Using randomness better in learning experiences

Or, how to go beyond simple dice, spinners and card draws

There are few concepts the average person associates with board games more than the idea of randomness. Rolling dice or drawing cards is part of Monopoly, Yahtzee and Catan. But not all randomness is the same. How you implement randomness — or whether you use it at all — can make a big difference to a learning game or playful learning experience.

Here’s a practical low-down on the different kinds of randomness and uncertainty and how you could introduce them into your learning game, or playful learning experience.

Randomness provides helpful uncertainty in your learning game

Chess would seem to stand as an example of a successful game without randomness. But a significantly better player will almost always beat a lesser player. Not 60 or 70% of the time but 99% or 100%. That’s not much fun for the weaker player. And, because there’s no luck, there’s always a ‘best’ move, so players can take a long time trying to figure out if they’ve found it.

If these ideas fit your learning experience, fine. But in games with chance, the odds are more even and less experienced players have more of a chance, which is often good for learning games. When outcomes are uncertain, there’s no ‘best’ move, so players are less likely to overthink things. And randomness adds tension and simulates the randomness of real life.

Randomness is just one kind of uncertainty

Before taking a dive into different random elements, though, ask yourself whether it’s really randomness you need. Randomness is actually just one kind of uncertainty. In fact, it can often be uncertainty that adds tension and depth to a game, rather than randomness specifically.

A different source of uncertainty is hidden information, which can fulfil some of the same functions as randomness. For example, you could have some face-down cards that can each affect things differently if they’re chosen and turned over. Then, if you wanted, you could have clues as to which card is which, or allow some players to look sometimes. And because the choice is with the player, it can feel fairer and give more agency.

Sports introduce uncertainty by means of player skill or performance being a factor. You don’t know what will happen — whether the favourite will win, for instance, because they may just find or not find their rhythm and best performance on the day. You could introduce this kind of uncertainty by adding a skill-based element like hitting a target.

Even Chess has uncertainty, without randomness. Will your opponent take the piece/sacrifice you’ve offered? Is their attack all it seems, or just a feint? Even Rock Paper Scissors has this kind of uncertainty — variations on its idea of one player choice interacting with another player choice to determine the outcome can be a great source of generating a result, instead of dice.

Be clear about when the randomness will happen

In Snakes and Ladders, there are no choices. You roll a dice and move that many spaces. A learning game shouldn’t be like this: it should contain meaningful decisions. The key question then becomes: should the randomness come before or after the decision?

When the randomness comes before the decision, that’s input randomness. Like if you have to draw some cards at the start of your turn and then decide what to do with what you just got. This kind of randomness tends to focus you on the choice you have to make — it’s more strategic, more like a puzzle.

When the decision comes first, that’s output randomness. Like when you decide to try something and then have to roll dice to see if you succeed. This kind of randomness can leave players at the mercy of unlikely or unfair-seeming results but some argue that it creates better tension, with a better ‘story’.

Which you use should depend on what you’re trying to do but just thinking about the two in relation to your game or experience can help you make a call.

Think about how each random event is linked to the last

If every random outcome is independent of the last and if every available outcome is equally possible, we call this ‘white noise’ randomness. Like if you’re creating a game around a series of elections and randomise voter turnout at each election by rolling a single dice: on a six, you get 60% turnout, on a 1 you get 10% and so on.

The problem with that is it feels very random and life isn’t really completely unpredictable like that. Things change over time, based on how they were before. Instead, you could get players to roll a dice that has faces saying ‘-3%’, ‘-2%’, ‘-1%’, ‘+1%’, ‘+2%’ and ‘+3%’ and change last election’s turnout by that much. Then you’d be using brown noise randomness — things change a little, in fairly predictable ways.

This tends to be more interesting to players but the best is usually what’s called pink noise randomness, where changes are usually small but there’s also a small chance of big changes. The stock market is like this. Most days it rises or falls a few points. But some days it falls off a cliff or jumps massively.

For more detail on white, brown and pink noise in games, this video by Geoff Engelstein explains it well.

You can generate ‘pink noise’ results fairly easily

The key to more interesting noise and its benefits is using dice (or other randomisers) to give a change to an existing value or state — not to create a new value each time. If you set things up so that small changes are most likely and bigger ones less so, you’re using pink noise.

So, for example, you could use a 20-side die, with a chart like this:

• 1–6 = +1
• 7–11 = +2
• 12–15 = +3
• 16–18 = +5
• 19–20 = +8

This means a little over half the time, the change will only be 1–2 points. But sometimes it may be as many as 8!

You could also:

• Let people roll a dice but say if you roll the highest number, you add a certain amount and then get another roll
• Use cards and stack a deck so that there are, for instance, very few ‘+10’ or ‘+20’ cards, but many ‘+1’ or ‘+2’ cards

Remember, if you use this idea, what you’re trying to do is make it so that the result is easy to predict with a reasonable amount of accuracy but impossible to say with certainty. Also that the resulting state of the game is likely to be similar to how things were before but may occasionally be wildly different.

This feels more realistic, is often more fun and opens up a space for strategy and tactics because people can plan for probable outcomes, or ride the outside chance of a less likely outcome if they want or need to.

Try to reduce ‘felt unfairness’

People don’t have a great intuitive grasp of probabilities. Many subscribe — consciously or unconsciously — to the Gambler’s Fallacy (the idea that if a number hasn’t come up in a while, that it’s more likely to come up next). And most people’s brains process a 90% chance as ‘more or less certain’. This kind of thinking can lead to things feeling unfair, even when they’re not.

You can avoid this by having some kind of mechanism that evens things out. Rolling many dice means that it’s much more likely to get a result close to the middle of those possible. Drawing from a card deck (or a token from a bag) means that each card or token will come up eventually.

Think about other ways to decide things

Remember that most of the time, randomness is a way of deciding something — how many squares somebody will move, whether they win a battle, how much a certain resource costs next turn. If that’s what you need — a way to decide something — there are plenty of interesting ways to do it besides randomness.

For example, to move, a player could have a limited number of cards and choose which one to play each turn. They can make the big move this turn or next turn but not both. And they can’t use it again until they’ve used all the cards, including the small or backwards moves. Other mechanics are plentiful, such as some of the ones on this list of game mechanics — check out in particular the categories ‘Movement’, Alternatives to Movement’, Alternatives to Turns’ and ‘Conflict Resolution’.

One of my favourites is limited resources having to be split between priorities: you have a number of tokens, representing cash or time. Do you spend them on moving to the space you want? Or do you save them for getting a result when you get there? Ideas like this can even be combined with randomness — you roll dice but you also have some one-shot ‘pips’ that can be spent to re-roll or increase dice results.