How to Win at Rock Paper Scissors (Real Strategy, Not Luck)

Short answer: You can beat humans at Rock Paper Scissors because humans cannot be random. Key patterns: beginners open with rock, players rarely throw the same move three times in a row, winners tend to repeat and losers tend to switch. Play the move that beats what your opponent's last result predicts. Try it on Rock Paper Scissors.

Why it is not a coin flip

Against a truly random opponent, nothing helps — you win a third, lose a third, draw a third. Against a computer using a proper random generator, no strategy exists.

But humans are not random generators. We are demonstrably bad at producing random sequences: we avoid repetition, over-alternate, and follow predictable emotional patterns after winning or losing. Every one of those is exploitable.

Pattern 1: Rock opens

Inexperienced players throw rock first far more often than a third of the time. It is the default — the fist is the most natural shape and the word comes first.

Play paper on the first throw against someone who has not thought about the game. This is the single most reliable edge available.

Against an experienced player, expect them to know this and open with scissors to beat your paper. At which point you play rock. This regress ends where you judge their sophistication to end.

Pattern 2: Win-stay, lose-shift

The most useful behavioural pattern, and it is well documented in studies of actual tournament play.

• After winning, players tend to repeat the move that just won.
• After losing, players tend to switch — usually to the move that would have beaten what they just lost to.

So:

• If they just won with rock, expect rock again. Play paper.
• If they just lost with rock (you played paper), they will often switch to scissors — the move that beats your paper. So play rock.

That second one is the most profitable read in the game and almost nobody defends against it.

Pattern 3: Three in a row almost never happens

Players very rarely throw the same move three consecutive times. It feels obviously non-random, so they avoid it — which makes the avoidance itself predictable.

If someone has thrown rock twice, they are unlikely to throw it again. That leaves paper or scissors. Play paper — it beats rock if they surprise you, draws against paper, and only loses to scissors. It is the best expected value against the remaining options.

Pattern 4: Announce and follow through

A genuine tournament tactic. Tell your opponent exactly what you are about to throw — then throw it.

Most people assume you are lying and play to beat something else. It works once, sometimes twice, and it is far more effective than it sounds.

Pattern 5: Scissors is underthrown

Across large samples, scissors comes up slightly less than a third of the time — roughly 29–30%. Rock and paper are marginally over-represented.

This is a small edge, but over a long series, favouring rock (which beats scissors and draws with rock) is very slightly positive against an unknown opponent.

How to be unpredictable yourself

Harder than reading others, because you cannot simply decide to be random.

• Pre-plan a sequence. Decide your next five throws before you start and follow them regardless of results. This defeats win-stay, lose-shift entirely, because your behaviour stops depending on outcomes.
• Use an external source. Digits of a phone number, the second hand on a watch — map values to moves. Genuinely random and impossible to read.
• Deliberately break your own rules. Throw the same move three times occasionally, precisely because your opponent assumes you will not.
• Do not react visibly. Frustration after losing is the tell that makes lose-shift readable.

Reading physical tells

In person, some players telegraph. Watch the hand during the countdown — a hand starting to relax often becomes paper, fingers beginning to separate often become scissors, a tight fist tends to stay rock.

This is unreliable and varies by person, but against the same opponent over many rounds it is worth watching for.

Does any of this work online?

Against a computer with proper randomisation, no — and that is worth stating plainly. Nothing beats true randomness.

Against another human, all of it applies. And playing against a computer is still useful practice for one specific thing: noticing your own patterns. Play thirty rounds and write down your throws. Most people are startled by how non-random their own sequence looks — long stretches avoiding one move, alternation that is far too regular, clear reactions to losses.

Fixing your own predictability is worth more than reading anyone else's.

Try it on Rock Paper Scissors, then move to Higher or Lower for a different kind of prediction game.

Related reading: Tic Tac Toe strategy and games to play with friends.