A single throw of Rock Paper Scissors between two strangers is essentially luck: if either player is truly random, neither can gain an edge. Over many throws it becomes a skill game, because people are not random, and a player who spots an opponent's habits can win more than a third of the time.
Why one throw is fair
Game theory has a clean answer. If you pick rock, paper and scissors with equal probability and no pattern, no strategy can beat you on average. That mixed strategy is the game's equilibrium. The catch is that it also cannot beat anyone; it only guarantees you break even over time. More on this on our game theory page.
So in a single decisive throw, like settling who pays for lunch, the game is about as fair as a coin toss.
Why many throws reward skill
Humans are poor at producing random sequences. We avoid repeating the same choice too often, we react to wins and losses, and we favour some throws. Each habit is something an attentive opponent can exploit.
The Zhejiang University study
In 2014, Wang, Xu and Zhou published a lab experiment with 360 students at Zhejiang University, split into groups of six who played 300 rounds each. They found that players' choices depended on the previous result. Winners tended to repeat their winning throw more often than chance would predict, and losers tended to switch to a different throw. The pattern is known as conditional response, or loosely win-stay, lose-shift.
What our own data shows
The State of RPS report publishes the dated WRPSA sample, throw frequencies and after-result transitions. Most play in that sample was against the computer, so it should not be treated as a universal account of human-versus-human play.
The report's players tended to switch after winning, whereas the Zhejiang study found a win-stay tendency in its laboratory setting. That difference matters: a pattern observed in one setting is a hypothesis to test against your opponent's completed history, not a guaranteed next throw.
Evidence from bot competitions
In 1999, Darse Billings at the University of Alberta organised the First International RoShamBo Programming Competition, where computer programs played long matches against each other. The winner, Iocaine Powder by Dan Egnor, did not try to be random. It combined history matching (finding moments in the past that matched recent moves and predicting what followed), frequency analysis, and random play as a fallback, and chose whichever approach was working best. The lesson: when opponents have patterns, prediction beats randomness.
Speed is not skill
There is one way to win every round, and it is not allowed. A janken robot built at the University of Tokyo reads the shape of your hand with a high-speed camera and forms the winning shape within about a millisecond, so to a person watching, both hands seem to land together. That is reaction, not prediction, and it is exactly what a late throw is. The count exists so both players commit at the same moment, which is why throwing late is a foul in the official rules. The robot is covered in Rock Paper Scissors robots and AI.
So which is it?
| Situation | Mostly luck or skill? |
|---|---|
| One throw, strangers | Luck |
| Best of three | Mostly luck, some skill |
| Long matches, same opponent | Skill matters |
| Against a truly random player | Luck; nobody can win on average |
If you want to test yourself, play against the computer and watch whether you fall into the habits above. Our strategy guide explains how to use them without becoming predictable yourself.
FAQ
Can you get good at Rock Paper Scissors?
Yes, over many games against people. You improve by noticing opponents' habits and hiding your own.
Is Rock Paper Scissors random?
The rules are fair, but human players are not random, so outcomes over many throws are not purely chance.
Is Rock Paper Scissors an Olympic sport?
No. It has never been part of the Olympic programme.