Jeff's World of Game Design

Every take a hot take
 Thumb up

The science of learning, and its application to teaching games

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: I work with lasersMicrobadge: INTJ: MastermindMicrobadge: I love Haiti!Microbadge: The Sands of Time
Disclaimer: This is a really long post! I decided to keep everything together for ease of cross-referencing, instead of spreading it out between two posts. And that has the advantage of keeping it self-contained, making it easier to use as a reference should someone want to do that. Anyway, it's long. But in my defense it covers quite a lot of ground!

When I studied physics in college, I would often check out a text from the library different from the assigned text. The texts covered the same topic but in different ways, and more importantly, often used different nomenclature. e.g. one text would use "V" for the electric potential, another would use the Greek letter "phi". My mind had to work a little to find the commonalities between them, and I had unwittingly stumbled upon a principle of learning: it's more likely you'll retain something you had to expend some effort to learn.

From gallery of jwarrend

A few weeks ago we talked about design thinking and finding better ways to teach games, particularly for publishers. I told Prof. Stephen Blessing about the post. He runs the Cognitive Gamer podcast, and indicated that the topic interested him enough that he might do an episode about it at some future time. But in the meantime, he suggested two books, which I've since read (well, listened to), and I think their findings, heavily steeped in scholarship, point to some lessons that might help us improve how we present games.

First, let's summarize the findings, and then talk about their application to games and rules learning.

colonistcolonistcolonistcolonist The Literature in Review colonistcolonistcolonistcolonist

colonist The Nature of the Problem

The problem isn’t just learning, it is retention. We start forgetting things we’ve heard almost immediately after hearing them, and it only gets worse as time elapses. So our learning strategy can’t just entail how to present info but how to help the learner retain it.

corn Learning styles

There is no real evidence to support the "preferred learning style" concept, i.e. that I'm an auditory learner and you're a kinesthetic learner. What's more important is to match the presentation of the concept being taught to the concept itself, i.e. to teach throwing a baseball kinesthetically rather than auditorially.

coffee Retrieval and quizzes

Retrieval practice is the most important and most effective way to make learning stick. This specifically means being quizzed or tested on information you've learned. In one study, a middle school social studies class was quizzed on 1/3 of the course material, another 1/3 of the material was repeated with summary statements at the end of class, and the final 1/3 was not repeated at all. At test time, the students did the best on the material that was quizzed, and there was no significant difference on their performance with material that was reviewed and material that wasn't reviewed.

indigo Motivation

Motivation to learn comes primarily from goals (sounds familiar!), which has two components: value and expectancy. Value means what learning will do for me, which can be attainment (I will feel good when I am adept at this activity), intrinsic (the act of doing this activity is pleasant in itself), or instrumental (learning this skill is a means to some other end that I'm pursuing, e.g. learning plumbing to make money as a plumber). Expectancy means how likely I believe I am to actually succeed at acquiring the information; a belief that it's too hard for me creates a low expectancy and is de-motivating.

Additionally, goal-directed learning leads to improved performance. An example is that music students who "practice just to practice" don't progress in the same way that students who practice with a particular goal in view -- e.g. to learn some particularly difficult piece. I can attest to this! I never got very good at guitar, and I suspect part of the reason is that I just played to have fun, but never practiced with an eye toward enhancing my skill; doubtless this was because I didn't have a good sense of what I ought to practice so as to improve, so I just played what seemed to me to be fun to play.

sugar Competence

The expert vastly overestimates the facility with which a learner will learn something, because the expert has internalized mental maps or knowledge networks -- systems that connect knowledge into a robust framework -- that novices will not yet have created. There is a progression from "unconscious incompetence" (I don't know, and don't know what I don't know) to "conscious incompetence" (I don't know but I know what I don't know) to "conscious competence" (I know but have to think about it) to "unconscious competence" (I know and don't even have to think about it, performing the skill is effortless). Think about how one progresses in learning how to drive, for example.

Someone who has achieved unconscious competence may not be able to put themselves in the position of one who is consciously or unconsciously incompetent.

tobacco Mastery and Cognitive Load

Mastery entails acquiring component knowledge, integrating that knowledge to perform tasks, and knowing the appropriate situation and appropriate component knowledge to integrate. For example, in a physics problem, I need to know not just what formulas to use, I need to know how to combine formulas to get to a solution, and I need to know which formulas to apply to a given problem. A novice may not be able to do all three of these at once because of excessive cognitive load (sounds familiar!)

Scaffolding can help with this. If I'm teaching you a skill, I can supply some of the components you need, or I can supply some of the necessary integrative framework, so that you don't become cognitively overwhelmed trying to do it all.

bluetaj: Prior knowledge

Learning sticks most effectively when students can connect the new knowledge to prior knowledge. When teaching a physics class, instead of diving right in with Newton's 2nd law of motion, I gave an example of what we might expect if Lenny Kravitz were to push the planet Mars, or Bruno Mars. You can rely on people's experience and intuition to teach them something.

But! If something they think they know is wrong, or if they apply knowledge that's inappropriate to the situation, this can actually make things worse. For example, a behavior professor said that students often fail to understand the concept of "negative reinforcement", which they assume even after being told otherwise means punishment, because of the word "negative". You can sometimes correct inappropriate knowledge by supplying a better mental model, e.g. by changing "positive/negative" from "good/bad" to "adding/subtracting".

graytaj: Repetition is bad. Repetition is bad.

Sorry, couldn’t resist that one.

Repetition, re-reading, and "massed practice" don't really work. Moreover re-reading the same text may produce fluency with the text itself, but not with the concepts being taught. These approaches produce short term gains but they don't "stick". Better is varied practice, practice after a delay, and "interleaving" different skills. And, per coffee, retrieval is best of all.

Moreover, it’s better to space out retrieval from learning, to let a little forgetfulness set in. When learning is hard and the brain has to work hard at it, the knowledge acquired becomes more durable. One professor gave an example of how she gives the students a blank sheet of paper at the end of class and tells them to spend 10 min at the end of class writing everything they remember from class. Most stall out at two minutes but she insists they keep at it for the full ten minutes, working harder to try to retrieve the details they may have started to forget.

colonistcolonistcolonistcolonist Toward Better Ways of Teaching Games colonistcolonistcolonistcolonist

Some of these principles can easily be incorporated into our approach to learning games, but the concern of design thinking is how to present games to aid learning. So, how can we apply these principles to games teaching, and in particular rule book construction? I think a few concepts come through.

d10-0 Learn by doing

Although the points above don't explicitly call out a particular mechanism for presenting information such as a game's rule set, I do think that corn is supportive of the point I argued for in the aforementioned Design Thinking post, namely that a "learn by doing"/"tutorial" approach is preferable to a "rules dump"/"lecture" approach.

One admonition that the books endorse is "practice how you play and you'll play how you'll practice". A batter who, in batting practice, takes 20 fastballs, 20 curves, and 20 change-ups, is less well-prepared to face live pitching than a batter who takes 20 pitches of a mixture, because in live pitching you don't know what's coming next. Practice that better replicates the performance environment prepares you better to perform in that environment. So too, perhaps, with games: learning to play while playing is more representative than being lectured about how to play, and lessons learned in the context of actually playing may stick better.

1 Quizzes not examples

Everyone at BGG has seen those promotional quizzes that publishers host when launching a new game (actually, quite pricey to run!) in which you get more chances at winning a free game the more rules questions you can answer correctly. These may or may not be good marketing, but as pedagogical aids these are exactly what rulebooks should include more of. Examples are fine and have their place but, per coffee, questions do more to activate learning and should be more broadly implemented.

And, per graytaj:, it's good to have some delay between when a concept is taught and when it is quizzed. If you first learn the movement rules, then the combat rules, perhaps a quiz question about the movement rules after the combat rules is appropriate, both because the delayed learning is more effortful and because you're interleaving the concepts of movement and combat. Both of these support graytaj:.

2 Use prior knowledge to your advantage

We often talk about how a game’s theme provides an aid to learning the game’s rules by situating the rules in a context already familiar to the players. The scholarship supports this very strongly, as bluetaj: indicates. If a game has a "build" mechanic, well, I know what building entails in real life and so I can use those existing mental maps to aid in my learning your game's build mechanic.

But it's only useful to the extent that the prior knowledge helps, which means that your theme-driven mechanics must actually connect with the knowledge people have. If there's a rule about "going to the market", but it involves rolling dice in a press-your-luck way, that doesn't actually leverage the mental maps players have about what "going to market" entails.

Relatedly, it can be a mistake (and one that LOTS of publishers make) to rely too much on familiar gaming concepts like worker placement or deck building. For one thing, some players won't have played those sorts of games and won't have those skills, and for another, the specific way your game implements those mechanics may be sufficiently different that prior familiarity with those mechanics is actually anti-helpful in learning your game.

I recall teaching a game of mine by saying how the auctions were "Dutch auctions but with a twist", and after I kept saying this a player said, "look, I don't know what Dutch auctions are, but since you're giving them a twist, this is doubly unhelpful, just tell me how your mechanic actually works!"

3 Break rules down into component skills

Better than The Lecture is to decompose a game’s rule set into components and teach those components individually per tobacco.

My teaching approach tends to entail presenting the turn structure and then walking through the turn options, but I now suspect it might be better to instead present the individual turn options as skills to learn, and build those one by one, and only then progress to how you select among those options using the turn mechanic.

An example might be to teach Broom Service not by starting with the turn mechanic (pick 4 cards, then first player plays one, then go around and follow, fold, or "trump"), but rather to pick one card and talk about what it does: this one gives you a green potion. But look, it has a "brave" and "weak" version, the "brave" giving you more. Ok now look at this similar card, which gives you brown potions, but is otherwise the same. And once players get that, now we're ready to talk about the turn mechanic and how we come to pick either the "brave" or "weak" action.

This is important because of sugar; as a player familiar with the game who understands the turn mechanic, it's easy for me to integrate the component actions into the larger construct of how we interact and compete against our opponents, but newer players may find it easier to progress from incompetence to competence if they don't have to start by thinking of that path in terms of the mental maps that an experienced player has already internalized.

4 Provide an integrative framework

My approach to teaching heuristics is usually predicated on the assumption that players want to acquire these for themselves, and this is especially the case for playtesting: you don't necessarily know what the right heuristics are! But I think this too may be misguided.

Players are cognitively overwhelmed by trying to perform the basic operations of the game while also trying to figure out what they're supposed to be doing to win, and so they fumble around to no particular purpose. And, we encourage this, telling them "just fumble around until the turn makes sense, the strategy layer can come later". In light of tobacco, players need more "scaffolding" to know how to connect their individual action-level decisions to turn-level and multi-turn-level tactical and strategic decisions.

What might this look like? In my game The Sands of Time there are six scoring categories, two each in the three "civ categories", and advances, buildings, and actions all are part of one of those three civ categories. So the simplest approach is, pick any scoring category, and then seek to take actions and advance and build in that scoring category's associated civ category. But the rules don't make this simple scaffolding explicit; they should!

The online tutorials for Root do this; you start as the Cats and learn how to complete simple objectives like building a sawmill or controlling three clearings, which in turn introduces you to the "build" and "move" mechanics, how to do them and what they look like when done.

Or as I've previously said, the species sheets in Sidereal Confluence give you some entry-level heuristics but also some entry level ideas about what you should be doing. If you're the Yengii and can invent a tech, try to get two cubes for it when you license it, that's a fair deal. So now I know that I want to invent a tech, and then try to sell it to my opponents, and that I should try to get two cubes when I do. That gives me a target.

By combining scaffolding (tobacco) with goal-directed learning (indigo), we help players to improve both their familiarity with the basic functions of the game AND with how to connect those functions to pursue goals, which will aid in their learning.

5 Anticipatory teaching

Another way prior knowledge (bluetaj:) can help is in giving players pre-emptive understanding of some tasks. Studies showed that asking learners what they thought the lesson was going to entail, even when they were wrong, helped them to better apprehend the actual content of the lesson. We can do that with games, too. "I'm going to teach you how combat works; what do you think I'm going to say next?" As they acquire knowledge about the moving parts of your game, they will better and better be able to anticipate how the next moving part works, but even when they guess wrong, the act of guessing will help make their learning more purposive and productive.

So, rulebooks should ask leading questions at the start of sections, not merely presenting the rules but encouraging players to think about what their intuition tells them "movement" or "combat" is going to involve.

6 Offer learning objectives

Every rulebook specifies the objective of playing the game: "to make the most points", "to be the first to build a space shuttle", etc. But few, if any, present the skills you'll need to acquire and master to play the game well. By decomposing a game into a requisite set of skills (tobacco, and Ch. 9 of You Said This Would Be Fun), players can better understand what they'll need to learn so as to become good at the game.

For example, soccer's goal is "to kick the ball into the goal", but the skills you'll need to acquire include running with the ball, kicking the ball hard, passing the ball, and taking the ball away from an opponent. In the process of teaching you how to play soccer by presenting you the rules, I also want to work into the conversation a presentation of these skills. If I only taught you the rules of the game, I haven't prepared you to actually play the game well.

So, if I'm teaching you Can't Stop, after telling you how to split your dice into two pairs, we're of course going to talk about how the track lengths tell us something about the relative probability of obtaining those results; 6 and 8 are pretty safe, 5 and 9 also reasonably so, 4 and 10 much less.

Or if I'm teaching you 7 Wonders Duel, instead of just giving you the basics of the three actions you can take (draft, income, wonder), maybe I'm also going to have you draft a red card, a green card, and a blue card, and talk about how each contributes to one of the possible ways of winning (military, science, points) so you understand that "victory" means winning along one of those three trajectories. And then at the first opportunity, I'm going to try to show you how to block, by pointing out a card that would be useful to me, and showing how you can use such a card for income as opposed to taking it for yourself.

Of course these are examples of a person teaching the game, but a rulebook tutorial can be similarly structured. "You win The Castles of Burgundy by having the most points, and you play by doing one of four things with your dice, but the skills you need to acquire to play well include knowing how to balance between filling an area early (more points) vs pursuing multiple building types (usually more efficient), and (etc)."

Or "To win Tiny Towns you have to make the most points from the cards, but first let's explore concepts like what it means for a building to be 'fed', and then talk about whether you want to get the feeder building in place early or late, (etc)".

The other advantage of presenting learning objectives is that you're telling players what they're going to get out of learning, which is motivating (indigo). In Ch. 10 of You Said we talk about how the scoring system is the contract between the designer and the player: pursue the goal, and you'll have fun. Knowing how I'm going to progress from incompetence to competence through pursuit of these learning goals helps motivate me by showing me tangibly what the value of this rule set is going to deliver -- attainment and intrinsic motivation, hopefully -- and it creates expectancy as to what level of effort I'm going to have to expend to achieve mastery, which in effect spells out the terms of the contract between me and the designer. "When you understand how each area of the player mat functions in Terra Mystica, you'll be able to play any faction by simply noting the specifics unique to that faction".

colonistcolonistcolonistcolonist Summary colonistcolonistcolonistcolonist

Doubtless there's much more to be said about how to use the scholarly literature in learning to devise better ways to teach games, but I hope this post will provide a useful starting point and will stimulate interesting thinking and interesting discussion about how we can better approach presenting game rules.

The great advantage of this literature, to me, is that it's systematic and rigorous. Whenever you talk about teaching games, everyone chimes in with that rulebook they liked or the way that they teach games and how well it works, and all that stuff is great, but the entire point is that as far as publishers are concerned, it's the wild west, and that means that whether you learn a game well or poorly depends entirely on whether you have a good teacher or poor in your gaming group. If I were a publisher, I would NOT want to leave my game's fate in the hands of amateurs like that, not if I could do anything about it. This body of literature strongly suggests that there are proactive steps publishers can take to do a better job.

The suggestions and ideas I propose are presented as food for thought, but readers who draw different strategies from the literature are still on the right path; ultimately, the literature is the thing. He who has ears to hear, let him hear!

colonistcolonistcolonistcolonist A Final Request colonistcolonistcolonistcolonist

This post required quite a bit of effort and even a few dollars of investment! And I think it's an important and worthwhile conversation. If you agree, please be encouraged to "like" the post to see if we can get it onto the BGG front page, and please also be encouraged to share it on social media to bring outside attention to the conversation. I'm by no means the right person to teach people about learning or about rulebook writing, but I hope that in engaging the topic, it will encourage readers to think for themselves and seek knowledge for themselves. Your simple upvote helps in this endeavor, so thank you!


If you enjoyed this post, please consider subscribing to the blog. If you didn't enjoy it, consider subscribing anyway! We talk about all sorts of stuff here and the next post might be more to your liking!
Twitter Facebook
Subscribe sub options Sun May 16, 2021 1:00 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}}