Jeff's World of Game Design

Nickel words from the pen of the designer of BGG's 7600th greatest game
 Thumb up

Cognitive load: juggling without hands

Jeff Warrender
United States
Averill Park
New York
flag msg tools
"You Said This Would Be Fun", a book about game design, available at Amazon and DriveThruRPG
Come talk design at the Jeff's World of Game Design blog!
Microbadge: You Said This Would Be FunMicrobadge: The Acts of the Evangelists - witnessMicrobadge: The Acts of the Evangelists - witnessMicrobadge: The Acts of the Evangelists - witnessMicrobadge: The Acts of the Evangelists - evangelist
Many readers may be familiar with the "rule of 7". This is the idea that a typical person can store about 7 pieces of information at a time, plus or minus two, but that it's often possible to improve on this by bundling information in "chunks" (for example, the area code of a phone number contains three digits but only requires one "chunk" of memory). And they may also be aware of studies showing that the brain can really only effeciently "multi-task" on two tasks at a time, one for each hemisphere of the brain; more than this can overload the brain.

Now, I have to say that in my experience the rule of 7 is, in my case, more like a rule of 2. Tell me three things I'm supposed to get at the store and it's a sure thing I'll forget one or maybe all three! This rule of 7 is no comfort to me, but I have found in our playtest group that one's cognitive capacity isn't strongly correlated to one's native inelligence, so maybe there's hope for the coarse-grained among us.

Recall that in the previous post I asserted that we should prefer to present players with decisions that are agonizing over those that are hard, and these rule-of-7 considerations provide an entry point to contemplate complexity in design from the player's perspective. This is important in an era where games are becoming increasingly complex (yes, I'm guilty of contributing to this), but not in the way you might think. It's fairly obvious to say that the more cognitively demanding a game is, the more limited its audience may be, because not everyone wants to think to excess. And it's also fairly obvious to say that we need to think about how much complexity a typical player can cognitively handle before reaching a frustration point. And we'll do this in this post, but what's even more interesting to me is the effect excessive cognitive burden can have on decision-making.

Stanford professor Baba Shiv reported a study of decision making under cognitive load. Participants were given a series of numbers to remember, and then asked to choose either a piece of chocolate cake or a salad. Participants given a long string of numbers were overwhelmingly likely to choose the cake, suggesting that, the more cognitively burdened the participant was, the more likely that person was to make the most impulsively appealing decision. Similarly, a Rutgers study investigated performance in a game under cognitive load. Participants played a Prisoner's Dilemma game, and they found that players who were given only a few numbers to remember converged to the optimal play over a series of games, whereas players given a long sequence of numbers to remember did not.

These studies, and others, point to an important finding: decisions made under cognitive overload are more likely to be impulsive and less likely to be analytical. How many times have we done this, or seen it happen: "Bah! I give up, I don't care. Whatever, I guess that I will play my magic pea shooter card to attack, uh, I guess, the sea dragon." The message for game designers is that making our games complex may not, in fact, result in deeper gameplay; for many players, it may actually result in substantially shallower gameplay, because the players, frustrated and overwhelmed by the options, will give up and make their moves impulsively rather than strategically.

There's a separate but related consideration: if the amount of mental effort the game requires seems out of proportion with the reward of expending such effort, or if the factors outside the player's control are significantly more important than those within the player's control, then players are unlikely to engage in the deep thinking that the game permits (or more likely, they just won't play in the first place). You're giving players a ton of things to think about, but there's no strong benefit to actually thinking about them deeply. This is the worst state of affairs, where you overload your players' cognitive faculties to no purpose.

The goal here isn't to quantify a threshold of exactly how much cognitive burden is appropriate for a game, but rather to be constantly looking for ways that we can reduce the cognitive load on our players, or to give them the ability to "chunk" information so as to unlock more of their cognition.

The first place to look is in the amount of mental energy spent on compliance. In playtesting games, one of the first things I pay attention to is whether I'm spending more time thinking about what I am allowed to do than what I want to do. Using brainpower to keep the rules straight and comply with them takes away from my ability to make decisions, especially for a "rule of 2" guy like me, and so I'm trading the fun part to follow an elaborate set of rules, and that's not very enjoyable. Bureaucracy, status phases, fiddly rules exceptions, these are all (generally speaking) bad, precisely because they encroach on the player's cognitive ability to weigh the available options and pick the best. Good component design, player aids, and appropriate thematic terminology can all help, but unneeded complexity should be stripped out whenever possible because of the role it plays in distracting the player from what's actually interesting.

Expanding on this point, one of my game design axioms is "design with incentives, not with restrictions", and cognitive load considerations show why this is important. Rules and restrictions that prohibit the behavior I don't want are well and good, but they impose a cognitive burden on the player: an extra rule the player must remember. So instead of "you may never have more than 6 territories", better is "each territory costs 1 upkeep". Now I'm creating a disincentive to excessive territory-grabbing, but I don't absolutely preclude the players from doing it. This puts the choice in their hands. If everything is balanced well, it will still be the case that players won't usually want to exceed 6 territories, but it's up to them to make this determination. Instead of giving them a rule they must remember (compliance-cognition), they're weighing benefits in tension with risks and costs (decision-cognition). This is the good kind of thinking.

Another way to promote the good kind of thinking is to use the game's visual presentation to minimize the number of things that the player must hold in mind. As a simple example, the boats in Puerto Rico have a number of goods that they each can hold, but the specific boats used changes at each player count. Imagine if we tried to remember, all in your mind, "ok, we're in a 3p game, so there's a 4, 5, and 6-capacity boat, and the player on my right just shipped 3 corn to the 6-boat, and there were already 2 corn in the boat so there's one spot left"; this would be hopelessly daunting. Game pieces like pawns and counters help us to easily see the game's geographical state so that we don't have to remember it (imagine playing chess entirely in your mind! Some people can do it, but it's not easy!). Similarly, then, printed components like the ships in Puerto Rico let the players easily enforce the rules about ship capacities without needing to even think about them.

Another nice effect is when components enable us to 'chunk' certain things together. For example, the icons on the buildings in Citadels actually convey three things: how much does it cost to build (and destroy, for the Warlord), how many VP is it worth, and what category is it (color of the symbols)? This information is all encoded in a single icon, making it easy for everyone to extract at a glance the information it provides without needing to think separately about each of these characteristics.

The most common manifestation of cognitive overload is "analysis paralysis", the effect whereby the amount of information and/or options that the game is throwing at the player is so overwhelming that the player takes a long time to sort through the options and make a decision. (Or worse, that cajoled by the other players to speed up, they throw up their hands in frustration and do something impulsive like attack the sea dragon with a pea shooter). As discussed above, we can ameliorate this by presenting the information economically, and encoding information in components rather than in rules. But it's worth asking what a player is doing when frozen in analysis. Sometimes analysis paralysis results from perfect-planning; from a player trying to plan out every possible move and countermove for the rest of the game. This is important in 2p games like Chess, but hopelessly hard in multiplayer strategy games, but players still try to do it. Can we encourage players to focus more on putting themselves in a good position and less to chart out the absolute perfect course?

If we can reduce the distance to the horizon over which the player is trying to look, we can reduce the number of permutations of possible outcomes the player is likely to want to consider. We can do this trivially by making the game more chaotic, either through the presence of randomness or the presence of tremendous flux between turns. This isn't necessarily bad; these merely make a game more tactical, and make long range planning more futile. But they also mean a player isn't, indeed can't be, doing much during other players' turns, and that leads to downtime of a different sort. And so, even in tactical games like Carcassonne or Citadels some strategic scope is desirable, especially if this is a long game and we want our players to play strategically.

Here, then, we can reduce the overload by restricting the number of options the player has available. For example, say we give each player five "country" cards. You can only act in one country per turn, by playing that country's card, and you only get your full set of cards back after you've used all five. Thus we will focus the player's attention on one area, and as an added bonus, will funnel that initial decision of which country to act in as the turns go by.

We can also help by simplifying the number of ways that points can be earned. In a "point salad" game like The Castles of Burgundy, the range of options can be explosive, but in an economic game like Chinatown, or a race game like Hare & Tortoise, or a route-planning game like Elfenland, there is a single goal, and the range of options that the player considers are not about which goal is best, but which is the best route to achieving that goal. This is an important point: while "multiple paths to victory" games are fun for their sandbox-y nature, players must think about what path to pursue AND how best to pursue it. This becomes easier the more focused the player becomes, and so in such a game, there can be incentives that encourage a player to stay on a strategic path once on it, because this also simplifies the cognitive process the player must use, and keeps the game moving briskly for everyone else.

Another axiom I try to follow was stated eloquently by Jonathan Degann: "Your scoring design IS your game design". Players will follow the incentives the game's scoring system offers. The corollary of this is that players will think about the things that the scoring system rewards them for thinking about. Simplifying our scoring systems not only relieves the players' cognitive burden, it also steers them to use that cognitive capacity to think about the right things.
Twitter Facebook
Subscribe sub options Mon Oct 1, 2018 1:37 pm
Post Rolls
  • [+] Dice rolls
Loading... | Locked Hide Show Unlock Lock Comment     View Previous {{limitCount(numprevitems_calculated,commentParams.showcount)}} 1 « Pg. {{commentParams.pageid}} » {{data.config.endpage}}
    View More Comments {{limitCount(numnextitems_calculated,commentParams.showcount)}} / {{numnextitems_calculated}} 1 « Pg. {{commentParams.pageid}} » {{data.config.endpage}}