Fairness is a Perception [Game Psychology]

(with thanks to Sini Merikallio)

Fairness is one of the most important aspects of a good game, but it is rarely straightforward. While a game’s rules might be balanced, the player may feel that the experience is not fair, and this is a source of design tension.

The reason is that fairness is not an objective quantity. It’s subjective. Games are fair when the player sees that his actions in the game are achieving tangible rewards, even if the game is cheating to provide them to him.

 

Fair Actions vs Fair Outcomes

A referendum in the United Kingdom proposes to change voting from the first past the post system to the alternative vote system. Political leanings aside, the interesting part of the debate (for this blog) is the argument over what fairness is. Both sides claim that their method of voting is the most fair, but they mean very different things.

The first past the post system argues that the vote at the ballot box is a fair action. You pick your horse, and the MP who gets the most votes wins. My vote carries equal weight to yours because we all get to check a box or pull a lever, and the result is the result. The elected MP represents the whole constituency, but the implicit understanding is that even though he may have not achieved an overall majority, his responsibility is to all. It’s easy to explain to the voter that everybody has one vote, easy to show that each is equal, and so the activity of voting feels fair.

The alternative vote argument, on the other hand, is about fair outcomes. It argues that every MP who gets elected should have won support from more than half the people in their constituency even if that means using preference-based voting. Their view is that Parliament should have a number of MPs that is more in line with the popular vote than historically has been the case. First past the post tends to allow minorities to rule the country (no ruling party has had more than around 40% of the popular vote for a long time) and so their position is that the outcome is unfair.

Outcomes are harder to relate to than actions (which is why the first past the post campaign is likely to win the referendum). Fairness is emotional, so from a game design perspective it’s the actions of the player that matter, not the outcome. The player needs to know that he is the active agent of change. If he feels at the mercy of the game, or that the game is toying with him or being opaque, then he feels like a rat in a maze.

 

Adjusting for Fairness

A rat in a maze is trapped by a series of arbitrary walls, switches and food pellets. While that may be perfectly fine for a rat, a human being is generally more perceptive and aware of their situation. The sensation is not one that is enjoyable. Far from feeling fair, the game that does this feels arbitrary and unfair.

Tiny Wings is an example of a game with this problem. It’s an iPhone game in which you play a little bird that can’t quite fly, but can swoop. Positioned correctly, the bird can use the curvature of the land as acceleration ramps achieve a sort of flight, and the action of doing this well is a real thrill.

The issue is that the game’s level and reward layouts are randomised, so it’s impossible to get to know the landscape of the islands better. You can’t reliably predict where the right zones are to drop down from the clouds, how to achieve sequences of boosts, and so succeeding in the game relies more on the luck of a good level layout than any particular act of skill. It places the player in the role of a rat hoping for a good maze.

I can see why the developer has done this. He is trying to keep the islands in the game interesting and continually challenging, to provide a refreshing casual experience. He is also trying to do so without having to spend a lot of time manually creating island layouts, which for any small indie game is an issue. The problem is that the result eventually feels unfair.

Tom Chatfield uses an example of how massive multiplayer games manage item drops to describe the psychology of rewards. A player may need to collect 15 items in order to complete a task, and those items have a 10% chance of appearing. The player gathers the items at a steady rate, but as he approaches the finish line, the last couple of items do not appear as the he expects. He thinks that the game is cheating, but it isn’t. It’s still using the same 10% chance.

But still the player feels as though the game is not being fair, and so dislikes playing. The solution is for the game to raise the chances of the last few items to drop to 15%, 20% or even 25%. It deliberately cheats on behalf of the player to make it more likely that he will complete the task. The actions now feel fair, even though the outcome is skewed, the player feels as though they are progressing toward something great by dint of their action.

 

Investment Issues

Racing games like Sega Rally and Mario Kart have long used active cheating to help poorer players catch up to better ones and keep the races close. This is great for new players, however for invested players (those who have played the game for a while) this sort of adjustment often feels unfair.

An invested player expects the game to have fair outcomes more than fair actions. They feel that they had to put a lot of work, time and money into the game, and so should everybody else. Invested players get angry if they feel that all of their hard work is for nothing. They hate the sense that other players are cheating, or that a game is cheating on their behalf, because the game and its world means something to the invested player. They don’t want that meaning to feel futile.

The concerns of new vs invested players are often very different. A new player tends to want the game to award them early and often and feel as though they are empowered, whereas an invested player wants to achieve things that matter to them. They are motivated by the personal goal of the epic win rather than the simpler wins of ordinary actions, and at some point their concerns need to be your concerns too.

This is why a game like EVE Online has attracted a significant core of die-hard fans. EVE is famously strict in its interpretation of fairness. It gives players a universe to play in, but essentially tells them that they are on their own. Players are free to wheel, deal, shoot, scam and otherwise compete with each other and EVE will not intervene. EVE has a very loyal base of users as a result, but is also imposing for outsiders. While it does have safe areas for new players to get to grips with the game, it’s a lot less forgiving on the whole than many other massive multiplayer games.

And for the EVE audience, that is exactly how things should be. CCP made the choice to look after their invested players more than their new players, and over the long term this has proved to be a very smart choice. For other games, the opposite choice is the right one.

 

Cheating for the Player

It’s important for a player to be able to achieve some amount of mastery in a game because that is how they become invested in it. In general, the best course of action to try and solve unfairness and promote mastery is to reduce randomness.

Whether through adjusting level content, or manipulating hidden factors to increase chances of winning, the result is the same: Less randomness and more control lead to greater opportunities to achieve, and so it all feels fair. Tiny Wings with fixed levels and a better implementation of the sun rule, for example, would be a much better game to play over the long term than it currently is.

It’s easier in a single or co-operative based game to manage the perception of fairness, but in a competitive game especially it’s very easy to get it wrong. At some point the game experience has to be about more than just the simple joy of action (as Tiny Wings illustrates) or else there is little reason to carry on, but at the same time there is the need to keep the new players feeling fair too.

Some games solve this quandary by allowing active cheating options to be turned on or off. Another solution is to have different game modes, or difficulty levels, that alter how the game plays.

It’s better, if you can, to not make players formal choices for which they have no context. It helps retain some sense of charm in the idea of the game if the player is kept away from switches and dials that tune the game, as they often serve to remind that the game is just a system. It is, but a little showmanship to cover over the components never hurts.

Comments

[Esotericsean]

Tiny Wings isn't procedurally generated (the maps don't change -- only their colors do), but they're so "complex" that it's nearly impossible to memorize them. I'm fine with this, though. I've played hours upon hours of Tiny Wings and have never felt cheated. Maybe it's the calming music and sound effects. That said, Mario Kart does frustrate me sometimes when you've worked your way to first and then you're blasted with a purple shell. Pretty interesting.

[Tadhg]

They don't change every time you play, but there definitely are some elements that move around.

In particular, the placement of items like the blue boosters changes. Some of the map elements also seem to be included or not occasionally. For example, on Island 2 there are usually two down slopes with small lips upward, but today there is only one (in my version at least). However the procedural elements are working is very clever, to the point that it might well be imagination on the part of the viewer sometimes to think they've moved when they haven't.

The net sense of the unpredictable effect is bad though. I noticed in particular that it was easy enough to get up to Island 6 or so, but then I hit a wall where there was no obvious way to improve, and sometimes I got there with sequences of boosts that were differently placed the following time. Sometimes there are two steep hills on Island 6, and again sometimes not. I also hate that the Sun rules pretty much kill any chance of succeeding past an island if you screwed the previous one up badly but made it. Again, rat in a maze.

What I think saves it (not to detract from Andreas' work) is that the act of getting the bird to fly is a real treat, and when you manage several in a row it's a thrill. The production values of the game are also truly lovely.

It's an interesting case regardless, as it's making the point that fairness is all about perception. By its nature it will be subjective, and often even plain wrong on the part of the viewer in terms of strict information.

Thanks for the comment!

[Jonathon Duerig]

I definitely agree that being apparently fair is extremely important.

I think that randomness must be manageable by the player to feel fair. Perhaps tiny wings would feel more fair to you if you could pick up boosts and still have a short time to use them for instance.

Whenever I design the basic systems of a game, I take pains to make sure that everything errs on the side of the player. When colliding with an enemy, only harm the avatar if they are clearly visually connecting. When colliding with a ledge, err on the side of letting the player make the jump. I try to do this kind of thing with every basic system.

Then when I balance for difficulty I make the enemies a bit bigger or the ledges a bit farther if I want to challenge the player. The player will not balk at missing a long jump in a platform game, but they will ragequit if they see their avatar apparently hit the ledge and fall through anyhow.

In a game world, there is no such thing as true fairness. There is only the appearance of fairness and a (hoprefully) balanced challenge curve.

This is a lot trickier for multiplayer games as you discuss, but I haven't had to worry about that yet. :)

[Account Deleted]

I'm a hardcore Mario Kart Player and I feel all the issues described by Tadhg, and I put a lot of thought to it.

In Mario Kart Wii, the probability for spiny shells do feel highly unfair, specially when sometimes you have like 30 secs of advantage and you receive three spiny shells in a row on the last half of the last lap. It just insane. But after considering some selection biases of my own and having faith that Mario Kart Wii designers are not "that bad", I like to believe that the probability actually does not change and it's just a perception problem.

Nevertheless, I was facing the same question: How would I change this probability to feel fair? What would be awesome is that any hurdle on the last lap should consider the possibility of, if affecting you, leaving you enough time to reach a particular position at the end of the race on the time remaining. For example: if you were consistently leading the whole race, last lap disasters feel highly unfair, but it would be a thrill to suffer them at a point that assuming almost perfect racing, allows to feel a "fantastic recovery" from the circumstances and win the race anyway.

That feeling sometimes can actually happen on multiplayer races, but how to modify the odds in the end in order to enable this "fantastic recovery" feeling? Well, that depends on the particular model of the race and the game data, but it shouldn't be that hard to do. Maybe for leading position is simple enough, but modeling a fantastic recovery for all the other positions, or worse, fantastic recovery for every player on a multiplayer race, seems a little bit harder.