A series of interesting decisions

EverythingElsePosted by Eskil Sat, March 21, 2015 03:58:20
While testing the micro design of my new RTS game EXO I'm currently designing the macro game, and I have been considering a number of different design concepts.

To begin with, since I am making a strategy game, it may make sense to define "strategy". Its easy to think that just because you move around troops from a top down prospective you are dealing with strategy, but in my definition strategy is not combat, it is setting something up that gives you an advantage later. This definition creates two core differentiators when designing a strategic element, the first being pacing. There needs to be enough time to take a strategic decision and then some time before it pays off. The second observation is that Strategy doesn't decide the outcome, it mealy tilts the odds, while action is the final decider.

Judging by this definition EXO is in its current state almost entirely is an action game and not a strategy game. This is why I'm considering how to add a new layer of strategy to the game, and in this article I will outline some of my intellectual work on the subject.

The athlete and the coach.

What is the most accessible sport, running or a card game with a 1000 different cards? Well most people can run, so in a sense running is a game that almost everyone can do. Its a simple and elegant game. A card game with a 1000 cards is hard to learn and will take a lot of time to get in to. On the other hand, if you start training to become a runner, your running wont improve fast, and your training will also provide diminishing returns as you get better. No matter how much you train you are very unlikely to ever be good enough to compete in the Olympics because very few people has the DNA required to do so. The card game on the other hand wont be limited in such a way. If you learn a single card every day then 3 years later you will have mastered the game.

Knowledge is easier to gain then skill. This is why MOBASs with hundreds of Heroes and equal numbers of items don't feel as intimidating as one may think. While playing Quake is easy to pick up, it still seams insurmountable to most people. Creating knowledge complexity is therefor a good thing, because it allows anyone to claim ownership of the game by gaining knowledge about it.

When a sports team does poorly the question of replacing the coach inevitably gets raised long before anyone talks about booting out the players. Why is this? Its because the Coach has knowledge not skill, and everybody things they have or at least can have knowledge too. It is clear to most of us that we will never be able to kick, run or throw like a professional athlete, so when they fail we tend to be forgiving, but we all think we can do the job of the coach since it is purely intellectual. If a coach has to decide what player switch out, its a decision any of us can make, so when the coaches gets it wrong its easy to label them incompetent, no matter how many factors they considered internally.

The coach is the spectators proxy, we all imagine ourselves being the coach when we watch a game, because they make the same decisions as we do in our heads. A game like 100 meter dash is very hard to have opinions about because its all just about athleticism. Our only advice to the participants would be "Run really fast!". As soon as we move up to longer distance running we can start having opinions about strategy of when save your strength and when to go for it. If you are designing a game for spectators, it makes sense to create a large space for audience participation by making the type of decisions made by a coach a large portion of the game.

A turn based game, forgoes the "athlete" entirely, and is therefor a much more comfortable and less stressful experience. But it also removes the ability to in a dramatic manor beat the odds, by executing a perfect play. Just as there is value in a game allowing for a player to carefully set up a trap, there is a value in allowing a player to think on their feet and improvise themselves out of a tricky situation. This is the balance between the coach and the athlete.

Possibility space.

One of the core things that you are looking for as a designer of any kind of game, is to create as many possible outcomes with as few rules as possible. Chess being the obvious example of a game that does this brilliantly. You can teach the rules of chess in a few minutes, but the possibility space of those rules can occupy a lifetime. Learning the game in chess is not about knowing what the peaces can do, but to see the possibilities afforded by what they can do. This makes the game both easy to get in to and hard to master. Ideally you want the spectator to instantly see the genius in the masters moves.

The way that i prefer to create a large possibility space is create many interlocking systems. If you have a game with 10 weapons, you have 10 choices, but if you have 5 weapon's and 5 armour sets, you have 25 combinations to chose from. Rather then increasing the length of the array of options you increase the number of dimensions of options. This also makes the game easier to learn. Even though you have 25 combinations of armour and weapons the player still only have to learn 5 weapons and 5 sets of armour.


How do one dividing the game in to segments of action and strategy? Strategy by its nature must be made over a longer time period, and should not immediately pay off. If you instantly get feedback from a strategic decision and the instantly can change your decision in reaction to the feedback then the decision ceases to be strategic. (This delay also often cause another design problem since it often makes the cause and effect less obvious)

Many strategy games divide their games in to early, mid and late game and gate abilities in such a way that it makes sense to wait to first tech up and then later attack. Early defence is deliberately stronger then early attacking units to slow down the pace of the game so that strategic decisions can be made. I kind of dislike this structure because it forces the game to play out in a specific order, I also don't like that many strategy game tuns in to 20 minutes of building, 30 seconds or fighting, the end.

Some games like MOBAS allow you to make important strategic decisions before the game starts (The pick and ban) and while this creates a lot of strategy it limits the possibility space, since players can change these decisions later in the game. The ideal game should allow you take multiple corrective decisions over the course of a match. MOBAs also control pacing using towers and other structures that are overpowered until the players has taken considerable time to level up. I find this a bit too rigid for my tastes.

Starcraft on the other hand is famous for its rushes, and Strctaft also allows you to make tech switches at any point, but it still limits you to a single race in game (Yes i know that Zerg can build Terran and Protoss units too). Early in in EXOs development i considered having multiple races in the game, but i now see that it is an ineffective way to create a large possibility space as choosing your race limits the combinations of units you can use. I find that its a missed opportunity that the most important strategic decision you make is a decision you make only once, as most players stick with one race for all matches.

These are games where multiple matches are stringed together so that players can redo their strategic choices between individual games. Counter-stirkes economy, or the limited supply found in Due process are excellent examples of this. I'm considering a tournament mode like this for EXO. Its understandable that its convenient to stop the action an let players think threw their strategies before the action restarts, yet it would still be better if decisions could be taken at any time.

To do this the games pace needs to fluctuate and at least partly be controlled by the players. Preferably a game should naturally have peeks of intense action and valleys of calm where the player can take the time to make more strategic decisions.

One of the core experiences of playing a strategic game is being able to think about it when you are not playing it. Your best strategies will be devised while in the shower, or in bed or while day dreaming at school or work. I think anyone who has ever been in to a strategic game has had the urge to play it just to try out some new strategy. The problem with this is that you don't want players to be able to always execute a strategy they have planed out before hand. The game becomes stale and not very exciting if players keep executing the same builds over and over again and again. If different builds counter each other too sharply, the player goes in to the game with one build, and fate decides if the opponent has chosen a build that is either strong or week against what the player is doing. The game turns in to Rock paper scissors and that is not a very interesting game. On the opposite spectrum you have card games where each game is played entirely differently depending on what cards are dealt. Ideally you want something in between where players can think up strategies outside the game, but where the game wont always be conducive to the execution of every strategy.


Balance is obviously very important for a competitive game, but i have started to think that rather then seeing balance as something good, maybe it is imbalance that is bad. Having a perfectly balanced game doesn't give you anything, its just that not having your game balanced will ruin it. Where as designers try to design games that have as many viable strategies as possible, players are trying to figure out a way to break the game by finding a single strategy so good it makes all other option worthless.

In a game like Starcraft with 3 different races almost all of the designers energy gets dedicated to keeping the game balanced. The designers cant just come up with a new unit idea and throw in in to they game, they carefully have to re-balance the entire game to take the new unit or feature in to account. To me it seems very inefficient to have a system where players constantly try to break the game and forcing the developer to constantly try to fix it. Therefor I think its important to build in to the game some sort of self balancing core mechanic that lets the designers be more creative.

A typical example of a self balancing system is where you divide a cake by having one person divide the cake and the other choosing one of the two peaces. MOBAs that employ a pick and ban system remove the spikes of over powered units, but does little to promote under powered units. A market based system where prices for less used items drop while the popular stuff rises would be even better.

I think its important to try to make player centric balance rather then opponent based balance. In Starcraft the Stim is an important upgrade for Marines in order to be able to counter speed Banelings. Stim makes the the marines able to go toe to toe with the Banelings, and is therefore balanced against Zerg, but having Stim is not at all balanced against not having it. There is no question IF the player should get stim, only how soon he or she should get it. If Stim was expensive enough or negated some other possible upgrade, the choice to upgrade Stim would be much more interesting.

I prefer strategic tech decisions that are different or temporary rather then just better. Lets consider a game with 3 different sets of armor. A traditional way to design it is to have armour level 1, armour level 2 and then finally armour level 3. The problem with this is that there are no interesting decisions for the player to make. The story in the game is already set. A more interesting way to design it, would be to have one anti fire armor set, one anti poisoning armor and a magic set or armour that makes you invulnerable for 10 seconds once a minute. Now the player has some interesting decisions to make and it makes sense to switch back and forth between the different armour sets many times as the game progresses.

Where I am.

Right now I'm considering a system where the map has a bunch of resource nodes, that are all "Plugged" so that the resources inside are inaccessible. The players can set a number of units to pop the cap of the resource. This requires some units to be present for a time. Once the node is opened, it starts giving the holder resources. The opening of the node gets announced to the other player, and by capturing the node once it is opened an opponent can steal the resource. Since the resource only flows for a short period, it creates a temporary focus point of the game. Once the resource is tapped out the players will move on to harvest another node or attack.

My current idea is to have three or four different resource types, and have each upgrade require one or two of the resource types. By making the resource type provided by the resource node unknown until the player has "popped the cap", I force the players to adapt their strategy to the resources dealt to them. I think this could be a good middle ground where builds can be pre-planed depending on different resource combinations, but where the player never can be sure what a specific game will bring in terms of access to the tech tree. I'm considering have the price of different tech options dynamically fluctuate in price depending on their popularity.

I want to thank Chris Thursten, Mahmud Din and Richard Lemarchand for their input.

  • Comments(6)

Fill in only if you are not real

The following XHTML tags are allowed: <b>, <br/>, <em>, <i>, <strong>, <u>. CSS styles and Javascript are not permitted.
Posted by theGiallo Sat, January 07, 2017 19:22:35

Hi there, I know this is an old post, but I feel the topic is not. In the last months I've being thinking about strategy games and I want to share a bit of my thoughts with you, hoping this can interest you or at least help my brain on the reasoning.

Thinking about RTS games, pretty much every game I can think of has very little strategic component and is mostly a tactic game. I see this dichotomy in what is usually called micro and macro. These games are full of micro and lack macro, and to me strategy happens at macro level, while micro is execution and tactic.

I think strategic decisions can be made on a very complex system, one that is not humanly solvable, but that can be "felt", i.e. on which hypothesis can be formulated and validated, with not 100% confidence, on which our intuitive brain can work and give us that "feeling" that leads our decisions. Making such a system is very difficult. The risks are the 2 extremes. One can build a system too simple, so that it is solved and playing becomes the sole execution of the solution. On the other opposite one can build a system so complex that no one can grasp a thing on it and playing becomes a meaningless series or random decisions with random outcomes.

Of the two extremes the former is better, because still the execution has to be performed. In fact I see here the cause of those games being focused on micro. Once you have a solved problem making your players perform the solution per se is not interesting, but making it harder can generate interest. An example of this can be seen in speed-chess, in which the time pressure limits the reasoning and shifts the play on a intuitive level, coming up beside the "feel" I was writing about.

Based on all this thoughts I theorized a kind of strategy game without micro, full on macro. Like being a general giving orders looking at a map, or being a king and looking at a really bad map and hearing reports from the fields. To achieve this a great AI effort would be needed. I think this would be very innovative, so innovative many will read this and not go for it. I would really like to research on this kind of game and attempt one implementation, but I would also like to play one, so I hope someone will try.


Aside from this: I like the idea of a self balancing cost. A way to obtain this can be using an auction, like in this MTG format .

Posted by RFDaemoniac Thu, September 10, 2015 07:57:54

think you lose out by having only a variety of choices that are equal. If having stim and not having stim are both equally valuable choices, then there is little to no progression in the game. Unlocking higher tech units, getting additional supply, getting straight buff upgrades all contribute to a dramatic arc in the game. There is a slow careful beginning with some pokes and information gathering. There is a midgame full of attempts to hit each other and to take advantage of mistakes. There is a lategame to conserve your resources and be much more methodical and careful.

The 20 minute buildup to a 30 second battle is certainly a terrible arc, but I think you've missed talking about the value in having a change in pacing and flavor over the course of the game.

Do you want the only arcs in the game to be about gaining a larger number of units? Rather, what else do you see contributing to a single game engagement loop?

Posted by NemesisLeon Tue, April 14, 2015 00:15:47

The Civilization series comes to mind. I've been playing Civilization : Beyond Earth recently, and it's pure long-term strategy. Yes, there are wars, but the long-term planning is most important especially on longer game speeds.

I see strategy as a competitive puzzle. You're trying to design a puzzle too difficult for your opponent to solve, and solving the puzzle your opponent provides means you win.

Also, chess is probably the #1 thing I'd point to for strategic design. It's all competitive puzzle-solving. Every strategy game changes over time.

One other thing to mention is competitive, professional Starcraft players in China or Korea. Starcraft is probably the best balance of action and strategy, and MOBAs are very close to that as well.

Despite the game's structure.... don't forget community reaction. The way people interact with each other. People keep getting over-competitive and hating on each other for "not being good enough"....

Posted by Hypocee Sat, March 28, 2015 22:32:47

Ah. Yes, market forces within a match are an entirely different matter! If your interface and theme fit around them they could be wonderful fun in the vein of M.U.L.E., Offworld Trading Company, Greed Corp, and Rise of Legends.

Posted by Eskil Sat, March 28, 2015 12:02:35

I think the implementation in Counter-strike was poor too. The reasone the everyone buys the Deagle is that ist simply the most powerful. If each pistol offerd a new skill or tactic nudging people to play them wouldent be as bad.

I do think you have a good point about the predicabillity of the game. The thinkg is that if a game is too predictable it will become dull as it plays out the same way every time. It is very much a balancing act.

But would not have to shift one a world wide scale it could be just whitin the game. Then you have way more predictabillity, and you would be able to do things like buying guns to deny the opponent a tactic. A economicly strong T side may buy loads of smokes, to deny the abillity for the CT side to defend a bomb site with smokes.

Posted by Hypocee Fri, March 27, 2015 02:03:38

Nnnnn... the alarm that goes off in my head is that this has been tried on a large scale once, in Counterstrike's dynamic gun pricing, and it was an utter disaster. I know there was a larger discussion on an earlier Crate and Crowbar, but I can't find it. I can only find the summary and reference to it at 34 minutes in .

At the first level of design, another way to describe a balancing price dynamic is "If players really like playing this game in one way for whatever reason, it is my duty to ruin that for them." It suggests that the designer can infallibly predict an "equally enjoyable" (as if that's a useful number) set of stats, timings, art, animations, real-world analogies, and whatever other factors apply for every weapon or tech branch. You know what, people like the Deagle because it goes BOOM and they've bought the Deagle for ten years. The camera in Spelunky can't be "fixed" into use by reducing its price because it doesn't cost too much gold; its cost is displacing something better from your hands. Only actual design work would bring it into rotation. In addition, trying to price things into equal use carries the assumption that the game is at its best when everything is equally common - no pocket builds, no risky combos for experts only, no situational responses.

Another level up, there's the huge hit to your design authority and community goodwill in refusing to promise consistency and knowledge to players. You've written about the attraction of knowledge before. It's the reason patch releases are a big deal in any multiplayer game. Counterstrike's example suggests that if you start essentially pushing hourly patches which are made by a mindless economic robot, a few responses will describe most of your (indie, tiny, fragile) multiplayer community:
1. Try to "fix" it through economic revolt. Organize meetups and spend two hours a day logging throwaway games to drive prices so they can play one "real" match with the values they want or think are "right". Fight your bans and countermeasures every day until you stop it or they get so angry they go to something else.
2. Get depressed. "Oh, the builds I like even though I know they don't necessarily win the most are turned off today, the day when I could have played. Well, I could either break entirely with my habits in order to please a robot that supposedly makes me a better player of this game that exists today, or I could go play any of the million other games or read some Internet. I think I'll check back next week to see if this has passed.' Go to something else, get distracted, never check back.
3. Suffer. Just put their heads down, grit their teeth and pay 5 times as much for the thing they like to do, for reasons they don't really understand or care about. Lose repeatedly, bored and frustrated, to opponents who are bored and frustrated that it's bizarrely easy. Sooner or later go to something else.
4. Do what you want. Parse the entire system to optimize their chance of winning a given match to two decimal points on a given day. Play solely to win at every moment. Either persuade or drive out the rest of your player base. Quickly settle to a stable local maximum in phase space with no fluctuations ever again, because after all you're designing a game that's shallow enough to be fully understood, and well-informed selfish humans in the real world have never been swayed by fashion or pursued economic bubbles.