Improving Player Choices

By Tracy Fullerton,Chris Swain,Steve Hoffman

Because it is simply one of the most powerful aspects of fun in gameplay, we need to look more closely at choice as an aspect of fun. What makes a choice interesting versus uninteresting? How can you design choices that are more interesting than not?

One of the most important aspects of choice is consequence. For a game to engage a player's mind, each choice must alter the course of the game. This means the decision has to have both an upside and a downside; the upside being that it advances the player one step closer to victory; and the downside being that it hurts the player's chances of winning. This concept seems simple, but you'd be surprised at how many games force the players to make choices that have no impact upon whether they win or lose.

Remember, the player wants one thing more than anything else, and that is to win. Anything you do that is outside this scope runs the risk of alienating your audience. So when Sid Meier says "interesting choices," what he means is that the game must present a stream of critical decisions that either directly or indirectly impacts the player's ability to win. No matter what you've been told in the past, drama and suspense in games seldom come from the storyline. It comes from the act of making decisions that have weight, and the more weight each decision carries, the more dramatic the game becomes.

As a designer, this is what you must strive for. But how do you make the choices in your game have significance? To start with, let's step back and analyze your game. What type of decisions are your players making? Are those decisions truly meaningful, or are they tangential to the main objective? To help analyze this, we use a tool we call the decision scale, shown in Figure 1.

Figure 1. Decision scale

If there are decisions in your game that seem "inconsequential" or "minor," you have a problem. Go back and rethink the choices you are giving your players. Is there a way to make those choices matter? And if there isn't, those choices need to be eliminated because they aren't adding anything to the game and are probably hurting the experience. Now take a look at the decisions higher up on the diagram. Is there a way to make some of your players' decisions fall into these categories? These are the types of decisions your players want to make.

But, unless your game is an arcade-style shooter, the decisions you ask your players to make shouldn't all be life and death. Nonstop action can get boring too--it's in the breather between waves of enemies that we can appreciate our accomplishments, anticipate the next wave, and steel ourselves for the battle ahead.

In order to create a truly engaging game, you want some peaks and valleys. Let the decisions rise and fall, and as the game progresses, ratchet up the tension by making the decisions gradually more important, until by the climax of the game, everything hangs in the balance.

Types of decisions

It's easy to say that games should have interesting choices, but why is one choice more interesting than another? The answer lies in the type of decision you ask to the player to make. If the player has to choose between two weapons, and one weapon is only slightly superior to the other, even though the player may be faced with a life and death encounter, the decision itself does not reflect this. To make this decision interesting, each weapon must have a dramatically different impact on the player's chance of winning.

But if the decision itself is too easy, then it's not a decision at all. If it's obvious that the player should use the golden arrow to slay the dragon, there's no real choice. Why would the player risk using anything else? This decision, although it appears to be life and death, is meaningless. The player will invariably choose the golden arrow, unless he doesn't know about its powers, and in that case, it's an arbitrary choice, not a decision.

The key to making this decision interesting is for the player to know that the golden arrow is the right choice, but also to know that if he uses the golden arrow now, he won't be able to use it later when he has to fight the evil mage. To make this decision truly dramatic, the player must be put in a position where both paths have consequences. If the player doesn't use the arrow now, his faithful companion, who is not immune to dragon fire, may die during the battle. However, if the player uses the arrow, it will be much harder to destroy the evil mage later on. Suddenly the decision has become more complex, with consequences on both sides of the equation.

Decision types

In the example of the golden arrow, the decision is a combination of the previous decision types. It's an informed decision because the player knows a lot about situation he is in, it's a dramatic decision because the player has an emotional attachment to his faithful companion, it's a weighted decision because there are consequences balanced on both sides, it's an immediate decision because it impacts the battle which is taking place with the Dragon, and it's a long-term decision because it impacts the future battle with the evil mage. All these combine to make the decision of whether or not to use the golden arrow a critical choice in the game, and this makes the game interesting.

Exercise: Decision Types

Take your original game and categorize the types of decisions you ask your players to make. Are there any hollow, obvious, or uninformed decisions? If so, try to redesign these choices.

Not all decisions in a game need to be as complex as the one with the golden arrow. Simple decisions are fine, just so long as they're not hollow, obvious, or uninformed. As a rule, you want to remove all nondecisions from you game, and a player should never be forced to think about anything unless it has some impact, either direct or indirect, on whether they win or lose.


Dilemmas are the situations where players must weigh the consequences of their choices carefully, and in many cases, where there is no optimal answer. No matter what the player chooses, something will be gained and something will be lost. Dilemmas are often paradoxical or recursive. A well-placed dilemma and trade off can resonate emotionally with a player when encountered during the struggle to win your game.

Game theorist John Von Neumann used dilemmas as a framework for studying how people make choices, and how conflicts are resolved in both game-based and real world dilemmas. Game designers can use the same methodology to study the choices in their own and other designer's games.

To understand dilemmas, von Neumann broke them down into very simple structures, called moves. Each move was diagrammed on a matrix, showing the potential outcomes of each strategy as they pertain to each player. To understand this concept more clearly, let's next look at a classic dilemma with a simple structure.

Cake-cutting dilemma

A mother wants to divide a piece of cake between her two children. In order to avoid arguments about how large a piece each child should get, she makes one child the "cutter" and one child the "chooser." The cutter gets to cut the cake, and the chooser gets to choose which piece. If we assume that each child wants the bigger piece (i.e., wants to "win" the game), we can diagram this conflict to show the potential strategies for each player, the dilemma they face, and the payoffs for each potential outcome.

Figure 2. Cake-cutting payoff matrix

As we can see, each child has two possible strategies. We know that it's impossible to cut the cake exactly in half; there will always be one crumb more or less on either side; but the cutter can choose to cut the cake as evenly as possible, or she can choose to cut one piece bigger than the other in an attempt to get the larger slice. Since we've determined that one piece will always be larger than the other, even if just by a crumb, the chooser also has two strategies. He can choose the smaller piece or the larger piece.

By looking at the payoff matrix created by combining these two possible strategies for each player, we can see that in this simple game, there is an optimal strategy for each player. Since we have said that each child will try to get the bigger piece, the chooser's optimal strategy is obvious--he will choose the larger piece. And, since the cutter is also trying to get the largest piece possible, she will try to cut the pieces as evenly as possible. The optimal strategies for each player meet in payoff #1: the chooser gets a slightly bigger piece.

The cake-cutting dilemma is an example of a zero-sum game. By this we mean that the total amount won at the end of the game is exactly equal to the amount lost. In this case, the chooser gains the crumb lost by the cutter. Because of the nature of zero-sum games, the interests of the players are diametrically opposed. What one player loses is gained by the other.

What von Neumann discovered in his study is that there is an optimal strategy for each player in games of this nature that will produce the best possible results in a given situation. He called this concept "minimax theory."

Minimax theory states that there is a rational way for players to make choices in a game, if we are talking about a two-player, zero-sum game. The optimal strategy for all players is to "maximize their minimum potential result." So, in the case of the cake-cutting example, while the cutter cannot "win" the game, her optimal strategy will still maximize the amount of cake she gets to eat.

Games that fall easily into optimal strategies may be interesting for mathematicians, but as game designers, they are often the kiss of death. If you present your players with a game as simple as the cake-cutting dilemma, they will always make the optimal choice and the game will play out the same way every time. How can we create dilemmas that are more complex, where the players must weigh the potential outcomes of each move in terms of risks and rewards?

A game that has a more complex payoff structure was created by two RAND scientists in the 1950s. Called the "prisoner's dilemma," it's a simple, baffling game that shows how games that are not zero-sum can create situations where the optimal strategy for each player can result in sub-optimal results for both.

The prisoner's dilemma

Two criminals commit a crime together and are caught by the police. For the purposes of our example, we'll call the two unlucky criminals Mario and Luigi. Mario and Luigi are held in separate cells with no means of communication. The DA offers each of them a deal and discloses that the same deal was made to his partner in crime. The deal works like this: if you rat on your partner, and he denies it, you can go free and the he gets five years. If neither of you rat on each other, the DA has enough circumstantial evidence to put you both away for one year. If you both rat, you will each get three years. Figure 3 shows the payoff matrix for each potential strategy.

Figure 3. Prisoner's dilemma payoff matrix

Using the same process we used to determine the optimal strategy for the cake-cutting dilemma, we can see that the optimal strategy for Mario is to rat on Luigi. If he rats, he gets either three or zero years. If he doesn't rat, he gets one or five years. The optimal strategy for Luigi is also to rat on Mario for the same reasons. If both players choose the optimal strategy, however, they will both serve three years--more years total will be served in jail than in any other resolution.

The hierarchy of payoffs in the prisoner's dilemma is as follows:

The actual numbers in this hierarchy are not important. What is important is that they ascend in this order: Temptation > Reward > Punishment > Sucker. If this hierarchy exists, the optimal strategy for each player will always result in a payoff that is less than if they had acted cooperatively. Now, we are talking about a true dilemma--what will Mario and Luigi do?

Exercise: Dilemmas

Does your original game contain any dilemmas? If so, describe these choices and how they function?

In a recent presentation at the Game Developers Conference, designer Steve Bocska of Radical Entertainment applied the hierarchy of payoffs in the prisoner's dilemma to a hypothetical game design in order to show the usefulness of game theory concepts to designing compelling dilemmas.1

Bocska imagines an online game in which two players are building and customizing spacecraft with a budget of $10,000. The game requires bartering and trading of raw materials, but at a high transaction cost: $8,000 of "shipping and handling" in a typical game round. A technology can be purchased that allows materials to be "transported" with no transaction cost--but in order for it to work, both players must purchase it. The cost of the technology is $5000.

Bocska asks, "Under these conditions, what is a player likely to do? If both players purchase the transporter equipment, they will reduce their transaction costs for the game from the usual $8,000 to a one-time cost of $5,000 for the transporter--a savings of $3,000. If, on the other hand, neither player purchases a transporter, the transaction costs throughout the game for each player will amount to the usual $8,000. What if only one player purchases the machine? With nobody else to connect the transporter to, their machine becomes effectively useless, resulting in them receiving the "sucker's payoff"--the cost of the equipment plus the added cost of continuing to barter using the traditional costly method ($5,000 + $8,000 = $13,000)." The payoff matrix in Figure 4 shows the results of the potential strategies.

Figure 4. Transporter game payoff matrix

Unlike the prisoner's dilemma, Bocska envisions a game in which the players can communicate--negotiating with each other when and if to purchase the technology. This complex payoff structure creates a dilemma for the players that can make for compelling strategic moments and potentially deceitful or cooperative decisions play after play.

This is exactly the type of situation you should strive to create in your games. If possible, give the players dilemmas as part of the core gameplay. Make sure to tie the dilemma into the overall objective of the game. If you can accomplish this, it will make the choices much more interesting.



Another format for structuring interesting choices in your games is by incorporating puzzles. Puzzles are solvable systems. They can contextualize the choices that players make by valuing them as moving towards or away from the solution. Suddenly, the act of rifling through a treasure chest takes on new meaning if you are searching for the key to open the door to maze, rather than just looting the castle.

Puzzles are also a key element in creating conflict in almost all single-player games. There's an innate tension in solving a puzzle. If you tie this into a system of rewards for solving the puzzle and punishments for failure, the puzzle transforms into a dramatic element. For example, take Myst, the best-selling adventure game of all time. It's essentially comprised of puzzles. It incorporates story and exploration as well, but at its core, it's really a glorified collection of interlocking puzzles.

The popular genre of first-person shooters is also puzzle-based, especially in single- player mode. Take Medal of Honor. You have to plant bombs, unlock doors, find medical kits in a labyrinth of rooms, and figure out how to use weapons and explosives in just the right way. The game is one giant action puzzle. The same holds true for many other single-player games.

You'll notice that we keep using the qualification "single-player." This is because in multiplayer mode you don't need puzzles to provide conflict. The conflict comes from the competition with other players, whether they are human or computer-controlled. But in single-player mode, especially when you are sent on a quest or mission, puzzles play an increasingly important role. That's why every game designer should also consider herself a puzzle designer. The better your puzzle design skills, the better your game will be.

One consideration when designing puzzles in your games is to make sure that the elements of the puzzle are woven into the fabric of the game. By this we mean that it advances the player towards his overall goal. If a puzzle doesn't enable progress, it's a mere distraction and should be redone or removed. A puzzle may also advance the storyline. You can use the puzzle to tell the player something about the unfolding plot.

If you can integrate your puzzles into the gameplay and the story, they won't feel at all like "puzzles," but rather like integral, interesting choices a player must make to progress in the game as a whole.

Rewards and punishments

The most direct consequences for player choices are rewards and punishments. Obviously, players enjoy being rewarded and dread punishments. Nothing is more natural. So when designing a game, a designer often emphasizes the rewards, while limiting the punishments. This makes sense; players aren't playing games to suffer the hardships of life. And in reality, you don't want to punish players so much that they stop playing your game. But often, the threat of punishment, if not the actual punishment itself, carries a dramatic tension that can add layers of meaning to even the most trivial choices a player makes.

Think of a game that forces the player to be stealthy, like Thief or Deus Ex. The tension, when you are trying to accomplish a task without being caught is tremendous. Getting caught and attacked, and let's face it, killed, is not fun. But that moment when you oh, so quietly pick a lock and sneak past the security bots without incurring any harm is made much more effective by the threat that the anvil of punishment was hanging over your head all the time (see Figure 5).

Figure 5. Being stealthy: Thief

Coming up with a balanced system of rewards and punishments is a way of making the choices in your games much more interesting for players. The type of rewards you offer can vary, but the best rewards are those that have utility or value in the game. When you develop your rewards system, use the following guidelines:

  1. Rewards that are useful in obtaining victory carry greater weight
  2. Rewards that have a romantic association, like magic weapons or gold, appear more valuable
  3. Rewards that are tied into the storyline of the game have an added impact.
    Make each reward count, and if it can both push the player closer to victory and advance the storyline, that's even better.

The timing and quantity of rewards is also critical. If you give a steady stream of small rewards, it can become meaningless. Players know the rewards are coming, no matter what they do and they stop caring.

Psychologist Nick Yee has studied the reward/punishment structure of an extremely addictive game system--EverQuest--and believes its addictive power lies in a behavior theory advanced by B.F. Skinner called "operant conditioning." Operant conditioning claims that the frequency of performing a given behavior is directly linked to whether it is rewarded or punished. If a behavior is rewarded, it is more likely to be repeated. If it is punished, it becomes suppressed. It is usually explained by using the example of a "Skinner Box," a glass cage equipped with levers, food pellets, and drinking tubes. Rats are placed in the cage and rewarded with a food pellet for pressing the lever, using reinforcement to shape their behavior.

Yee writes, "There are several schedules of reinforcement that can be used in Operant Conditioning. The most basic is a fixed interval schedule, and the rat in the Skinner Box is rewarded every five minutes regardless of whether it presses the lever. Unsurprisingly, this method is not particularly effective. Another kind of reinforcement schedule is the fixed ratio schedule, and the rat is rewarded every time it presses the lever five times. This schedule is more effective than the fixed interval schedule. The most effective method is a random ratio schedule, and the rat is rewarded after it presses the lever a random number of times. Because the rat cannot predict precisely when it will be rewarded even though it knows it has to press the lever to get food, the rat presses the lever more consistently than in the other schedules. A random ratio is also the one that EverQuest uses."2

While this might seem surprising, if you relate it to your own actions in games and in the real world it begins to make more sense. Have you ever sat down to play "five minutes" at a slot machine and looked up to realize you'd been there, determinedly pulling that lever for several hours? In many ways, Las Vegas is simply a giant Skinner Box.

We may all be just rats in a cage, but there is one type of reward that is very powerful and that can't be delivered like a pellet, and that is peer recognition. Humans crave acknowledgement for their achievements, and there's little that can motivate us more. Especially in multiplayer games, if there's a way for you to make the players, even the ones who aren't winning, feel recognized for their efforts when they do achieve a goal, then you will have a much stronger game.

Many games do this through the Internet, tracking scores or providing tournaments and ladders. There are more immediate ways to provide recognition, in the moment, as well. One is to track and broadcast the players' achievements during the game, highlighting and dramatizing each success for everyone to see. If it's an online strategy game where one team is pitted against another, make it clear when a player pulls off a brilliant maneuver. Let his comrades know exactly what happened and how it impacts the victory conditions. If it's an online RPG, allow the players to show off their conquests to the world, either in the form of legends, artifacts, or admirers who follow them about.

Exercise: Rewards

Analyze the rewards system in your original game prototype. Look at each reward and determine if it is useful, romantic, and/or tied to the storyline. How are rewards timed? Does the timing reinforce the player's desire to continue playing?


The Skinner Box example works well for game mechanics that are repetitive and apt to become rote. For larger, more complex choices, however, the more clearly you allow players to see, and anticipate, the consequences of their actions, the more meaningful their choices will be.

In chess, and other games with open information structures, the entire state of the game is visible to both players for evaluation. There's nothing hidden. If players are experienced, they can calculate out moves dozens of turns in advance and see exactly what will and will not happen. The anticipation that players feel in a situation like this is heightened by the knowledge of when they will be able to capture a piece or get in a particular position.

Can games with closed or mixed information structures create anticipation? Definitely. Real-time strategy games often use limited visibility to offer the player a glimpse of the opposition, but only while her units are posted in enemy territory. Since the game state is always changing, the view quickly becomes outdated, and the player winds up making decisions based on only partially accurate information (see Figure 6).

Figure 6. Warcraft III: fog of war

In this example, players accept the lack of information as one of the conditions of the game and understand that their job is to maximize their position given the limited information they have available. In fact, the lack of visibility can increase a player's sense of tension. With the knowledge that the game state is in flux, players feel compelled to act swiftly to counter anticipated enemy moves. In many ways, the hiding of knowledge has added a new dramatic twist that is lacking in the completely open strategy games.


Surprise is one of the most electrifying tools at a designer's disposal. People love to be surprised, especially when they feel they should have anticipated the event. Too many surprises will alienate players, however, so, how do you know when to use surprise and when to telegraph an event?

A surprise outcome to a player's choice can re-invest them in the game--perhaps they thought they were going to find 20 gold pieces behind door number three, but it turns out to be a trusted friend ready to join their journey instead--a much greater reward.

Surprises may feel random to players, but in a good way. The trick is to find the right balance between the randomness of surprise and the importance of making player choices meaningful. Take the example of a real-time strategy game, where you might send a simple foot soldier up against an ogre because he's all you've got. The foot soldier has strength of one to five, while the ogre has strength of one to 20. Odds are that the ogre will win. But there's always that chance, no matter how small, that the foot soldier will prevail.

Randomness, and surprise, in this case adds a level of drama--the tension of not knowing exactly how a highly probable event will play out. Will this be a David and Goliath story or just another dead foot soldier? In most well- designed games, the element of choice remains dominant. If every choice a player makes results in random effects, they will feel like their choices have no meaning. But keep surprise in mind; used judiciously, it can create a wealth of fun and excitement.

Exercise 10.5: Surprise

Are there any surprises in your game? Try taking one type of choice and adding an element of surprise to the outcome. How does this affect the gameplay?



Nothing is quite as satisfying as seeing the choices you make result in progress. It's part of human nature to derive joy from the act of advancing towards a goal. The small payoffs along the way are often sweeter than the final victory. The same is true in a game. Allowing players to feel they are moving forward is the best way to draw someone into a game and keep them engaged.

One approach for structuring progress is to design milestones for the players. These are small goals along the way to the grand goal of winning. Advertise these milestones to the players so that they know what they're striving for, and reward them after each accomplishment.

Many games do this well. In Medal of Honor, the milestones come in the form of missions. They give you a map and let you see where you're headed and what you have to achieve to get there. This helps the player feel like they're making progress throughout a long campaign. The same is true for games that use story to block out their single-player levels, preparing the player at each step and setting out clear and obtainable objectives, and then rewarding the player at the end of each sequence with graphics, praise, and another chapter of the narrative.

Figure 7: Medal of Honor

No matter what the game, whether it's an arcade shooter or a simulation, providing a path for the player to follow gives a sense of achievement. Be creative in finding new ways to represent progress for your players. Don't limit yourself to just one system. There's no reason you can't measure progress in several ways at once.

When you consider the pacing of progress that players can make in the game, you might also consider the typical amount of time a player spends with a game. Veteran game designer Rich Hilleman with Electronic Arts says that their designers plan "mini-arcs" of about one hour into the overall game progress. This is because that is what they have found the length of time the average gamer plays for in a single sitting.

At the end of each mini-arc, the designers try to make sure the player encounters a "memorable moment" of gameplay, which makes sure they will return for another play session. These mini-arcs, when aggregated, form the overall dramatic arc of the game.

Exercise: Progress

Take your original game prototype. Is the ultimate goal clear? Is the player always moving towards this goal? Make sure that you have milestones established along the way. Does your system help motivate the player to reach the final goal? Describe how.

The end

By "the end," we're not talking about when a player dies; we're talking about the moment when the play completely resolves. After investing hours, days, weeks, or even months, this is the instant when your most loyal players deserve a reward for all their effort.

Multiplayer games have their own reward built in: the satisfaction of beating the other players, or, if you have created a cooperative, unilateral, or team interaction structure, the satisfaction of having worked together to beat the game or the other side.

But what of your single-player game? After all the conflict, struggle, and time invested, make sure to give the player a satisfying reward. Too often, the end of all that work is a fluffy animation, where the hero is showered with praise and adulation. If you're going for an ending like this, why not build the reward into the story? Make that animation a moving moment in your hero's quest for whatever he lacks.

Exercise: Endings

Is the ending or resolution of your original game prototype satisfying? If not, how could you make it even better?

Fun "Killers"

For all your efforts, you may have implemented some features that "killed" the fun in your original concept. Here are just a few we've seen come up in first-time games over and over.


Micromanagement is a classic example of forcing your players to make decisions that don't feel important. Game designers can fall into a trap of believing that the more control they give the players over their universe, the happier they'll be. On some level this is true. Hardcore strategy gamers love control. They want to tweak everything and dissect each element of the game. But there is a fine line between granting your players control and burdening them with chores.

As a designer, how do you know when you've given them enough? Start with the basics. Is the task necessary? Make sure the decision the players are given is not obvious, hollow or uninformed. If it passes these tests, it still may fall into the micromanagement trap. Micromanagement takes place when a task becomes repetitive or tedious. If you're asking the players to make too many small decisions and not enough large ones, then you have a problem. The best way to know this for certain is to bring in fresh playtesters. If they complain that it's too much work or too tiresome, this is a red flag.

The solution in almost every case is to simplify your game. Micromanagement comes from breaking up a task into too many small pieces. The overall impact of the combined decisions may be important on a strategic level, but each individual decision is too burdensome to be worth the effort. The solution is to combine the micro-decisions into one macro- decision. For instance, if deploying an army requires choosing what weapons each unit will use, what supplies they're going to carry, what form of transportation they'll utilize, and what route they'll travel, you're probably asking too much of your players. You can solve this by making some of these choices for them. Set default values that make sense and leave only the most important decisions, like what route to travel, up the players.

In addition to eliminating lesser decisions, you can give the players ways of automating certain tasks, like resource management, troop deployment, and logistics. This provides players with the degree of control they desire. Some players may chose to do everything manually, while others will prefer not to deal with the details. You'll find that different people want different things from your game, and the more flexible a system you can provide, while still keeping the game relatively simple, the better.

Exercise: Micromanagement

Are there any elements of micromanagement in your game? If so how can you streamline the choices players make so that they are not bogged down by unimportant details?


Some games fall into a pit of stagnation, where nothing new seems to be happening for a long period of time, choices stay at the same level of importance and impact.

A common source of stagnation is repetition, where players are caught doing the same task over and over. For instance, if the players are forced to fight the same type of battle repeatedly, the game can feel like it's at a standstill. The players may actually be advancing in levels or moving closer to their ultimate goal, but the actions they're performing are so repetitive that they mask any progress being made. In this case the solution is twofold. First, you should vary the type of action being performed. Next, you need to communicate how each action is bringing the player closer to victory.

Another type of stagnation is a balance of power. For instance, you have three players competing to conquer the world, and whenever one player gets ahead, the others gang up and smash the leader down, thus creating an endless cycle where no one is able to achieve victory. The solution here is to create a condition that tips the balance of power so far in favor of the winner that she can defeat the combined strength of the opposition.

A third type of stagnation is the endless loop syndrome. This type of stagnation occurs when a game device traps the players in a cycle. For instance, in a business simulation game, a player may get caught in a trap where all his profits are being eaten up by debt payments. No matter how long he plays, he cannot get over this hump. One solution is to shake things up with an unexpected event, like a windfall or natural disaster that will either push the player over the hump or knock him into bankruptcy. Of course, you could also tweak the game so that players never get stuck in this type of situation. Give the player debt relief or jack up interest rates--whatever it takes to move the game in one direction or the other.

The last type of stagnation is where it feels like nothing is happening, because nothing is actually happening. In other words, no progress is being made, either because the game is poorly designed or because there's no clear goal. An example of this might be an adventure game with a poorly defined objective. The players roam around but have no idea where they're supposed to go or what they're supposed to do. In this case, the solution is to go back and design the game so the objective is clear.

Exercise: Stagnation

Is there any point in your original prototype at which the game play stagnates? If so, determine what is causing this problem--do you have a repetitive loop? A balance of power? How can you break the cycle and improve the progression of the gameplay?

Insurmountable obstacles

Another problem area to avoid when designing a game is insurmountable obstacles. Despite the name, these may not actually be impossible situations, they just seem that way to a certain percentage of players.

Whether this occurs because of a dearth of information, a missed opportunity, or lack of experience or intuition, the result is always the same: your players wind up banging their heads against the same obstacle over and over and over. Look at your watch--how long now before they shut the game off in frustration, never to return again?

Most of us have been trapped by insurmountable obstacles at one time or another and wound up going in circles looking for that hidden doorway or secret panel. As a designer: make sure that and the game has some way of recognizing when a player is stuck and providing them with just enough assistance to make it past the obstacle without diluting it's challenge completely. Of course, this is easier said than done. Nintendo adventure games, such as those in the Zelda series, are typically good at providing info when players are stuck. Game characters are placed in strategic spots to provide clues and other information that will help you overcome the obstacles. Like with other variables, clues have to be balanced to provide an appropriate level of difficulty for the players.

Building this kind of intelligence into the game is costly and time consuming. Sometimes, it doesn't need to be that sophisticated. In his presentation at the 2003 Game Developers Conference on making games more fun through user-testing, Microsoft User-Testing Manager Bill Fulton used an example from the opening moments of Halo to illustrate how a task that seems obvious to the designer may seem like an insurmountable obstacle to the player.

Immediately after the introductory tutorial of this first-person shooter game, your character is asked to follow a guide character to the bridge of the spaceship you are on. Of course, you do, but seconds later, the guide character is killed in an explosion right before your eyes, leaving you trapped behind a partially open doorway with no guide, and no clue how to open the door.

As part of his presentation, Fulton showed a videotape of just one of many user tests in which a playtester stumbled around the corridor, pressing every button on the controller, trying everything he could think of to open the door, all the while talking out loud about how he didn't know what to do. This went on for several minutes until it became clear from his tone of voice that if this player were at home, he would have given up--only five minutes into the game. As Fulton pointed out humorously, "I hope you all recognize this as 'not fun.'"3

The goal of the Halo designers, in leaving you guideless, trapped behind the door, was to create a sense of confusion and vulnerability that lasted only a few seconds for dramatic purposes. They assumed the player would immediately realize the door wasn't going to open, see the alternate exit to the corridor they had planned, and be on their way. User-testing proved that most players needed a little help past this obstacle.

In the final product, a second explosion, timed a few seconds later, drives the player instinctively away from the half-opened door that will, in fact, never open. A text prompt pops up showing the player how to jump over objects, and a carefully designed floor mat points toward another opening in the corridor. The opening is blocked by a set of pipes, but if you know how to jump, that's no problem at all. With just a few modifications, the opening moments of the game were changed from an exercise in frustration to an exciting scene, filled with drama and tension.

Arbitrary events

As much as random events can be used to good effect in certain circumstances, like fortuitous surprises and unforeseen dangers, badly designed randomness can be the downfall of a game. Many games involve some form of randomness--we've seen how randomness can affect combat algorithms in real-time strategy games, and how it can stop movement mechanics from becoming predictable in boardgames. These types of randomness add to gameplay.

But there's a big difference between utilizing randomness to change up gameplay and allowing for totally arbitrary events to disrupt the player experience. For instance, if you've spent weeks building a sophisticated character in a role-playing game, and then suddenly a plague, for which there is no cure, kills your character; you had no chance to defend yourself, and all your hard work has gone down the drain. It's hard not to feel cheated. We all know that life is full of unexpected events, some of which are devastating. So why shouldn't games include them?

The problem is that, as in life, good surprises are welcomed by players, but bad surprises are not. So, how can you include random events that are negative in nature without alienating the player? Whether it's a meteor storm that levels a city, an economic fluctuation which bankrupts a company, or a surprise attack that wipes out an army, you have to make sure it fits into the players' expectations of the game. Prepare your players in advance for the possibility of such an event and give them options to mitigate the damage--just don't tell them when it's coming or how bad it's going to be.

If we take the example with the plague, you should warn the player about the possibility of diseases, and allow them to purchase an antidote in advance. If they choose to ignore the warning signs and take no action, then when the event does come, it's their fault, and they'll know it.

A good rule of thumb is to caution your players at least three times before hitting them with anything catastrophic. Random events that have a lesser impact require smaller warnings, or even no warning at all. It's fine for a player to learn through experience to expect events of smaller consequence. But the bigger the impact, the more of a heads up you should provide. If you follow this rule, the events won't appear arbitrary, and your players will feel like they are in control of their destiny.

Predictable paths

Games with only one path to victory can become predictable. Linear or simple branching structures often lead to this type of predictability. If you want to add a greater sense of possibility to your design, consider treating the structure in a more object-oriented approach. Giving each type of object in the world a simple set of rules for interaction, rather than scripting each encounter separately often leads to creative and unusual results.

An example of this type of thinking is Grand Theft Auto III, which has a level structure and story line that the player can follow, or he is free to wander the world, stealing cars, committing crimes, or running a taxi service if that's what he wants to do. While wandering the world doesn't advance the player very far in terms of the overall game objectives, it does give the sense that the world of the game is responsive and unpredictable. At any moment, you might attract the attention of the police and wind up in an unscripted high-speed chase. Simulation games are other examples of this type of design--games like the SimCity series can evolve in many directions, all based on the choices of the players.

Figure 8. Sim City

Another way to keep game paths from becoming predictable is to allow players to choose from several objectives. For instance, in Civilization III, the player can choose between six paths to victory: conquest, space travel, cultural advance, diplomacy, domination, and overall score. Each choice takes careful planning and will cause the player to weigh each choice anew, making the game not only interesting the first time around, but extremely replayable. Simply choose another path to victory and the game takes on an entirely new twist.

Not every game has to have the scope of a Civilization or a Grand Theft Auto, but when finding the balance between too much possibility and too much predictability, it's usually best to err on the side of greater possibility.

1 Steve Bocska, "Temptation and Consequences: Dilemmas in Video Games," Game Developers Conference 2003.

2 Nick Yee, "EQ: The Virtual Skinner Box,"

3 Bill Fulton, "Making Games More Fun: Tips for Playtesting Games," Game Developers Conference 2003.

End notes


Return to the full version of this article
Copyright © UBM Tech, All rights reserved