft's success as an esport.

Vicente Cartas Espinel
profile image
I think that the success of Warcraft and Starcraft doesn't lie in the game mechanics but in the multiplayer implementation. has been since the start much stronger, stable and full of features than Relic Online. click and play in less than 30 seconds hasn't happened in Relic Online at all: when CoH launched it wasn't strange waiting more than 20 mins to find a match. That didn't happen at W3:RoC launch (although probably Blizzard sold much more than Relic).

The flawed online killed the high level play. I think that now Relic has all the keys (good game mechanics, networking experience,...) to make their next game (DoW2) really competitive in multiplayer.

Steven An
profile image
Great article. CoH and DoW definitely have a different feel compared to SC. SC is so solid and crisp, whereas CoH and DoW, while they can be fun, feel a lot more fuzzy. Your article articulately explains those vague adjectives! Kudos to you, James. I'd love to see more articles like this.

I think another important factor is just cultural inertia. Can you imagine introducing a new kind of basketball and expecting it to take off? People don't want to spend precious time getting good at a different RTS that may not take off professionally. And nor do fans want to re-learn the intricacies of another RTS so they can enjoy watching the matches (even to enjoy StarCraft matches, you need to know a lot about the subtleties of the game).

But I think Relic is doing just fine making commercially successful RTS's. Their RTS's don't demand the large amounts of dedication needed to really get into StarCraft (especially these days, when everyone playing is ridiculously good).

I do hope that Blizzard is paying attention to these issues you bring up. Because if anything will displace SC, it's gonna be SC2. That's something people can get behind, and if they do a good job with the game and make sure to address everything you mentioned, we'll finally see professional RTS go 3D - that would make for much cooler spectating!

Alun Rees
profile image
I agree in most of the points, having been a semi-pro SC player myself (Quite a few years ago, of course.). I mainly disagree, however, with the "visual clarity" issue.

Once you start playing at pro or semi-pro level, the amount of effects, clutter and sheer confusion in the screen no longer affects you in any way. What is confusing to casual players is simple and deterministic. In the same way that a high level player in Starcraft can micro-manage individual zerglings out of a horde of one hundred while under fire from dozens of Interceptors flying in and out of Carriers, the same skillful player can definetely make out every important detail out of any one second of game in DoW or CoH.

The other points, however, are extremelly valid:

A pro player simply will not stand for randomness. W3 was vastly inferior to Starcraft when it comes to professional competition, and the reason for that lies squarelly on damage randomness. In DoW and CoH, randomness took an extra step with random morale, random hit chances, random abilities, random untriggered special attacks...

The AI does hinder high level play a fair bit. I do think, however, that this is something that pro players can make up for. SC pro players have learned to compensate for the fact that their dragoons will bump into each other stupidly and walk around in circles instead of moving in the direction indicated. The same pro player could micro the heck out of a squad of soldiers to override this AI to a large degree. This, of course, is less so in CoH.

Things that were touched on but not sufficiently so in the article were the "Macro" game, and the Squad functionality itself.

In SC, from the very first second of game you have to control individual builder units to maximize resource output, and from that point until the end of the game, you will continue to build them and pay them more than a fair bit of attention. W3 and later DoW/CoH have removed this entirely. There is a set amount of builders that are useful, and once set to their tasks, you can pretty much forget about them. Very often, in SC, high level players can beat casual players without using a single fancy trick, but simply by playing the Macro game in a far superior way (and thus outproducing the standard player). In DoW, W3 and CoH, the resource production gap between a pro player and any moderately experience one is almost nil.

Being Squad-based means that you cannot control each individual unit to maximize damage output and "dance" them individually to minimize the amount of dead units (Which means each unit stays alive, and thus dealing damage, longer). This, once again, means that a moderately experience player "dancing" his squad will probably be almost as effective at it as a pro player doing the same.

However, it is worth noting that I do not believe all these changes are for worse. They made for a better game to play when all you want is fun (As opposed to a sport). Playing Starcraft is, pure and simply, exhausting. I have often found myself on the breaking point after a trio of 10-minute matches. With DoW and CoH, I could play either of them for hours without breaking a sweat. I believe to a large degree that must have been the designer's intention.

Max Haider
profile image
The game Dawn of War is based on, WH40k, has plenty of randomness, and yet people go to those tournaments and write enormous tactical articles.

You never know if a particular shot in basketball is going to go in. Somehow there's still a high level of competition in the NCAA and NBA.

Football coaches have no direct control over how their players execute a play, and yet they manage to get called geniuses sometimes.

CEOs can directly control anyone manning the phone lines or working the assembly lines for their businesses. A natural disaster might wipe out a factory at any time. Still, they get offered millions of dollars to do what they do.

The expectation to control every detail results in more of what I'd consider low level play than "high level" play. If I watch a competition, I want to see big decisions getting made, not micromanagement. I like observing competitors who can understand the odds and overcome, instead of quitting because something is left to chance. The idea that an RTS without physical skill is just a TBS is ridiculous. It's real time because people get to simultaneously make decisions with a time limit. A real "real time" war game would have troop deployments and resource acquisition take months or years. You'd sometimes have days to make strategic decisions. And you'd never tell any individual troop which window to shoot from. The physical skill of the general doesn't matter at all, and yet it's all happening in real time.

I'm having a very hard time figuring out what the author means by "high level play". As best as I can tell, the author means game play that is attractive to the sort of person that would practice a game to the point of making a livelihood off playing it. (Which is, unfortunately, extremely different from high-level brain function or high-level decisions). If that's what high level play is, then I hope Starcraft reigns forever as the last RTS with high level play. Any game that caters to such play is going to be much less enjoyable for me than a game that actually makes decent use of the random number generator and streamlines or removes uninteresting things like worker allocation. Aren't computer games supposed to be fun? Should an RTS sacrifice its appeal to a simulation or a casual gamer to please the small segment of the population that wants to earn a living off their tremendous clicking speed?

Jay Wilson
profile image
I was the lead designer on Dawn of War, so I have a few thoughts on the subject.

Alun Rees's comment was dead on:

"Playing Starcraft is, pure and simply, exhausting. I have often found myself on the breaking point after a trio of 10-minute matches. With DoW and CoH, I could play either of them for hours without breaking a sweat. I believe to a large degree that must have been the designer's intention."

Exactly right. :)

The design philosophy of Dawn of War was aimed at making a more casual, more fun, less eSports RTS. I personally thought it was foolish to try and outdo Starcraft, because even if you make a better Starcraft, who cares because Starcraft is awesome and no substitute will do! I had no illusions that we could take away Blizzard's audience, I wanted us to find our own audience.

We didn't outsell Blizzard, but then who does? We did make a very successful, well reviewed, and well liked game. That's not too shabby.

That being said I think Dawn of War and Company of Heroes have their place in serious competition. Many professionally played games have elements of luck and chance. Poker, for example. But it's not what competitive video gamers are primarily focused on, with good reason.

Video game competition is a fairly new thing, with a ridiculous number of games to choose from. It's natural for competitors to focus on fairness and high degree of skill when choosing the games that define the competitive market.

On this front I think this article is dead on. To play Starcraft at a competitive level arguably requires more skill than any other RTS. And that makes it the perfect competitive RTS.

It also makes it extremely intimidating to casual users who want a fun RTS to play, which is what Dawn of War was aimed at.

Erick Nivel
profile image
@ Alun Rees

Sorry to burst your bubble, but you were never anything at StarCraft. Semi-pro and Professional status come only from those living in South Korea, and only a few have been there (none of them being you). Nice faking though, everyone thought you were so cool.

@ James

Great article, you did many communities proud by bringing this to light in such a professional writing. Cheers.

Erick Nivel
profile image
To add a slight edit to the last; only a few non-Koreans have ever been there as semi-pros/pros.

Alan Zhan
profile image
Great article. You really captured how Starcraft has developed. Professional, detailed, and accurate.

Andy Davis
profile image
I feel I have to disagree with your assertion that Company of Heroes is inherently less capable of supporting high-level competitive play than StarCraft. To begin, I must say that, as an experienced CoH (and SC) player, CoH has a rather steeper learning curve, and this can warp the observations of a player who is not experienced with CoH.

For example, you make the point that CoH lacks the visual clarity of SC. On one level, you are correct. SC has a very clear, succinct interface with little clutter and few distractions. However, once a player becomes familiar with the tactical map, unit symbols, etc., I feel that the CoH interface offers just as much, if not more, useful information to the player. Specifically, although there are explosions going off, and smoke covering the battlefield, the unit symbols are always clearly visible, as they are always on top. A player has to actually try to lose track of a unit in a battle.

To address your point about the micromanagement involved with resource gathering, I would say that, essentially, you are comparing apples and oranges. CoH does indeed have a resource system, based on map control. For those who aren't familiar with the game, let me explain. Each map is divided into sectors, which can be low/medium/high resource sectors, provided either fuel or munitions (manpower, the third resource, is primarily produced by the HQ, so your manpower income is nearly static).

There are several parallels that can be drawn between these two systems. With the SC model, a player can certainly choose to gather one resource at the expense of all others. There is a similar choice made in CoH, although it is more a decision of degrees rather than absolutes. The player can choose to prioritize fuel over munitions, or vice versa, depending on the strategy being pursued. I would also say that the absence of worker micromanagement in CoH is offset by the constant amount of "raiding" on the opponent's territory. Most, if not all, of the battles in a CoH game are fought over controlling a particular resource sector, or group of them. Harassment, which is almost nonexistent in SC, is nearly constant in CoH.

The randomness factor in CoH is hard to quantify, and even more difficult to explain to someone unfamiliar with the game. It's certainly true that you can lose an engagement because of an unlucky (or lucky, depending on which side you're on) grenade, or lose a tank retreating because it's engine was destroyed. It's true that the randomness inherent in the game slightly diminishes the players' control over the game. However, it's important to remember that this affects all players equally. You may lose the first skirmish due to an unlucky grenade, and win the second for the exact same reason.

Also, the randomness emphasizes something that SC does not: immediate adaptation. "Oh crap, my armored car just hit a mine I didn't know was there while trying to escape, I need to bring up a G43 squad to slow the Rangers pursuing it so I can get it out safely and repair it." These kinds of situations occur every single game, and I would say that it is a player's response to these types of situations that determine which players are top notch, and which are just good. So, the randomness actually separates the chaff from the wheat, as it were.

The AI in CoH, specifically the infantry AI, can do some pretty wonky things. It's important to remember that the AI makes decisions for each man individually, to anywhere from 2-6 individuals are nominally acting as a unit. Bugs (of which there are a few) aside, the infantry AI is fairly predictable, though. Of course, a player can make use of this more realistic AI. If you know that dropping a mortar shell near them will make them slow their fire and drop for cover, executing a mortar drop while taking advantage of their reaction to charge them with an SMG squad is a perfectly legitimate (and useful) strategy. The more familiar a player becomes with the game, the more successfully they can predict, adapt to, and make use of the AI.

To be fair, SC and CoH are two very different games, with two very different design philosophies. I personally prefer CoH, but I certainly realize that SC has enjoyed much better support and, to a lesser degree, better balance.

profile image
@ Andy Davis: I'm sorry, but it's very obvious you have little experience with Starcraft. :) Harassment is at least as important in Starcraft as it is in CoH. Ever heard of mutalisks?

Dan Gray
profile image
I have to say I found the comments much more interesting than the article.

It has always seemed entirely obvious to me that neither DoW or CoH were intended to be hardcore competitive titles. In the same way as all the other non-competitive RTS titles since the beginning of the genre, the focus was on fun over competition. The author seems to be under the impression that RTS titles preceding and including Starcraft were ultra-competitive.

@Jay - Thanks for taking the time to post. :)

Sam Fold
profile image
Have to say I strongly disagreed with this article, namely because I was involved for years in another competitive RTS which had no base building and a decent degree of randomness: Myth 2: Soulblighter. The second RTS made by Bungie software before they became very famous with Halo.

For a game with only moderate commercial success Myth 2 had a competitive multiplayer community for years after release, even after Bungie stopped supporting the game, the community carried on with Tournaments. In fact, there are still tournaments being played now, more than twelve years after the release. Sure it's not as huge as Starcraft but I think it would be considerably larger if Bungie still supported the game as Blizzard does with SC.

Specifically in the Myth games, each person in multiplayer starts with a number of points with which they can buy units. Different units are worth different numbers of points. So you could start off a match with fewer more powerful units or a large bunch of weaker ones. You have around 30secs to a minute to pick your team then the game starts. You cannot get any new units once a game starts.

There is zero base building, whether you win or not depends on your skill at micro-ing those 10-30 units. Due to the game design, terrain makes a huge difference. If your archers are on a hill they have a big advantage, similarly, if you have dwarves and are fighting in water, their molotov cocktails are useless unless you are skilled enough to hit the enemy directly in the body of a unit.

This is where I think Starcraft is lacking, it's always either 1 or 0. There are very few degrees of variation which made the game somewhat lacking for me.

In Myth, one of the ways in which it was very advanced for the time was the use of a full 3D engine with realistic physics. ie: you could walk your archers half way up a hill and their arrows would correspondingly go further, at the top of the hill they would be even more effective. The arrows themselves would move faster when fired with the help of gravity, given you a massive advantage over an opposing player who was at the bottom of the hill.

Given that each object in the game was modeled with realistic physics and you could also aim where you wanted, it was possible to try and predict an opponents movements aim your archers or dwarves to fire at the ground where you thought an opponent might move to. This led to archer battles where players would move individual archers to avoid arrows and purposefully aim near, but not directly at an opponents units to encourage them to move in a certain direction for a tactical purpose.

So, even though you could superficially say, "oh there's no base building so less skill and technique is needed". It's nowhere near as simplistic as that.

Also, in Myth 2, there is randomness and it's a canard to say that this somehow rewards less skilled players. In the end, the best players always win. Randomness makes each game as a whole less predictable and more fun. The way SC works, if a player shot a basket, it would go in the hoop, which would make basketball a whole lot less interesting.

Basically SC makes up for the fact that it doesn't use realistic physics or a fully 3D world and zero randomness by adding complexity through base-building and many different unit types and abilities. Just because it's different it doesn't make it any better or worse, it's just personal preference. I never got into SC multi much because I don't find base management or base building fun at all. I loved Myth because it was all about how effective you were at pure battle skills and controlling each unit as well as possible.

One final example: There was one particular game-type in Myth called Assassin. Each player had one or more units which were "Assassin" targets and the object of the game was to kill as many assassin targets as possible in the time limit.

Now on one particular multiplayer level, the assassin targets were molotov cocktail throwing dwarves. When thrown and detonated, a molotov pretty much instantly kills a dwarf. However, this particular map was mostly set on a shallow lake with an island in the centre. Each player usually took 6 dwarves only in the unit trade (only other unit allowed was a useless slow melee unit) . This map/gametype because a classic because of the pure micro skill needed to take your six dwarves and try and kill as many opposition dwarves in the time limit. Mainly because you needed to aim and control your dwarves perfectly as the cocktails would extinguish if falling in the water (and the map was 75% water). Even with the good amount of randomness in Myth, a good player would win or come second in this map every time. However these games could turn into strategic marathons lasting 20+ mins, even though each player only had 6 units.