Global Araç
Rock Paper Scissors
Siz
?
vs
CPU
?
Galibiyet
0
Mağlubiyet
0
Beraberlik
0
Rock-paper-scissors (called janken in Japan, shi-shi-shi in Korea, ro-sham-bo in some parts of the US) is one of the few games mathematically proven to have no winning strategy when both players play optimally. The game-theoretic Nash equilibrium is to play each option with exactly 1/3 probability — under that strategy your opponent cannot exploit you, and you cannot exploit them; expected outcome is exactly 1/3 wins, 1/3 losses, 1/3 ties for both sides. Any deviation from uniform random play (favoring rock, alternating predictably, mirroring the opponent) creates an exploitable pattern.
This implementation pits you against a CPU that draws each move from a cryptographically- strong uniform random distribution (Math.random or crypto.getRandomValues depending on browser support). That means it plays the Nash-equilibrium strategy perfectly — it has no exploitable bias and no memory of your previous moves. Against this CPU, your long-run win rate converges to exactly 33.33%, regardless of strategy. The variance over short series is the only interesting math: in 10 games you might win 7 by chance (binomial distribution); in 100 games you’ll be near 33; in 10,000 games you’ll be within ±1% of 33.33% with high confidence.
The interesting human-vs-human research: studies of championship-level RPS (yes, there’s a World RPS Society) and studies of casual players show humans are terrible at uniform random play. We underplay rock, overplay paper after a loss, mirror our opponent’s previous move, and show strong gender patterns (men open with rock more often, women with paper). A skilled human or simple Markov-chain model can beat unsophisticated humans at 55-60% over a long series. To beat THIS app you’d need to find a flaw in the random-number generator — which is, by design, not findable.
Nasıl Kullanılır
- Click rock, paper, or scissors — your move.
- CPU instantly plays its move (uniformly random) and shows the result.
- Track your wins/losses/ties across the session.
- Play a long series (50+ rounds) to see your win rate converge toward 33%.
- Reset to start a new tally any time.
Ne Zaman Kullanılır
- Quick decision — flip-coin substitute when more than two options are needed.
- Demonstrating game theory to someone (Nash equilibrium, mixed strategies, expected value).
- Killing 30 seconds of waiting time without committing to a longer game.
- Teaching kids about probability, fairness, and that no strategy beats truly random play.
Ne Zaman Kullanılmaz
- High-stakes decisions — the CPU is fair but “random” isn’t the same as “considered.”
- Multi-player tournaments — this is 1v1 vs CPU only.
- Pattern-recognition skill-testing — the CPU has no patterns to read; you can’t practice human-vs-human strategy here.
Yaygın Kullanım Senaryoları
- Educational use — demonstrating the underlying concept
- Onboarding a colleague who needs the same calculation/conversion
- Verifying a number or output before passing it on
- Quick use during a typical workday
Sık Sorulan Sorular
Is the CPU exploitable?
No. It picks uniformly at random using browser-grade randomness with no patterns to read. Even if you play 10,000 games, your win rate will converge to 33.33% with no strategy able to push it higher. The only way to beat this CPU is variance — the natural ups and downs of small samples.
Why is uniform random optimal?
Game theory: rock-paper-scissors has a unique Nash equilibrium where both players play each option with 1/3 probability. Under that strategy, neither player can change their strategy unilaterally to gain advantage. Any deviation creates a counter-strategy. So if you’re facing an optimal opponent, you can’t do better than 1/3 wins on average — and the optimal opponent against you plays uniform random.
Why do I lose more often than win?
Variance. In any short series of an even-money game, you’ll have streaks of losses by chance. Over 10 games, expected wins is 3.33 with standard deviation around 1.5 — so winning anywhere from 1 to 6 games is well within normal range. Run 100+ games and you’ll be very close to 33%.
Are there variants like rock-paper-scissors-lizard-Spock?
Yes — the RPSLS variant from Big Bang Theory (originally invented by Sam Kass and Karen Bryla in 1995) extends the cycle to 5 options, reducing tie probability from 33% to 20%. Same Nash equilibrium principle: optimal play is uniform random across the 5 options. This tool implements the classic 3-option version.
Can humans beat each other consistently at RPS?
Yes, slightly. Skilled players read tells (eye flicks, hand position, breathing patterns) and predict using Markov models of opponent behavior (what they played last, what worked last, gender-specific opening patterns). Top RPS Society champions hover around 55-60% win rates against casual players in long series. Against another expert, they’re back at 50% — predictability is the limiting factor.
Is this game culturally universal?
Variations of RPS exist worldwide and date back to at least the Han dynasty in China (3 to 200 BCE) where it was called shoushiling. Japan’s janken (rock-paper-scissors) is the modern direct ancestor; Western adoption came in the early 20th century. Some cultures use different objects (frog/snail/slug in some Asian variants) but the cyclic-dominance structure is universal.