Good game design is not enough–in order to build a good game, the designer must also think about the agents that will interact with the system he or she has created, the players. Players are the most important and integral component to your design, and you need to think about how they will interact with what you’re building as well.
I started thinking about this problem when we designed early versions of Millennium Blades. The game was fun, but after only 2 hours it was exhausting. Players felt mentally stretched and fatigued by the amount of information they had to store and process at once. I started thinking about why that could be, and what could be done from a designer perspective to account for it.
The key isn’t really in simplifying a design. Simple games can be overwhelming, and the actual play of Millennium Blades was pretty simple (it could be explained in about a minute). It’s another property that, for the sake of this article, I’ll call information density. In a more dense game, we are asked to hold and manipulate more pieces of information at one time. In a less dense game, we need to hold and process less.
So what is the average human capable of? Scientific studies tell us that our Working Memory can hold a small number of “chunks” of data, depending on their complexity. The amount of working data a person can process (in terms of strategic data) is usually around 4 pieces, with children and old adults having a slightly lower capacity.
In a previous post, I talked a bit about Objective-Oriented Design. With 4 pieces of strategic data to store, the average player can evaluate a goal chain about 3 elements deep–that is, they can evaluate the effectiveness of any move bringing them closer to victory about 3 steps on that objective tree. I’ll reference that post a lot in this one, so if you haven’t read it yet, take a quick break and check it out.
Experienced Players vs. Inexperienced Players
When you deal with similar information structures a lot, you inherently learn to simplify and collapse them. This is why gamers are capable of playing more complex games–they can collapse portions of the tree that are redundant or which they have already resigned themselves to. For example, on the Catan decision tree, the player doesn’t need to evaluate the 2nd level of the tree, because he knows the only way to Build cities is to build towns.
Another way experienced players can optimize their play is to keep a portion of their objective tree in long-term memory rather than working memory. For example, the player may decide that “Build Cities” is going to be at the bottom of his decision tree and be the main objective (at least as far as his working memory is concerned). In this way, he can reduce his decision tree down by a significant number of levels, and make much quicker turns.
Casual players who are new to board gaming (and video gaming to a similar extent) do not have these practiced memory shortcuts, and must evaluate the entire tree for each decision. Even experienced players can grind to a halt when thrown into an unfamiliar system (such as teaching a card gamer to play a miniatures game, or teaching a board gamer a complex LCG or CCG), because they have not forged the mental tools and shortcuts for quickly evaluating new shapes of decision chains.
Practicum: Reducing Density
So what’s to be done to make a game more accessible? Well, that depends. As we’ve seen from the section above, you can afford a more complex game if your players are already familiar with your conventions.
For designers, we call these conventions “mechanics”. Things like Leader-Follow, Trick-Taking, Role Selection, Worker Placement, and the like are all names for the different shortcuts that we’ve created to work in a certain paradigm. Worker Placement is a fairly obtuse concept, when explained from scratch. However, for players who understand it, an entire Worker Placement system occupies only a single chunk of data. To an unexperienced player, though, it may occupy the entire working memory.
One reason gateway games are so important is to help new gamers establish these paradigms, by focusing exclusively on a single mechanic. For players who understand mechanics, these games seem derivative and uninteresting because they occupy so little of the Working Memory space, but to a player who hasn't developed those processing tools, even gateway games can appear complex.
If you’re designing for experienced players, then using familiar mechanics can be a key to reducing your game’s density. Familiar mechanics let you pack a lot of gameplay into a very small amount of working space.
If you’re designing for new players or casual players, putting your objectives close to the bottom of the tree is the best solution. Take great casual games like Dominion and Sushi-Go. The Dominion objective tree is pretty easy: Buy Stuff -> … -> Buy Victory Points, with the … being any number of “Buy Better Stuff” steps. In Sushi-Go, the objectives are printed on the cards themselves. Keeping your final scoring objective only one or two levels deep on the tree is a great way that these games reduce density and stay casual.
Working Memory vs. Short Term Memory
So far, we’ve only talked about Working Memory, but there’s another kind of memory that people have too: Short Term Memory. Short Term Memory lasts for about 30 seconds, and holds 7 chunks of data (plus or minus two).
Have you ever noticed that fast players are very fast, and slow players are very slow? If a turn takes more than about 30 seconds, it’s probably going to take more than a minute and a half, and the limitations of our short term memory are largely responsible for this disparity.
Short Term Memory is where we store the operands we’re processing–individual choices, moves, cards, or the results of our calculations on these things.
Two keys to reducing the density of your game are centered around short-term memory. The first is giving players the ability to hold all the pieces they need to process in short term memory at once. The second is making the processing simple enough to complete in under 30 seconds.
Our short term memory holds about 7 pieces of data, and for a casual all-ages game, we assume only 5 pieces. Furthermore, we need a spare slot to hold the results of our calculations. Beyond that, if calculation is going to require counting or multi-part evaluation, we need another slot to hold our intermediate results. Thus, for a casual game, the player should be choosing between about 4 pieces of data on a 2-level decision tree without any intermediate calculations.
Alternatively, for a more complex game targeted at experienced gamers, we can operate on about 5 pieces of data on a 3-level decision tree. It doesn’t sound like a big difference in numbers, but it’s the difference between Sushi-Go (Make Sets -> Score Points) and 7 Wonders (Gather Resources -> Build Buildings -> Score Points). Also notice the degradation of viable options in these two games. In Sushi-Go the pool of moves to consider is arbitrary at first, then very narrow, where in 7 Wonders, just about any card is as viable on your second draft as it was on your first, and many synergies exist between the cards in various ages, so forward-thinking is necessary.
When designing a game, it’s not enough to think about the game itself. You have to be conscious of how players are going to be perceiving and interacting with the information that your game presents. A lot of factors can contribute to this, and game design is only a small part of it. Graphic design, the presentation of the rulebook and cards, and the way the game is taught are all factors in how a player will interact with and understand the game. However, understanding the processing power of players can help you to make your game the best possible experience for its target audience.
- How deep is the objective tree in your game?
- How much short-term memory does the game demand from players?
- Based on this, what is your audience? Are you missing the goal range for the audience you wanted? What adjustments would you make to bring the game back into their preferred range?