Abstract
Problem: How should game designers implement punishment mechanics so they feel fair and don't drive players to save-scum or quit?
Approach: Tim Cain examines the carrot-and-stick model of player behavior, drawing on his experiences with EverQuest, World of Warcraft, Fallout, and The Outer Worlds to identify what makes punishments acceptable versus rage-inducing.
Findings: Effective game punishments follow three principles: they must be proportional to the offense (fair), they must be communicated in advance (telegraphed), and they should be bundled with rewards so the player accepts the negative as part of a larger package.
Key insight: The best punishment systems don't feel like punishments at all — they feel like consequences the player chose, because the game warned them, gave them agency, and paired the stick with a carrot.
The Carrot and Stick Model
Game design uses what psychology calls the carrot-and-stick approach: rewards (carrots) and punishments (sticks) guide players toward the intended experience. Every designer wants players to have the most fun possible, and they use a combination of both to steer behavior.
Some games lean heavily one way. Tim contrasts two MMOs from his experience:
EverQuest was all stick. Tim tells a story from 20 years ago: he gained a level and received a teleport spell to a city on a new continent. He teleported there, walked outside expecting a newbie zone, got killed by a high-level creature, lost experience, lost the level, lost the teleport spell, and respawned on the original continent. It took him hours of running and ships to get back to where the spell had taken him instantly. That's a brutal punishment loop.
World of Warcraft felt like all carrot. The sticks in WoW are very small and the carrots are very big. Blizzard guides players by rewarding desired behavior (XP, items, quest completion) rather than punishing undesired behavior. Doing the "wrong" thing simply doesn't reward you — which many players interpret as a waste of time, a soft punishment in itself. This is positive reinforcement, and it generally works very well.
When Negative Reinforcement Is Needed
Sometimes positive reinforcement alone isn't enough. Negative reinforcement has its place when it prevents behavior from escalating.
Tim gives the example of attacking someone in a town in an RPG: guards come after you, but if you didn't kill anyone and put your weapon away, they stop. That's negative reinforcement under the player's control — a warning system that prevents you from doing something worse (like killing a townsperson, which makes the guards attack until you're dead).
The key distinction: negative reinforcement should give the player a chance to course-correct before consequences become permanent.
The Save-Scumming Problem
A critical difference between tabletop and computer RPGs is the save game. In tabletop, you can't reload. In computer games, players do it constantly.
Tim identifies two groups: a small group that reloads to avoid all punishment, and a much larger group that has a threshold. They'll accept losing a little money, but if the game takes XP? They're reloading. Everyone has this threshold, and designers must account for it.
This is also why stories, classes, and systems designed for tabletop don't translate directly to computer games — it's a fundamentally different medium.
Principle 1: Make Punishments Fair
Punishments must fit the crime. If a player does something small and gets a huge punishment, it feels unfair.
Tim describes several unfairness patterns:
Accidental triggers with disproportionate consequences. Clicking on an NPC to talk but accidentally attacking them, then guards killing you. Picking up an item that's marked as stealing, triggering everyone to attack. These are tiny actions — sometimes completely accidental — that trigger massive negative responses.
UI-caused punishment. Tim recalls an Ultima Underworld story: he tried to give an item to an NPC, but the NPC's animation turned them away, causing the item to fly past into lava and be permanently destroyed. The game's interface tricked him into losing a critical item. When the UI causes the punishment, players feel cheated and will almost certainly save-scum.
The rule: if a small or accidental action triggers a huge punishment, players will reject it as unfair.
Principle 2: Telegraph the Punishment
Players should never be blindsided by punishment. Having a consequence hit you that you had no way of anticipating feels unfair because you had no opportunity to factor it into your decisions.
Tim gives concrete examples of good telegraphing:
Warnings before trespassing. Instead of instantly attacking the player for crossing an invisible boundary, have an NPC say "Hey, you need to get out of here. You're trespassing." The player then has agency: they can back up and leave, or push forward knowing what will happen. If they're then attacked, they feel ownership of that outcome.
Environmental cues. If a game has jails and you see prisoners and guards everywhere, you can infer that doing something illegal might get you jailed. When it happens, you expected it and possibly even planned for it.
The principle: the player should be able to say "I knew that could happen" after any punishment.
Principle 3: Bundle Punishment with Reward
The most effective way to get players to accept punishment is to glue it to something positive. Tim designed this way across multiple games:
Fallout's trait system. Every trait came with something good and something bad, inseparable. You couldn't get the benefit without accepting the drawback. Players chose these trade-offs willingly at character creation.
The Outer Worlds' flaw system. Years later, Tim evolved this concept into mid-game flaws. The game detected patterns — you keep getting attacked by robots, you keep falling and taking damage — and offered to formalize that into a flaw (making it happen more often or be harder to avoid). The trade-off: you get extra perk points. The punishment was paired with a reward, and crucially, you could always say no.
Quest consequences. Even without formal flaw systems, quests can bundle punishment and reward naturally. Rescue the prince and you get money, but the kidnappers now hate you. Steal the necromancer's orb and complete the quest, but the necromancer becomes your enemy. When the punishment is telegraphed as part of the quest, players accept it as a natural consequence.
Summary
Tim's three principles for effective game punishment:
- Make them fair — punishments should be proportional, never triggered by accidents or UI quirks
- Telegraph them — players should always be able to anticipate consequences before they act
- Combine punishment with reward — package negative outcomes with positive ones so players feel they made a trade-off, not that they were penalized
References
- Tim Cain. YouTube video. https://www.youtube.com/watch?v=kLbRnZDEYX4