Applying Rock Paper Scissors to Video Games

If you were ever a child (which I hope you were) or heck, even if you're an adult (which you probably are) then you've likely heard of Rock, Paper, Scissors (RPS) - or Roshambo, as some call it.

Some use it for evil in order to determine who pays for lunch, while others prefer to use it for good to pass the time with friends. Some video games have even started adopting it in order to determine the outsome of events.

But the problem is that the game doesn't always work when applied to video games.

Full disclaimer: I'm no mathmetician. Probability and statistics have never necessarily been my forte, so take whatever I say with a grain of salt and that in mind.
 

In the Yu-Gi-Oh Trading Card Game, which player goes first can be very important.

The first player cannot make attacks on their opening turn and is commonly left to play defensively to prepare for their opponent's turn (during which which players may start attacking freely). Depending on your deck and starting hand however, opening the game may be your optimal strategy.

The official rulebook for the Yu-Gi-Oh Trading Card Game states that in order to determine which player goes first, players must "play rock-paper-scissors or flip a coin", the winner choosing which order play begins in.

Yu-Gi-Oh: Legacy of the Duelist attempts to emulate this mechanic by having the players play RPS before the game, which works well enough for online play - but what about the standard exhibition duels with AI opponents?

Well...the application is less effective.

Initially I played the game quite a bit and had very few problems with the way turns were determined - after all, turns were being decided and duels were being had. Who was I to complain?

In fact, it was my roommate who pointed out the glaring flaw to me: we had never seen a tie.

I doubt the game was rigged (since the Legacy of the Duelist's AI seems to just choose a winning or losing 'throw' at random) and obviously the last thing anyone would want is an AI picking the same option as the player over and over again. Regardless, I felt cheated that the game was reading my movements, as well-intended as it might have been.

As I was amazed to discover while doing research, a lot more goes into the psychology and strategy of RPS then most might realize. So when an AI opponent chooses a throw arbitrarily from a winning or losing option it essentially boils the outcome down to a coin-toss and nullifies what differentiates RPS from a coin flip as a means to determine odds.

RPS mechanics work better for Legacy of the Duelist's online play where, players are more likely to stick to certain patterns or throws. Yet only one round of RPS is played to determine the outcome, so it always goes to the player who wins the game first. If there are no ties and one single round, then there's no opportunity for players to recognize patterns and act on them.

Furthermore, by playing RPS against a real player online it does allow for verbal tactics to develop for players, but also removes many of the physical nuances and tells that help players develop a strategy against too.

In the mobile edition of Yu-Gi-Oh, Yu-Gi-Oh: Duel Links, the starting turn is simply determined by a coin toss.

Now let's talk Yakuza 0.

Yakuza 0 is a strange game with lots of strange minigames to be had, ranging from watching erotic videos to singing karaoke. One of its more eccentric minigames, the Catfight Arena, allows the player to place bets on scantily clad women who wrestle one another into submission.

Each fighter has different stats and abilities which makes them better or worse in certain conditions. The player doesn't take direct control of the women as they fight. Instead they play RPS. Each potential throw they may choose has different levels of effectiveness depending on which fighter they choose. After the throw choices are revealed, the player can mash a button to help cheer on the girl and help her win the grapple.

Unfortunately how the game determines the outcome is unclear. Some players believe the game reads their moves and that scrolling through the list of throws is the most effective means to confuse the AI, others theorize that the outcome is entirely random and it is a complete game of chance.

Obviously, this could very well mean that the RPS game is working fine. The AI could be making choices without bias and simply choosing options. Shouldn't this be the ideal execution of a RPS game against AI?

Well...no.

Just like Yu-Gi-Oh, the player can't make an informed choice because they have literally no information to work with. A few round would reveal that the AI doesn't appear to choose an option based on it's current fighter's abilties, nor does it make choices dependant on patterns or the actions of the player.

So with nothing to help them decide a course of action, players are left to make decisions based on guesswork alone and effectively render the minigame a coin toss.

Furthermore, the outcome may often feel predetermined to the player. Some fighters have the tendency to push back arbitrarily during button mashing while some characters have the ability to win automatically when their health reaches zero, effecitvely rendering the player's victory a failure. Isn't that a nice surprise?

Divinity: Original Sin has a mechanic called 'dual dialogues' which addresses many of the issues Yu-Gi-Oh: Legacy of the Duelist and Yakuza 0 do not.

Dual Dialogue appears between players if they choose opposing stances on the narrative choice at hand. In this example, one player has chosen to let a man steal a fish while the other chose against it.

All characters have three personality traits (Charm, Intimidation and Reason) which they can use in these types of conversations. After the opposing view is established between players, they each choose a trait to use in the RPS game. Depending on the context and player's charisma + personality trait, the amount of points they score upon a successful win of a turn will be different.

After entering the RPS game, players simply choose one of three face buttons to choose an option - the first player to ten points wins the argument and their dialogue choice acts as the response to the narrative choice.

This use of RPS is more effective for a number of reasons:

1) Real Opponents: While this mechanic can be used against NPCs (wherein AI throws are determined arbitrarily much like the previous two examples, unfortauntely) the game offers more opportunities to use it against other players. Not only does this allow for pattern recognition but it's the only effective method of keeping the nuances of RPS in tact.

2) Multiple Throws: Since character's charisma traits can't simply be raised to 10, it gurantees one player must at least win two rounds. This forces mutiple throws to be had wehich in turns allows the possibility of recognizing patterns of throws to help players make informed choices on future throws.

3) Meta: This mechanic is incorporated into a game with local co-op and online play, so the potential for metagaming exists. Players have opportunity to feint opponents by announcing fake throws or (if local) look at their opponent as they decide in order to read them, just like the hand game.

So what have we learned?

It's possible to incorporate RPS into a game, this much is true. But unless a human plays against another human an AI may struggle to create an experience which doesn't feel like sheer random outcomes for the player. Especially without multiple throws, the game of rock paper scissors loses much of its meaning.

Again, I'm no mathematician so there may be some probability to all these mechanics I may have overlooked or not picked up on.

Yet as a player first and designer second, there are plenty of times in which I'd rather not see my illusion of choice be broken. Sometimes if you need to determine an outcome in a video game, it can be okay to throw of a virtual coin.