How to Win at Rock Paper Scissors
Rock paper scissors is a three-way choice, but against a human opponent it is not pure luck. People cannot generate randomness, and the habits they fall back on are measurable. Here is what researchers have found and how to use it.
In theory it is perfectly balanced
The game is a standard example in game theory. The three options beat each other in a cycle, so no single move dominates. The optimal strategy is to play each one exactly a third of the time, completely at random.
Play truly at random and your win rate is locked at one in three no matter what your opponent does. You cannot win more, but you cannot lose more either. Game theorists call this the Nash equilibrium.
The catch is that humans cannot produce randomness. Even when you feel like you are choosing without thinking, patterns show up in the sequence. Against a person, theory and practice diverge.
Win-stay, lose-shift
This is the best-documented pattern. People tend to repeat a move that just won and switch after losing. A large 2014 study in China involving 360 participants confirmed the effect statistically.
Exploiting it is straightforward. If your opponent just won, they are more likely to repeat that move, so play whatever beats it. If they won with rock, play paper next.
If they just lost, they are likely to avoid the move they played. That raises the odds of the other two, so pick whichever of those is better for you.
Rock is over-represented on the first throw
Surveys repeatedly find rock played more than a third of the time on the opening throw, and the effect is strongest among men and newer players. The clenched fist is the most natural gesture and reads as aggressive.
That makes paper a slightly favourable opening. The catch is that this fact is widely known, so an informed opponent may deliberately avoid rock.
Scissors appears least often, because the hand shape is more complicated and less instinctive. So when you are torn between two options, lean towards the one that handles rock.
Nobody plays the same move three times
People feel transparent when they repeat themselves, so after two identical throws they almost always switch. In genuinely random play, three in a row happens one time in twenty-seven and is entirely unremarkable.
You can use that. After an opponent throws the same move twice, the next throw is unlikely to be that move, which narrows it to two possibilities and lifts your odds from a third to close to a half.
To avoid being read yourself, deliberately throw the same move three times. It breaks the model your opponent is building.
Actually playing at random
The only reliable way to be unreadable is to stop deciding with your own head. Prepare a random sequence in advance and follow it, and no amount of analysis will push your opponent above a third.
The tool on this site picks each throw with unbiased randomness. It also works as a practice partner for spotting your own habits, since playing a long run reveals how heavily you lean towards particular moves.
If something actually matters, do not settle it with human rock paper scissors. Whoever reads patterns better has an edge, so it is not the pure coin flip people assume.
Why it breaks down with a group
With three or more players throwing at once, ties become common. If all three moves appear, the cycle means nobody wins, and that happens often.
The tie rate climbs with group size, and past five players most rounds resolve nothing. That makes the game a poor choice for selecting one person from a group.
For that job, a name picker or a wheel is far quicker, because neither degrades as the group grows. Rock paper scissors works best with exactly two players.
Try it right here
You can try what this article describes without leaving the page.
You
❔
Computer
❔
Pick your move
Wins
0
Draws
0
Losses
0
Open the full tool
Full screen, saving results, and the rest of the options all live on the tool page.
FAQ
Can you really read an opponent?
Not perfectly, but you can gain a statistical edge. Win rates move from about thirty-three percent to around forty percent, and the difference accumulates over many rounds.
Does the tool see my move before choosing?
No. Its throw is generated independently of your selection, which is why the win rate converges on one in three over time.
I keep losing several in a row. Is that odd?
It is normal. Losing five straight has about a 0.4 percent chance, but three straight is about 3.7 percent, which is common. Short sessions exaggerate streaks.
Does the same apply to variants?
The core structure carries over, but variants where the winner keeps initiative amplify the psychology. The player holding initiative repeats their previous move more often, so patterns become even more pronounced.
What should I use to pick one person from a group?
Not this. Ties make it inefficient. A name picker or a wheel resolves in one step regardless of group size.