Saturday, September 15, 2012

Negative Reinforcement

The other side of satisfaction is dissatisfaction. It can be a useful tool.  "Now, Alex, why on earth would you ever want players to be dissatisfied with your game? Isn't that defeating the whole point of a game? Games should make people happy!"

Actually, that's completely wrong. Games that only make people happy have a short lifespan. Eventually, the amount of positive stimulus you have to give the player eventually exceeds the production capabilities of your game. Bayonetta is a wonderful example of this.

I greatly enjoyed Bayonetta. However, the game grew linearly in both difficulty and epicness. The reaction times required to beat each incoming boss also increased.  The result was that my hands were seized up in painful cramps that forced me to regularly put the game down. Furthermore, the story events that took place in the game keep accelerating into absurdity. Youtube "Bayonetta Final Boss" if you don't care about spoilers.

Useful Uses for Negative Reinforcement

If you consider the pacing the macro level of a game, dungeon or encounter, you don't want players to be going balls-to-the-wall nonstop for the entire experience. To cater to their human nature, you want luls, breaks and breathing periods between moments of intensity. Players, however will continue to naturally seek higher and higher levels of intensity until they breakdown from exhaustion. 

You need to give them a hint that pushing forward harder is wrong. 

The first use of negative reinforcement that came to mind for me was attacking the walls in Legend of Zelda: A Link to the Past. Early into the game, you learn that bombs can destroy walls. 

Some walls have obvious cracks which indicate they can be destroyed. Other walls cannot.  To reinforce this, when you stab cracked walls with your sword, the wall makes a hollow noise. However, if you stab an indestructible wall, it makes a high pitched, unpleasant tinging noise. 

This generally discourages you from stabbing walls randomly to check for hollow spots but allows you to confirm you suspicions of a cracked wall without randomly spraying bombs everywhere.  

That unpleasant tinging sound carried over into other places.  When you fight bosses in the game, the same sound is used to indicate you are attacking him or her in the wrong way.  

This negatively reinforced sound thus shifted from an exploration tool to a boss fight feedback mechanic.  In general, Nintendo games are filled with these kinds of small polish point, which dramatically help players learn the boundaries of the game faster.

Dragon's Dogma similarly uses cheap deflecting noises and not the deep, visceral sounding wound audio effects when you're attacking a boss in a non-weak spot. 

Higher Level Uses for Negative Reinforcement

Early on in WoW's development, the design team wanted to penalize players who spent a lot of time grinding in the game and ignored the story-based nature of the questing. Similarly, there was a strong desire for players to not play endlessly. 

This lead to several systems. The first, was the rested system, where players eventually got 50% of the experience they would normally gain through killing monsters.  The hope was that players would be encouraged to do quests (which were decoupled from the rest system) and stop the endless camping of Owlkin in Winterspring. (I am looking at you, Zaibach)

The next system was the introduction of durability. Early on, without a durability system, players would zerg difficult camps of monsters endlessly, exhausting themselves, frequently getting frustrated and eventually giving up out of rage instead of looking for a different objective.  Bosses in dungeons were regularly kited to the entrance of the dungeon, where players would zone in, suicide on the boss, then corpse run back repeatedly. 
Durability made that strategy expensive, added an upper limit to the effectiveness of that strategy and generally served to be a very cheap death penalty.  Keep in mind that in this era, death penalties frequently consisted of XP and level losses.

Over time, durability became an increasingly smaller penalty, as repair bots, vendor mounts and geeves became prolific. The psychological effect remains in place - with many players leaving groups and quitting after a long series of wipes.

Negative Reinforcement can Backfire

You need to be very careful where you  use negative reinforcement. In small, well-placed doses, it is highly effective.  Used too often, players feel like their freedom is restricted. Too harsh and players feel punished for honest mistakes.  Too visible and players will constantly rally for its removal. 

Players hated the original rested system, complaining it did all of the above.  Some design teams would have panicked and removed the system. However, resistance to an idea doesn't always mean the idea is wrong. Instead, they rebuilt the system to give a 100% bonus while rested, rather than a 50% penalty while exhausted.  

This lead to the blizzard catch phrase "Make it a Bonus".  Generally, the concept of taking a systemic penalty, baking it into the system, then periodically granting players the ability to bypass that penalty.  It's a rather ingenious philosophy.

The LFG/LFR/LFD systems all use it - granting you a ton of rewards for the first completion of a dungeon, raid or battleground. Then far smaller rewards for continuing to play.


  1. An interesting sub-point: When using negative reinforcement to discourage behavior, it's vital to firmly connect the reinforcement to the behavior. If players don't understand why something bad has happened, they won't know how to avoid it.

    PS - I'm loving the blog. Thanks for taking the time to organize all of these awesome thoughts. :)

  2. This post makes me think of the Bloodlord Mandokir fight in ZG as an example of negative reinforcement being extremely confusing. I remember the first time doing that fight and getting decapitated I was so confused. With death being one of the biggest penalties in wow (at least during an encounter) the first time I got decapitated I spent the next few minutes trying to figure out what in the world I had done wrong to cause this penalty.
    I often try to think about things from the persective of a hyper casual player who doesn't study the game or research anything. Luckily my mom plays wow, and nearly everything confuses her. So that's generally an easy litmus test. But I think about that same situation and how it confused me, an experienced player, and I wonder what in the world my mom would have thought in that situation. Would she ever have realized that this was NOT negative reinforcement and instead the use of death as a mechanic?

  3. I would add "never punish a player with negative reinforcement for game defects". For instance if a player randomly disconnects during a PVP scenario and receives a debuff that prevents rejoining it won't take many of those before the player abandons the game entirely.

  4. Also, loving the blog. Keep it up! :)

  5. Alex, your blog is awesome. I'll say it again: It's concise, clear, and concrete. A world of difference from most blogs written by "game designers."

  6. The only real difference between negative reinforcement and a lack of positive reinforcement is the player's expectations. If they expect something, getting it feels meaningless, but not getting it feels like a punishment. If they don't expect it, though, then failing to get it doesn't feel like a punishment any more - it just feels like the standard game. Unfortunately, controlling the player's expectations can be pretty difficult...