Perceived Fairness

Fairness is everything in video games. Difficulty tuning, randomisation, matchmaking, there are so many things that can make a game feel unfair, which can lead to bad feedback from players.

Today we’re going to look at randomisation, and how what the numbers say aren’t always how it feels to players.

The 99% Shot

If you have played a tactics game with percentage based attack outcomes, you’ve been here before. You go to attack an enemy and it says 99% chance to hit, and it misses…

If a shot is 50/50, you feel great if it hits, but don’t get upset if it misses. We’re prepared for that outcome. At very high percentages to hit, you have already mentally banked the win and started planning for it.

Though mathematically correct, humans are very bad at perceiving probability. We unconsciously round that 1% down to 0. When we read 99%, we’ve already assumed victory.

The Losing Streak

Gamblers fallacy persuades us that a short sequence will reflect a long term average. For example, we expect to win 5/10 coin flips, or to roll a 6 at least once if we roll dice 6 times.

Because of this, we see repeated failure as something that shouldn’t happen, that it must be a bug, rather than chance working as intended.

Even worse, negative events like this stick in our memory very easily. Players are much more likely to remember a streak of bad luck over any amount of average results.

Given that we want to include chance, but don’t want to upset our players too much, what can we do?

Expectation vs Outcome

We can fudge the numbers here behind the scenes to rebalance the game for perceived fairness. It means that we aren’t showing players the true chance to hit, but they are experienced what they expect to with those numbers.

For example, we may increase the chance to hit at 95%+ so that they never miss, so we let players keep the promise their brain made to them about their decision.

We could also break bad luck streaks, but guaranteeing a good result after X bad results. You will see this a lot in gacha games that give players a good drop after X to make sure bad luck doesn’t make the player drop the game.

Fantasy

We should of course also consider the fantasy of the game and how it interferes with player perception. Would a highly trained soldier actually miss a shot from close range, with a shotgun? Probably not, so then we perhaps need more deterministic rules that can influence the chance.

Alternatives

We might also consider some alternatives. For example, deckbuilders don’t tend to have accuracy, but use a ton of randomisation. However, as I covered in my input vs output randomness article, a lot of the chance here is in the drawing of cards. This is actually great, because while you may not get the card that you need, as you draw down from the deck the possibilities narrow, meaning that instead of true randomness, you get delayed determinism.

Maybe we could apply this principle to xcom damage or some other plain % success chance? If there are a bad of tokens labelled 1-6, you might pull a 1, but you know that your chance to pull a 6 just increased.

An example of something similar in action is Citizen Sleeper. Each day you roll a pool of dice, which could be good or bad, but it is up to you how you spend them. Rolling a bunch at once mean your chance of a total dud is lower, and you get to choose where your good and bad luck will go. This keeps risk management and chance, but puts the control into the hands of the player, which I think is really smart.

Conclusion

When working with chance, keep in mind that human beings are very bad at understanding why bad luck comes their way. Most likely if you are using unaltered percentage based outcomes, you will run into bad feedback from players, so consider some more advanced approaches!