Rock Paper Scissors Online is a deterministic three-choice game that runs entirely in your browser against a transparent local opponent, and the exact winning route scores exactly 1,000 points in three legal moves. The opponent is not a language model, a remote service, a hidden random generator, or another person: it is a small counting rule that reads only the moves recorded in your current run, displays the source of its next prediction before you choose, and shows the counter it has already queued. That visibility is what turns a familiar hand game into a short prediction puzzle you can study with the screen in front of you. Because every round is reproducible, the same move history always produces the same opponent counter, and the same winning three-move sequence always lands on the same 1,000-point finish.
What "Rock Paper Scissors Online" Actually Runs in Your Browser
Every button press, key press, and counter decision in Rock Paper Scissors Online happens locally in your browser tab. The page loads with no account, no upload, no API call, no model download, and no new dependency, and it does not write to any server. The opponent only sees the moves you have entered into the current run, so two different browsers playing the exact same sequence of moves will see the exact same counters in the exact same order.
That single property — same history in, same counter out — is what makes the game inspectable. You can pause before clicking and still know which move the opponent has already locked in for the round, because the queued counter is rendered on screen before your legal choice is recorded. After your choice lands, the page updates the prediction source, the new history, the wins and losses column, and the queued counter for the next round, all in the same view.
This is also why the word "learning" on the screen has a narrow meaning. Nothing is trained, persisted, or inferred about you. There is no profile, no personalization across sessions, and no claim that the opponent is optimal or unbeatable in any strategic sense. Restart simply clears the in-memory history and returns the opponent to its fixed opening prediction, which is Rock with Paper queued to beat it.
How the Prediction Rule Reads Your History
The opponent's next move is decided by a small priority list, not by sentiment or style. Before your next legal choice is recorded, the screen already shows three things: the source the rule will use, the player move it is predicting, and the opponent move it has queued as the counter. The source line tells you whether the rule is using a transition sample or a full-history frequency, which is useful when you want to know exactly why a particular counter was chosen.
The rule works in this order. If your latest recorded move has appeared at least once earlier in the run, and at least one of those earlier appearances was followed by another legal move, the opponent looks at those following moves and picks the most frequent one. If no usable transition sample exists for your latest move, the opponent falls back to the most frequent move in your complete history. When counts tie, the rule resolves them in the disclosed order Rock, then Paper, then Scissors.
| Condition before your next choice | What the rule predicts |
|---|---|
| Your latest move has appeared earlier with at least one following move | The most frequent move that historically followed it |
| No usable transition sample exists for your latest move | The most frequent move in your complete history |
| Two or more counts tie under either rule | Rock first, then Paper, then Scissors |
Once the prediction is fixed, the opponent queues the move that beats it. A legal repeated choice — pressing Rock twice in a row, for example — is treated as a new learning sample rather than an ignored duplicate, and ties are recorded for the same reason: a tie can shift the next frequency or transition lookup, even though a tie changes neither wins nor losses. Invalid keys are atomic no-ops, so pressing an unrelated letter does not pollute history or change the queued counter.
How to Play Rock Paper Scissors Online and Finish on 1,000
Rock Paper Scissors Online takes a few clicks to learn and a single repeatable sequence to finish. The whole run plays out on one screen, so the steps below are also the rules of the game.
- Read the visible prediction source, predicted player move, and queued counter before your next choice is recorded. These values are fixed for the round.
- Choose Rock, Paper, or Scissors using the three large buttons, or press 1 or R for Rock, 2 or P for Paper, and 3 or S for Scissors. Ties still update local history.
- Play the exact sequence Scissors, then Paper, then Rock. The opponent plays Paper, Rock, and Scissors in response, so all three rounds are wins.
- Watch the score update as 300 plus 300 plus 300 plus the 100-point completion bonus, landing on exactly 1,000.
- Press Restart any time to clear the current run, return to the opening Rock prediction, and try a different history from the same screen.
The same three-move route always produces the same final score because the opponent is deterministic. Two losses at any point also close the run immediately, and a 1,000-point finish freezes later choices until you Restart.
Score Formula, Ties, and Run Limits
The scoring in Rock Paper Scissors Online is fully disclosed on screen and follows a fixed arithmetic. Each win awards 300 points and each loss subtracts 100, with a hard floor at zero so the score never goes negative. Reaching three wins adds a one-time 100-point completion bonus, and two losses at any point in the run close it immediately. After a completion or a defeat, later choices are frozen until you Restart.
The exact arithmetic for the perfect route is simple and worth writing out. Three wins at 300 points each give 900 points, and the 100-point completion bonus adds another 100, for a final score of 300 + 300 + 300 + 100 = 1,000. There is no partial credit for an interrupted run, no streak multiplier, and no hidden adjustment for difficulty.
| Event | Score change | What it does to the run |
|---|---|---|
| Win | +300 | Counted in the wins column; recorded in history |
| Loss | −100, with a floor of 0 | Counted in the losses column; two losses close the run |
| Tie | No change | Still recorded in local history |
| Third win reached | +100 completion bonus | Added once; freezes later choices |
Ties are recorded in history but never change the score. This matters because the rule that decides the next counter reads from history, so a tie can quietly influence a later prediction without ever appearing in the wins, losses, or score column. Repeated legal choices are recorded the same way: a second Rock is not collapsed into the first one, it is a new sample that the prediction rule will see on the next round.
When to Follow the Route and When to Test the History
The 1,000-point finish in Rock Paper Scissors Online is useful as a baseline, but it is not the only thing the page is for. Because the prediction source, the predicted player move, and the queued counter are all visible before each round, you can deliberately play moves that are not on the winning route and watch the source line change in response. That is the easiest way to confirm the rule is doing what the on-screen explanation says.
For example, the disclosed failure route is Rock, then Paper, then Paper, then Paper. Rock loses to the baseline Paper, the next two Papers tie because the rule still falls back to the opening Rock frequency and queues Paper, and on the fourth Paper the rule switches to a learned transition that predicts Paper and queues Scissors for the second loss, which closes the run. The same kind of inspection works at any point in the run, which is why Restart is worth using when you want a clean history to test against a specific theory.
The page also tells you when nothing is at stake. Once the run reaches the third win or the second loss, later choices are frozen, so you can study the final state without accidentally starting a new round. Restart is the only way to return to the fixed opening prediction of Rock with Paper queued, and it clears the in-memory run state only, leaving your browser storage untouched.
Reproducibility and the Word "Learning"
Calling the opponent a pattern learner can sound like marketing, so it is worth pinning down what the page actually does. The word learning here means deterministic counting within the current run: every legal move, including ties and repeated choices, is added to a small in-memory list, and the next counter is picked from that list using the priority rules shown earlier. Nothing is trained across runs, nothing is saved to a server, and nothing about your play is carried into a future session.
This narrow definition is also why the game is reproducible. Given the same sequence of legal moves, in the same order, from a fresh Restart, the opponent will always produce the same prediction source, the same predicted player move, and the same queued counter. If you want to verify a claim about the rule, you can replay the exact same history and watch the screen match your prediction, which is more useful for inspection than it is for studying any psychological skill.
Rock Paper Scissors Online also makes no claims about optimal play or cognitive ability. The score is a fixed arithmetic result of three wins and a completion bonus; it is not an assessment of pattern recognition, working memory, or reaction time. Treat the route as a deterministic puzzle you can solve in three moves, and treat the rest of the page as a transparent rule book you can read with the game open.
Related reading: Symmetry Game: Read the Mirror Rule to Score 1,000.
Related reading: Interactive Synonym Quiz: Practice Verified Word Match.