The following was published in the April 2012 issue of Game Developer magazine…

A zero-sum game is one in which the gains of any one player are balanced out by the losses of all the other players, such as winning a pot of chips after a hand of poker. Using strict game theory terminology, many competitive games are not actually zero-sum. Scoring a field goal in football, for example, does not take three points away from the other team.

However, more loosely speaking, the phrase “zero-sum mechanics” can mean that hurting one’s opponent is as equally valuable as helping oneself. In a typical RTS like StarCraft, a rush strategy, which aims to destroy the enemy’s economy as soon as possible, is just as viable as a boom strategy, which focuses on building up one’s own economy. If one can quickly wipe out the enemy’s first units, it’s irrelevant what level of development one’s own troops ever reach.

Thus, whenever a game rewards the player equally for hindering the enemy as for strengthening herself, the game has a zero-sum mechanic. Most team sports (basketball, soccer, football, etc.) share this characteristic; the defense, which prevents the opposition from scoring, is just as important as the offense, which does the scoring.

Competitive games are firmly rooted in this soil. Fighting games balance protecting one’s own health with taking away the health of the opponent. Strategy games encourage countering an enemy’s plans as well as perfecting one’s own. Shooters combine killing as many enemies as possible while also fulfilling some parallel goal, such as capturing a flag or checkpoint.

Zero-sum mechanics, in fact, seem to be the default choice when designing competitive games. However, their ubiquity masks the many, many problems with this type of gameplay. Indeed, zero-sum mechanics are, at best, a necessary evil and, at worst, a wrongheaded approach to game design that turns away many potential players.

The Zero Problem

The problem with zero-sum mechanics is that they require a negative experience for someone – watching a devastating combo annihilate one’s character in Street Fighter, watching one’s buildings crumble in Age of Empires, dying and respawning over and over again in Team Fortress. One player’s pleasure results from another player’s pain.

In fact, competitive games do not require that another player must suffer. A game’s rules determine the frequency and intensity of player interaction; ultimately, the designer decides how players will interact with each other during play. Indeed, competitive games are even possible without players being able to affect one another at all – consider parallel sports like golf or bowling, for example, or online games with asynchronous leaderboards like Bejewelled Blitz or Burnout Paradise.

The most important distinction is whether a player can lose their current progress or if they can only lose the ability to continue progressing. In the former case, the game mechanics have a zero-sum feel as losing one’s progress is usually a painful experience and often a sure route to a loss. In contrast, one of the defining traits of the Eurogame movement (epitomized by games like Ticket to Ride and Settlers of Catan) is eschewing such direct, zero-sum player conflict in favor of limited, indirect interaction which will not destroy a player’s progress.

For example, in worker placement Eurogames, such as Agricola and Caylus, players take turns choosing exclusive abilities; the competition emerges from players jockeying for position to determine who gets to grab the best jobs first. If a player knows his opponent needs food, choosing the food job for himself can seriously damage this opponent’s fortunes. However, this tactic is qualitatively different from actually destroying an enemy’s farms and killing his villagers in Age of Empires.

In the former case, the setback may only be temporary; in the latter, the player suffers a heavy emotional loss and has little chance of recovery. In fact, a player who spends too much time trying to disrupt his opponents in a game like Agricola can often dig his own hole as each precious action has significant opportunity costs. In contrast, damaging an opponent early in an RTS has little downside; wiping out another player’s economy can actually buy valuable time to grow one’s own much larger.

Balancing a RTS game to not reward destroying another player’s economic base as soon as possible is extremely hard. Indeed, RTS games suffer heavily from a dominance of zero-sum mechanics, which encourage the rush. Many players adopt “no-rushing” house rules to manually rebalance the gameplay away from destructive raids and towards building up for the endgame.

Further, many RTS games end with a whimper instead of a bang because the end goal is usually wiping out the enemy’s forces, which means that the outcome is obvious halfway through the match. In Ticket to Ride, during which players race to complete routes before running out of pieces, the dramatic tension is a consistently rising slope. In contrast, the dramatic tension of StarCraft is an arc which rises and then falls, and – unfortunately – the downward side of this arc is simply a sequence of painful events for the loser.

However, zero-sum mechanics need not be endemic to the RTS genre. Consider economic games, like the Anno series or Railroad Tycoon or even M.U.L.E., in which the primary goal is the acquisition of wealth; because the players are in a race to see who grows the fastest, the games need not encourage – or even allow – players to attack one another.

Alternate competitive mechanics are possible in military RTS games as well. Warcraft 3 introduced the creep – neutral characters who occupy the central area of skirmish maps and who players race to kill for the rewards and experience points. Perhaps a new RTS could take this mechanic a step further and make the game focus solely on killing creeps?

Removing the Negatives

Many competitive games solve the zero-sum problem by severely limiting interaction, so that players can only affect each other under certain circumstances. In Mario Kart, for example, racers can only shoot one another after picking up limited-use shells from certain locations; even then, players will only get the most powerful shells if they are trailing in the race. Even in a cutthroat RTS, a player can only attack after first building a barracks, then training troops, and finally moving them into position.

Thus, limiting player interaction is a powerful tool for minimizing negative emotions from zero-sum play. Games with similar themes and rules can dramatically change their feel depending on what sort of interaction is allowed. For example, Travian and Empires & Allies are similar asynchronous strategy games played over months of real-time about developing a military and then attacking one’s enemies. However, an important difference separates these two games with what happens when players invade each other’s cities.

In Travian, attacks are strictly zero-sum; resources captured by the attacker are taken from the defender’s stockpile. In Empires & Allies, however, combat is actually positive-sum; the resources captured by the attacker are conjured from nothing. Furthermore, while units which die in Travian are removed from the game, defending units in Empires always stay alive, even after a defeat.

Empires quietly belies players’ expectations for combat – that a victory requires a defeat – and this design choices pays off by making the game more accessible and less emotionally draining. In contrast, Travian uses the traditional approach that one player’s gain requires another player’s loss; accordingly, this design choice creates a nasty world full of brutish players with short tempers.

Many designers instinctively assume that conflict must be zero-sum, but this prejudice may be keeping their games from reaching a larger audience. The emotions players experience during a game are real enough, so a mechanic that requires at least some players to suffer should be used carefully.

Adding the Positives

Sometimes, alternate solutions are blindingly simple. In the board game 7 Wonders, players compete along multiple axises – earning victory points for science, civics, buildings, wealth, and military. The default way to implement military in such a game would be to allow players who invest in an army to attack other players’ units, buildings, or resources. 7 Wonders, however, employs a very different approach.

The game is split into three epochs, and at the end of each epoch, players with the largest armies receive positive points while the other players receive negative points. Furthermore, the total point distribution is actually positive-sum, so that losing combat does not hurt a player as much as winning combat helps. Thus, the military strategy does not drown out all the others and is appropriately balanced; a strong military cannot prevent an opponent from winning with strong technology because military victories do not require the loser to forfeit her progress.

Indeed, the spirit of positive-sum gameplay can benefit other aspects of game design. Puzzle Quest, for example, avoids a manual save system by ensuring every combat is positive-sum; players can never lose an item during combat and will always gain at least a little gold and experience from each battle. Thus, a player is always better off after combat, whether a win or a loss, so the game can constantly auto-save into a single slot. This feature, which would be hardcore if paired with a traditional zero-sum design, instead removes the need for a load/save system, which can be a barrier to entry for new players, thereby expanding the game’s potential reach.

Ultimately, zero-sum mechanics are still a powerful tool for game designers as they can unlock primal emotions. Sometimes, allowing players to destroy each other is exactly what a game needs. However, not all conflict need be zero-sum, especially since that design choice has significant disadvantages. Losers need not suffer so that winners can triumph.