RTS games: “all about speed”?

I just finished reading this article from “In Machinam” about the Real-time strategy game Starcraft, which states that:

an RTS is about clicking speed, in the same way that an FPS is about how well you can aim; that’s what makes Starcraft different from Civilization.”

While I agree that Starcraft and most of its clones are by and large games of reflex, I don’t believe that all RTS games must be by their very nature: why not make issuing orders cost resources?

I mentioned the “Berkeley Overmind” a while back. The AI is a great technical achievement, but what really makes it unstoppable is an exploit. Its signature unit-type, a sort of flying organic gunship called a “Mutalisk“, was balanced with humans in mind: humans are unable to use the unit nearly as well as the AI, which is quick enough to babysit each individual: dancing it in and out of range of attackers.

The AI thus gets more than its money’s worth each time it purchases one.

Arguably this is how all good players get their edge: producing units costs in-game resources, but issuing orders is free. Units cost a fixed amount of resources, but the value that you get out of them depends on the order they are issued. That orders should be free colours every interaction with the game system: it’s only natural that play should become all about taking advantage of this limitless resource!

But what if issuing orders cost money? In another RTS, Archon, it is possible to change the present by issuing orders to units in the past. This costs “chrono-energy” however, of which the player has a limited, regenerating amount:

This simple mechanic prevents click-spamming, at least when it comes to changing history.

So let us imagine an RTS game where all commands have an associated resource-cost. Players are forced to choose the “if, who, what, when and where” of their commands with great care or risk bankrupting themselves. This brings the emphasis away from the “real-time” part of RTS and towards the “strategy”.

An arbitrary constraint perhaps but games are, by their nature, full of arbitrary constraints. The turn-based strategy game “Solium Infernum“, one of my favourite games of any genre ever, imposes a limit on the number of orders that can be issued per turn, and competitive chess imposes a global time-limit for each player. These limitations only serve to turn the gameplay experience into something much finer, with more interesting trade-offs to consider.

Similarly the passage from turn-based to real-time doesn’t have to ruin any strategic gameplay that might have existed. Unless, of course, spamming orders costs you nothing ;)

That RTS games should change is a matter of opinion, but do you think imposing a cost on order would actually change anything? If not, why?

9 thoughts on “RTS games: “all about speed”?

  1. Freelancer Epic

    I often see complex game design being very simple games with different time scales nested within each other.
    Take a team-based action game, à la TF2. In these games you have a millisecond to second gameplay which consists of shooting and dodging attacks, in order to bring individual fights to a positive conclusion (you alive and ProSnyper96 dead). This is the Action Timeframe.
    This is nested within a multiple second timeframe, in which you chose how to navigate a building, try to outflank your opponents, pick your engagements, and manage your ammo in order to have a full gun but not be reloading as an engagement begins. This is the Tactical Timeframe.
    Then you have a minute-scale timeframe where you decide which objectives to push, which weapons to carry and try to group with friendly forces. This would be a Local Strategy Timeframe.

    Most action games end here.

    In strategy games you also have another layer, where you produce units, try to locate enemy troops, attempt to outnumber the enemy and exert domination over key map resources. A Global Strategy Timeframe.

    I think there’s one step higher, which would be the Political Timeframe, where you negotiate war and peace conditions. (Think Civilisation)

    Meaningful gameplay occurs when the player can identify which timeframes he has access to, and what he can do in order to best play each of these mini-games and how they link together.

    In “real” wars, this is also the case, and the chain of command flows in a similar manner. Presidents generally don’t run into buildings with assault rifles, nor do they even tell individual soldiers how to go about an engagement in a battle. They sometimes hit on the Local Strategy Timeframe (think pres Obama overseeing the Navy Seal assault on Osama’s compound), but mostly stay between the Political and Global Strategy timeframes. What presidents do is give orders to generals, who in turn give orders to lieutenants on the field, who in turn give orders to team leaders, who give orders to soldiers, who do their best to not die and eliminate ProSnyper96.

    So give your players a limited amount of orders they can issue, but make it so they can take advantage of military hierarchy. It should cost the same to say “You there! Cross that bridge and shoot the sniper who entrenched on the hill” to say “Hey lieutenant, I really need you to have your guys take control of that city” or even “OK, in this war we’re going to limit civilian casualties, I don’t want the UN bothering us, and I really need France to join us”.

    That way, players will try to limit their involvement to where they are needed. And when a key engagement occurs, a battle outside enemy HQ where a nuke is about to be launched, they might feel the need to micro manage that fight, at the cost of losing sight of the global strategic plan for a while.

    1. Wilbefast

      I completely agree with this “lens” of yours :) It’s interesting that when it comes to games with strategic elements (and I include TF2 in this category) AI systems generally use “hierarchical games” to model the problem. So choosing the next point to capture and avoiding enemy fire are different games played at different levels as part of a global strategy.

      As for the suggestion about control: in Achron units are organised hierarchically, so a single order can be passed down to large number of units from a top-level commander. I’m not sure I’d do exactly the same thing, but the idea would basically be to cap players’ APM to a set value, so sending a group of units across the map and sending a single unit a meter to the left would cost the same amount (1 action).

      I’ve found RTS games that try to fight micro’ by introducing a lot of automation tend not to give you enough control to feel that you’re really in charge. For instance, Globulation 2 has you place attack- or defend-flags to define your global strategy http://www.youtube.com/watch?v=HR0NEttH1Jc but it’s difficult to be much more fine-grain then “everyone attack this now!”.

      A few interesting action games that combine all the various levels of gameplay you talked about: Savage 2 https://www.youtube.com/watch?v=PfVdHg-Sq68 and Natural Selection 2 https://www.youtube.com/watch?v=ZCsgeVfiklw call themselves “RTS-FPS hybrids”. I’ve only seen videos of NS2, but Savage 2 has a top-level commander placing buildings, researching technology and issuing orders to player-controlled “lieutenants” who other players are attached to in a squad-like structure. As you get better at the game you progress from playing only the second-to-second gameplay.

    2. JungleEd

      All the theory seems to be covered between treating time as a resource viewing real time games nested sub games, but how about the inter-game time. Building a great deck wins Magic the Gathering tournaments. Endless hours of study create chess masters. And friends earn you points in Cow Clicker. Each of these involves strategy, money, and time. Not wanting to spend arduous amounts of time leveling up a character or making superficial friends I’ve considered an RPG that circumvents interaction with the real world. Let’s call the game “The Hours”
      1.     The first rule of “The Hours” is don’t talk about real life in “The Hours”
      2.     The game can only be played one hour a week. (Now it’s like network T.V.)
      3.     The game play varies enough to be impossible to prepare for.
      I also don’t enjoy the action timeframe. So I’ve imagined a hybrid between real time and turn based. Imagine an RTS plays for 3 seconds, gives one player a chance to react and then plays for 3 more seconds. Also having a global timer would be nice. I’ve skipped to many dinners over 3 hour Star Craft games.
      I’ve also thought about a hybrid between AI and human play that could eliminate the action timeframe. Start with a game that has a few basic building blocks that can form larger units, add and API, and let the players assign scripts to various objects in real time.

      1. Wilbefast

         Sorry for the late reply (I’ve been at ECAI!) interesting thoughts, especially about this “meta-game” layer on top of all the others. As for the 3 second RTS idea: do you know “Sins of the Solar Empire”? Apparently the time taken to travel between systems (area where tactical battles take place) was intended to create a sort hybrid real-time-turn-based strategy game. Id est you can micromanage orbital engagements, but once units are moving between planets or systems they can’t stop and can’t turn back until they’re arrive. Frozen Synapse is another interesting example.

  2. TopRamen789

    There are a few questions I have..
    Would each unit have their own order cost or is each order cost the same? Also, would the cost of orders scale with the unit selection?
    How much would it cost?

    I think a cost for unit orders would change a lot. In my opinion don’t think it is for the better. Players will want to make their units as efficient as possible and could contribute to ‘deathball’ (a common term used for StarCraft 2 to describe a critical mass of units) strategy. Which, imo, creates a very unsatisfying spectator experience as well as a playing experience. The deathballs clash a couple of times around the map and game over.

    Also, I don’t really see why click-spamming is a problem in RTS games. Yes, if you are a faster more skilled player you’ll probably win (this is why some of us play RTS games, we kind of love the twitchy feel). But decisions do matter in RTS games. A good example is StarCraft 2. Which emphasizes decision making over mechanical ability at a certain point.

    Starcraft 2 is about momentum. Getting ahead just a little bit. Then letting that advantage snowball over time.
    The way that players do that is by playing with strong mechanical ability and making the right decisions. This means good micro/macro, knowing all of the races, making the right units to counter your opponent, optimizing build orders, and learning how to counter build orders (i’m sure there is more i’m leaving out too).

    I suggest watching a Pro Starcraft BW game, and I recommend visiting TeamLiquid.net.

    I think if you watched a game of say… Flash vs Jaedong in SC:BW you might see how it could be difficult to have each unit order cost some resources.

    1. Wilbefast

      The idea is to cap the player’s APM to a fixed maximum value, the cost would be subtracted per action, not per unit. I’d go for something like 20 APM to start with, or 1 action every 3 seconds. If you only have 1 action every 3 seconds you *really* have to think about  what that action will be. However these kinds of things need to be tested and tweaked around during prototyping.
      It’s a good point about “death-ball” degenerate strategy though: there’s
      certain a risk the game would devolve into a case of “throw all units at
      enemy and wait”. I’ll have to think about how this could be avoided…

      What I want to stress is that I’m not trying to make a better Starcraft. I’m trying to make a completely different kind of RTS for those who don’t like “the twitchy feel” and want something more slow and ponderous. I never said there wasn’t an element of strategic thought in the -craft games, but memorising a build order and executing it very quickly generally beats having a genius plan at 20 APM. Hence it’s only when both players have similar speed (ie. pro games) that strategy need come into play. By capping the maximum speed, player are *forced* to compete based on strategy.

  3. diamondleague68

    Civilization actually requires heavy strategy, while Starcraft is a reflex based RTS without much need for actual strategy. 

    If anything, Starcraft and some SC clones in esports ruins the young gamers perspective on strategy, for they do not understand the true meaning of actual strategy.  Even worse, many can’t even distinguish the difference between tactics and strategy.

    Base formula wise; Starcraft = 90% tactics, 10% strategy.  While games like Chess, Civilization, and Total War = 90% strategy, 10% tactics.  In other words, SC (both 1 and 2) is easy mode in the realm of strategy, the “real time” part being more accurate than the “strategy” part of the title genre.

      1. Wilbefast

        I’m not a top-tier Starcraft player, so it’s not really for me to judge. It does seem to depend a lot on properly executing a pre-set “opening book” strategy (ie. build order) while identifier and countering the other player’s, which is rather like Chess in a sense. The late game is often one of attrition and economic superiority, more closely resembling investment-based games like Plants vs. Zombies: bad investments result in a loss of advantage, and the game has a great deal of positive feedback so this tends to lead to a downward spiral.

        Take a look at Extra Creditz “Depth vs. Complexity”: http://www.penny-arcade.com/patv/episode/depth-vs.-complexity
        What is suggested is that the time granularity of the decisions (ie. APM) of a game directly influences the complexity of the decisions that can be made, and the complexity influences the depth of possible strategies. This would suggest that turn-based games do indeed have the potential for a far deeper experience, if they play their cards right.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>