The foundation of my multi-player experience came from vanilla deathmatch titles like Quake III Arena, or Unreal Tournament . In such titles success or failure appears entirely dependant on player skill. However on closer inspection there is a subtle form of positive feedback in action. When you spawn, or respawn, into the world the starting weapons you have are not as powerful as the weapons you can find in the environment. Therefore your primary instinct upon entering the world is to locate these powerful weapons. In an encounter with an opposing player, you may die and therefore respawn with the default weapons, or you may win and continue with the weapons you have, likely more powerful ones than you started with. Over time this leads to a situation where the players who are on average more success in combat are at any given time likely to be in possession of the more powerful weapons and thus be at an advantage. As players who are consistently dying are more likely to have only the default weapons. Success leads to an advantage and therefore an increased likelihood of future success.
On the other hand when players respawn they are returned to full health (In Quake III Arena itself, they gain a boosted health level that decays over time), whereas the players that survive comabt will probably have had their health reduced. Though this reduced health can be recovered through health pick-ups until those are located the surviving player is at a potential disadvantage compared to the respawned player. This serves as a slight negative feedback loop leveling out the playing field between successful and unsuccessful players. However, besides suicide, player death has no direct impact on overall score so the number of times a player dies is of little import beyond bragging rights. In a deathmatch game, the negative feedback loop has no direct influence on the score. The more successful you are the better your chances of being successful in future. Of course it’s not as straightforward as that, as second order factors such as the layout of the level, and player psychology, come into play. Still the core positive feedback loop remains a key influence on the play experience. Consider how different a match of InstaGib Unreal Tournament feels to the standard game.
Recently I’ve been spending a lot of time playing Team Fortress 2. It’s only after several hours of play, and listening to the Commentaries, than I’ve started to realise the extent to which feedback loops are used throughout the game. Some are explicit and have clearly been designed by Valve, while others are implicit and could have been either designed or emmerged from the other mechanics within the game. Together they have subtle and wide ranging affect on the play experience.
When you die in vanilla Team Fortress 2, you have to wait a pre-set time before you are able to respawn, the length of time is altered depending on whether your team is winning or losing. This is an explicit positive feedback loop that means that if the winning team is able to kill a number of enemies they will out number the opposition for a time and have a higher chance of achieving their objectives. Initial success leads to continuing success.
Another positive feedback mechanism, which is more implicit and might not have been directly designed, exists for all classes but is most noticeable for, the Heavy, the Pyro and the Engineer. The first two classes are offensive and have primary weapons that burn through ammunition (In the case of the Pyro quite literally), so playing as either of these classes it’s possible to run out of ammunition rapidly. Though there are several mechanisms to replenish that ammunition the most interesting is that of gathering the dropped weapons of dead players. If a player waste their fire and doesn’t manage to kill anybody they will diminish their ammunition supply and be weakened, however if they are successful they will be able to gather the ammunition dropped by their defeated foes. Though this is true for each class the rate of fire of the Heavy and Pyro make the affect that much more pronounced. The Engineer is in a similar position, as they can salvage metal from the dropped weapons and use that to repair, or upgrade their turrets making them more powerful and therefore more likely to kill more opposing players, providing more metal and leading to further repairs and upgrades. Accurate shooting pays for itself; again initial success leads to continuing success.
Boink!
A stable game rarely makes for a fun game as anybody who’s been stuck in a stalemate on CTF-2Fort will attest to. These feedback loops are not foolproof but generally they prevent rounds from degenerating into a war of attrition. The effects of the feedback loops described, and the various others present in Team Fortress 2, are subtle, they combine to slightly destabilise each match. When one team starts to pull ahead their success is rewarded and reinforced leading to an increased likelihood of future success; the losing team has to work even harder just to survive. This leads to a game that is unlikely to result in a stalemate. A game that is much more dramatic, that reinforces the Team Fortress 2 notion that “a draw doesn’t mean everybody wins, but that everybody loses” and that leads to valiant charges and last ditch defenses.
Without these feedback loops in place, the game would be more prone to stalemates especially between teams of similar average skill levels.
It natural to want to feel that skill is sole determinant for success in a multi-player title, and obvious feedback loops systems can be off putting. At the same time games of pure skill are often not as dramatically entertaining as those where the balance is inherently unstable.
