Rock Paper Scissors - A Method for Competitive Game Play Design
January 23, 2007 Page 2 of 5
Following the advancement of strategy of two RPS players eventually leads us to a well-defined strategy. One player may decide to perform one attack over and over. The defender will recognize the pattern and react to it. The attacker will then realize that he must change his attacks.
The defender will realize that the attacker is no longer doing the same attack every time, and must also react by attempting to predict what the attacker will do. To avoid being predictable, the best thing that the attacker can do is to attack as randomly as possible. If the attacker is able to attack in a truly random fashion, this eliminates the defender's ability to predict and react to the attacks. This is the "dead end" of a simple RPS implementation of a game.
Since this article specifically discusses games which implement fast-paced game play, speed becomes an issue. Attacking quickly is more effective than attacking slower because it makes prediction on the defender's part even more difficult.
Therefore we have the following conclusion: The end strategy of a simple Rock Paper Scissors game is to be random and fast.
Statistically, each attack will tend to occur just as frequently as another, given that each is equally effective. However, since speed is a consideration the faster attacks are used more frequently than slower attacks.
This strategy manifests itself frequently resulting in an undesirable experience for players. Fighting games which are built with complex mechanics such as Tekken or Soul Calibur require a player to invest time to learn a character's moves. Although the most effective way to play these games in the long run is to memorize how to perform moves and when they are effective, the immediate desire to win prevents many players from becoming skilled at the game.
Instead, beginning players button mash as this seems to be more effective against other beginners. The strategy, although not formally recognized by the player, is the same - be random and fast. The result of the match generally depends on luck and the players do not have much control over what is happening. Since the players are not learning anything, they are not experimenting or seeing what the game has to offer. Generally a button mash-fest results in bored players relatively quickly, so it's pretty obvious why this is an end strategy which game designers should avoid.
Namco's Soul Calibur II
Although RTS games do not fit the criteria of being fast paced, there is an artificial reaction time imposed by the game since buildings and units take time to build. Resource scarcity further increases the amount of time required to build desired units. If a defender in an RTS does not know which kind of attack his opponent is planning his only option is to diversify his troops.
As mentioned before, units in games like StarCraft are effective against certain other units, and to defend against all possible kinds of attacks, a defender should build a variety of troops. Similarly, an attacker must consider the defense that the defender is building. Without any knowledge about a defense the attacker's best strategy is to build a variety of units as well.
The result is that both the attacker and defender will build a variety of troops. Since time is critical, they will be following the same strategy - be random (or in this case, diverse) and fast. Therefore, many RTS games are reduced to games of memorization and methodical repetition.
mind that the RPS model must be present to reach this end strategy. If
all attacks are not equally effective or if there is one dominant
strategy then players will exploit that and the game will not not even
reach the point of being an RPS game.
In the end, the game is no longer about decision making, but rather performing a series of steps as efficiently as possible. This is another situation that a game designer should attempt to avoid, as this tends to not be as fulfilling of an experience for players as one where success is achieved through decision making and improvisation.
Signals - Overcoming RPS Shortcomings
The strategy of being random and fast is problematic because it does not create an enjoyable experience for players. RPS games become very mechanical. In fact, being random and fast is a very simple strategy to formalize and even implement in AI. Fortunately, signals are a relatively simple solution to this problem.
A signal is an action or behavior which indicates that another action or behavior is going to follow. The word "introduce" is not entirely accurate because whether there is effort on the game developer's part or not, there is almost always a signal present to indicate an action.
For example, if an attacker intends to shoot his opponent in Halo, he must face and aim at the defender. This is a signal to the defender that he will likely get shot at, although the animation for aiming is included to make the game more realistic and not necessarily to enrich game play.
In the boxing game Fight Night 3, the haymaker does not execute immediately when the player pushes the button. Instead, the boxer plays an animation before striking. The pause between the execution of the haymaker and the actual punch connecting is a signal. The defender can react to the attack if he is fast enough by dodging, blocking, or counter attacking with a faster punch.
Electronic Art's Fight Night Round 3
Signals are almost always present in one form or another. What is important is not that the signal is there but that it is obvious enough to be noticed, and that there is enough time between the signal and the attack that the defender may react.
The timing of signals is critical to the way the game is played. If the amount of time separating the signal and the attack is too long then the defender will be able to react successfully every time. If the signal is too subtle or occurs too close to the actual attack then the defender will not be given a chance to react.
The soccer game Super Mario Strikers used signals to give the defender a chance to react to the “Super Strike” move. The captain of each team can perform a powerful shot at the goal which cannot be stopped if performed correctly. To perform the shot the attacker must press the B button at the right times as indicated by a green and yellow ticker. During this time, the defender can react to the signal by tackling or using an item on the attacker. The length of the signal is short enough to make this an effective attack, but long enough to give an attentive defender the chance to stop the attack.
Signals introduce a level of skill to the game that is not present in simple RPS games. Now rather than guessing or acting randomly, the defender has the opportunity to successfully react to attacks. Taking this one step further, signals allows a defender to recognize patterns. Skilled Fight Night 3 players are able to identify punches by their animations and react accordingly with blocks, parrys, or dodges. In general, if a defender knows a good defense or counter attack then he will be able to defend against attacks by learning their signal/attack relationship.
This learning is especially valuable because it gives the player a sense of accomplishment and control. Rather than relying on luck he can now be attentive to what is happening and through his own effort become more successful at the game. This built-in reward system is very important in keeping players interested in a game, which is obviously desirable as a game designer.
The presence of obvious, well-timed signals can greatly enrich the play experience. Keep in mind that for a signal to be effective in enriching the play experience, the player must be aware of the RPS relationship between the attacks. There are three criteria which must be present at this point:
Each attack must have a known and effective defense or counter attack.
Each attack must present a signal that is detectable.
Each attack must occur after the signal at a time interval which allows the defender to react.
If these three are true, then given a certain number of repetitions, the defender should have a realization. He may think, "Oh, now I see! Whenever his character does this action, it means he is going to attack like this. Next time, I should try this counter attack."
As mentioned before, this realization will keep players playing the game. The player is now thinking about "next time" rather than about quitting because he doesn't stand a chance. Again, the focus is not on the presence of the signal, but rather that the signal is obvious enough to be noticed by the player and timed such that a defender can react to it.
Page 2 of 5