Random Headshots and the Action-Reward Model

Possibly the most frustrating games (video or not) are those whose core gameplay mechanic depends on randomization. From the roll of the dice in Monopoly to the lucky headshots in Counter Strike, these games provide the possibility for a player to defeat another player of much greater skill. Often, this initially leads from angry, blood-boiling cries of “OMG NOOB! YOU'RE SO LUCKY!” to questions like “why does this skill-based game allow random noobs to dominate a veteran player? Isn't it a bit of an oxymoron for skill games to have luck?”

In my previous entry I suggested that luck and randomization play a central role in games like Nanaca Crash. The thing is, Nanaca Crash is an extreme case of how luck and randomization can benefit the gameplay. Indeed similar concepts play a role in contemporary games such as Counter Strike, where the main skill is learning to aim and fire at the head. In CS, there are two streams of interactions that allow the player to perform these headshots: 1) learn to control the recoil of your gun (i.e. use skill) or 2) spray and pray (i.e. use luck). These two separate streams also separates the novice players from the veterans. The veteran will usually perform headshots via stream #1 whereas the novice will most likely follow stream #2. What's interesting is how two entirely different streams of interactions allow players of any skill level perform the most deadly action in the game.

For the novices, every random headshot gives them a glimpse of what it's like to be a veteran player. That is what keeps the player coming back for more. It's like being offered free broadband Internet access for three months and then having it taken it away. No doubt you'll feel the urge to shell out the extra coin lest you risk falling back into the dreaded depression that is dial-up. Similarly, with CS, after being offered that one shining moment of making the perfect headshot, you think to yourself, “maybe I can do it again next round!” Then next round turns into next next round, and suddenly you've been playing for hours.

Thus randomization and luck is a useful tool for games that do not provide an explicit action-reward gameplay model. What I mean by an “action-reward gameplay model” is the gameplay model whereby explicit rewards or upgrades are given to you as you complete more of the game. A good example would be Gran Tourismo where winning races gives you money which you can use to upgrade your car. The action-reward relationship is explicit.

It can be argued that the action-reward model in CS is explicit in that players are rewarded cash for making kills, but skilled players will often max out their arsenal early in the game. This renders the weapon purchasing system irrelevant and suggests that the action-reward model in CS is not explicit. Instead, the draw of the game for skilled players is being able to dominate the enemy with repeated headshots. Consequently the relationship between the action and the reward must be related to making headshots. Randomization is one way to reward novice players with these headshots (the “glimpse of what it's like” via luck) while still rewarding veteran players in greater frequency. Thus, randomization supports an action-reward model that is implicit.

Here are some more examples:

• Any Sports Game: Scoring a random goal


• Street Fighter 2: Pulling off random spinning pile drivers with Zangief


• Texas Hold'em: Flopping a set with seven-deuce offsuit

-j