Survivor Prediction Game Rules

Play the historical challenge

Historical Challenge is the default mode. The first pack covers all 13 competitive episodes of Survivor: Borneo (Season 1, 2000). The reunion is not a prediction round. These are predictions made now against recorded outcomes, not entries made before the original broadcasts. Play at your own pace; historical rounds have no weekly broadcast cutoff.

New historical runs are prediction duels against Scout, a fictional rules opponent. His calls are committed when each checkpoint opens and use only earlier votes and challenge records, never your picks or the hidden results. Each reveal compares both calls and their duel points. More points wins that council. Win more councils than Scout across all 13 episodes to win the season. Equal council wins is a season tie; total duel points do not break the tie.

You have three Double Reads for the entire run. Before locking, optionally select one question: a correct call earns twice its normal duel points; an incorrect call earns zero and still spends the Read. You can keep all Reads for later. Each episode can use at most one, consumed only when the lock is accepted; changing a draft spends nothing, and retrying an accepted lock never spends again. Double Reads are your advantage over Scout, who earns ordinary points. They do not change original weighted accuracy or historical outcomes. Older accuracy-only saves retain their original rules; New challenge starts the new duel.

Choose real names or stable player aliases before starting. Season and tribe names stay visible. The checkpoint includes the cast still in play, earlier challenge results and previous recorded ballots. The game table shows only opened checkpoints and explicitly revealed results. The first episode has no earlier results, so the opening predictions have limited evidence. Knowing the season can help; this is a personal challenge with no competitive leaderboard.

Answer two or three questions per episode. Departure calls weigh two points and challenge calls one; the finale jury winner weighs two. Combined reward/immunity is scored once in episode 1. Episode 7 has no reward prediction. Episode 11’s reward scope is omitted because the archive includes differently scoped reward records. In the finale, select both castaways voted out across its councils, predict the last immunity winner and choose the jury winner. The runner-up is not part of the voted-out set, and the finale jury call is not an earlier season-winner pool.

Tap a castaway or tribe to place a prediction; single choices automatically open the next empty card. For the finale departure set, select the two voted-out castaways and use Place departure set; the runner-up is not included. The set shows its selected count against two, with missing or extra picks explained. The jury-winner card names your current departure set and lets you revisit it. Before sealing, a winner also in that set is flagged alongside any immunity overlap; you may edit either card or keep your guesses. Tap any hand card to change it. Study departure opens earlier ballots; Study immunity or Study reward puts that challenge first. A complete hand offers Study camp to continue at your last camp stop. Back to my hand returns to the same card without changing your picks. Earlier records brings prior votes received and individual immunity wins beside the cast, or recent challenge records beside the tribe choices. The first camp has no past evidence. Shields reflect your own immunity prediction; if it conflicts with a departure call, the hand offers an advisory cross-check without blocking submission. Review the visible hand before Seal my predictions. One accepted lock saves the whole set and ends editing. Reveal each actual result individually; future calls remain hidden. The final card of every council stays visible above its outcome; earlier cards are available in the recap. Live feedback tracks leads and comebacks using revealed points. Next-card stakes show what each correct call could earn. Island journey records completed wins, losses and ties; expand it to revisit an opened episode. The merge and finale introduce their changed prediction formats when you reach those checkpoints. After the whole episode is revealed, the next checkpoint opens with the newly known records. Correct exact sets receive their full weight; incomplete or extra selections earn zero. Weighted accuracy uses total earned weight divided by possible weight across fully revealed episodes. Correct calls and best hit streak are personal statistics, without extra points.

In departure comparisons, expand Who wrote this name? to follow a recorded voter with Study voter. Only legal current castaways can be studied; Not in camp and Other read explain unavailable names. Studying changes only that comparison card, keeps your hand and other read, and spends no Read. You can follow another voter and use Back to return one step to the original list and button. Cancelling Change read keeps the path; manually changing or browsing a comparison card ends it. Only Pick or Keep places a single prediction. Past ballots do not establish alliances or predict future results. The temporary investigation path is not restored after refresh.

Historical and fictional runs are saved locally in this browser until replaced or site data is cleared. My local saves can export both runs to a file and preview an import before replacing them. There is no cloud sync or automatic transfer of older server saves. When browser saving is available, unsealed historical picks, the current question and your Double Read choice return after a refresh or game-mode switch. A restored hand remains editable; only sealing submits it and spends a selected Read. Accepted predictions and partial reveals remain authoritative. Starting a new historical challenge replaces that historical run and clears its draft; fictional practice remains separate. Source links appear after an episode is fully revealed and may contain later-season spoilers and real names even in alias mode.

The reviewed historical snapshot and original fictional rules are bundled for client play. Normal play reveals one result at a time, but browser developer tools can inspect bundled data; this is a personal challenge, not a protected competition. Saves record legal actions and rebuild scores on import. Unsupported data versions or invalid files do not replace current runs. Live picks are not currently offered.

Play Camp Duel

Camp Duel is an optional fictional practice game. Outread a clearly labelled fictional opponent across five original fictional episodes. Rook studies challenge and reward reports; Sage studies the camp and idols. Their calls are fixed when the round opens, before your choices. They never read your predictions or the hidden episode record.

Use two regular scouts per round. Selecting a report category shows its purpose for free. Open report spends one scout and permanently opens that report; rereading is free, and reports cannot be exchanged. You also have two deep reads for the entire campaign: after regular scouts are spent, the Open report button explicitly costs one deep read, at most once per round. Spend it to settle a difficult call now, or save it for later.

Choose four predictions, one card at a time. Exact-set cards allow multiple selections; tap a selected card again to remove it. Check the four-answer review and use Edit to revisit any prediction before Lock in my calls. Locking saves all four and ends editing for the fictional round. Unsubmitted choices are drafts and are not saved by a page refresh. Each result flip reveals only that call and both submitted predictions. Reloading restores accepted calls, opened reports, deep reads and revealed results.

The ordinary published question weights apply equally to both sides. The greater earned resolved weight wins the match; equal weight is a tie, and Void scores for neither side. Win more rounds than the rival to win the campaign. Prediction accuracy stays separate from match wins and independent winner calls. Your choices affect your information and duel performance, not the fictional camp outcome.

After each completed round, including a loss, choose one camp or challenge lead. That report opens free in the next round alongside two fresh scouts. The choice can change until you enter the next camp; then it is fixed. Mid-season entry gives a three-round comeback and no past match wins. Camp Reads and the fixed Rules walkthrough remain available from the format selector or fresh-duel dialog.

Play Camp Reads

Camp Reads is an optional private practice format. It is an original fictional mini season, with six castaways, five episodes and a final two. New seasons change tribe assignments, alliances, one or two councils, idol protection and the jury winner. A personality is not a fixed fate. These scenarios are not real Survivor events or a replica of an official season format.

Before each prediction round, read the camp briefing and optional cast dossiers. You have two scouts to spend on four report categories: camp whispers, challenge rehearsal, idol watch and reward conditions. Each opened report uses one scout and is saved in this browser. You cannot exchange it for another category. Reports distinguish observations, rumours, authenticated possession and schedule checks. Camp reports may identify a target and a fallback: cross-check immunity or idol protection before calling the departure set.

Investigating is optional, and you can still edit your predictions until the cutoff. Scouting never adds points or changes the episode result. Submit all four picks, finish the fictional round, and unlock a story recap plus feedback on your read. Sealed reports stay sealed after locking. A new round restores two scouts.

Your personal goal is at least 70% overall accuracy across three or more scored rounds. First read means completing one scored round; Clean sweep means one 100% round; Camp Reader means meeting the season goal. Only unlocked, non-Void results count, and hiding results or a correction can change these milestones. They are private achievements, with no extra points or prizes.

Individual immunity and reward questions ask for a specific player from round 3. The final exhibition award gives no protection from the separate jury decision; the last departure question asks for the runner-up, and both winner pools use the jury winner. A reward with a failed event record is Void, while a cancelled reward is Not applicable. Every episode has scored questions. The original Rules walkthrough is also available in the format selector and retains fixed examples of all-Void rounds and a result correction. A season report appears after the finale is unlocked, with a clear replay action. Winner pools remain separate from weekly accuracy.

Enter a published round

Live episode picks are not currently offered in the game. The following published-round rules describe the retained online format, separate from current local historical and fictional play. Each is labelled Live or Fictional. Before opening a question list, confirm that you are ready to see cast information through the previous episode. This permission does not unlock the episode results.

Complete every question and submit before the shared UTC cutoff. The cutoff is five minutes before the verified earliest broadcast used by the round. You can update an open entry; all accepted submission versions remain in your personal history. Locked rounds cannot be reopened or backfilled by joining later.

Published results require a sourced result proposal and publication after the verified episode end. Corrections preserve the original settlement and recalculate every affected entry. Unlock results after watching; hide them again at any time. The personal summary includes unlocked results for the selected season and remains separate from private scores and Winner Pools.

Community picks appear only after locking and at least ten eligible submissions. They use each player's last accepted entry. Exact-set option percentages may sum above 100% because one player can select more than one castaway. These describe user choices, not real-world winning probabilities.

Published Winner Pools

Published Winner Pools are listed separately from weekly rounds. The Original pool and Second-Chance pool each have one season-winner question, their own opening and shared UTC cutoff, and separate accepted submission histories. The later pool opens only after the original cutoff. Missing the original never prevents you from playing future weekly rounds or an available second-chance pool.

Your winner result is Correct, Incorrect, Void or Not entered. It never adds to weekly accuracy or the three scored rounds needed for a non-provisional weekly record. Current anonymous results are private; official public standings require verified accounts.

The reviewed season winner cannot be released before the verified finale end. One editorial result and any correction apply to both published pools in the same transaction, keeping previous versions. If the reviewed winner is absent from a pool’s frozen options, that pool is Void for all participants.

Each pool needs its own cast-viewing confirmation and result unlock. Unlocking a weekly finale or the other pool does not disclose this pool’s result. Published anonymous picks can be restored in this browser for up to 180 days; keep the browser cookies. Current local five-round runs remain separate and stay until you replace them or clear site data.

Try the Rules walkthrough

Choose Rules walkthrough in the practice format selector, then start a five-round fictional season, or join at round 3 to try entering midway. All castaways and results are original fixtures, not real Survivor data. Each round opens with a local practice cutoff 30 minutes later, using this device’s clock. Submit all four answers and save any edits before the cutoff.

Finish round fast-forwards the fictional episode and locks picks immediately, along with any open winner pool. Your last accepted entry is scored; unsaved edits are not. Reveal fictional results when ready, then open the next round. The five rounds demonstrate a single Void question, multiple eliminations, no elimination, not-applicable outcomes and an entirely void round.

Your accepted entries and their versions are saved locally and restored in this browser until you replace the run or clear site data. Export a save file for a backup. This has personal results, not multiplayer standings. Overall accuracy combines resolved question weights; the average round score averages percentages. Both include only unlocked results and exclude all-void rounds.

The original Winner Pool locks when you finish round 1. A separately labelled Second-Chance Winner Pool opens at round 3. Each also has a 30-minute local practice cutoff and locks when its current round finishes. The fictional winner is revealed only after you unlock the finale; winner picks never add to weekly scores.

Clear questions before broadcast

A weekly round contains 3–5 questions with explicit outcomes and settlement rules. Questions must be verifiable from reviewed episode events. No leaked results, boot lists, or betting odds are used.

A question cannot assume there will be exactly one elimination. No elimination, multiple eliminations, and not-applicable results must be handled where relevant. The fictional season uses four questions per round and original fictional results. The Published rounds section shows available question lists; when none are open, choose the historical challenge or fictional practice.

Two question types

Each question has a published weight, defaulting to one. The question wording, options, weight, and rule version are frozen when the round opens.

Locking and accepted entries

Open entries can be edited before the shared cutoff. A live-round deadline will be five minutes before the verified earliest broadcast used by the game. All regions use the same instant, shown in their local time zone.

The last complete, valid submission accepted before locking is the entry used for scoring. Late, queued, or offline requests cannot replace it. A device timestamp does not prove a submission was accepted.

Weighted accuracy

Round score = 100 × correct resolved question weights ÷ all non-void question weights. Results are displayed to two decimal places.

Example: with four equal-weight valid questions and three correct answers, the round score is 75%. Missing a round does not count as an incorrect submitted round, and it does not earn points.

The summary shows round score, overall accuracy, and rounds entered. Average standings require three valid rounds; earlier results are provisional. Cumulative comparisons are limited to clearly labelled groups with the same starting point.

Void questions and corrections

When a question cannot be fairly resolved, it is marked Void and removed from every player's scoring denominator. If all questions in a round are void, the round does not create a zero-score failure and does not count as an entered scored round.

Example: if one question is void in a four-question round and you get two of the three remaining questions right, your score is 66.67%.

Reviewed corrections keep the original resolution, the new version, its source, and the reason. They update every affected player's result consistently.

Join later, start fairly

New players can enter future open weekly rounds, without altering locked history. A Winner Pool uses its own shared cutoff and separate standings.

Any second-chance winner pool starts separately. Later predictions made with more information are not mixed into an earlier winner pool.

Results and community picks

Current-season results will remain behind viewing confirmation. They must not leak through page text, metadata, shared previews, or background requests.

Live community pick percentages may appear only after locking and at least ten eligible submissions. They describe participating users' choices, not real-world winning probabilities. There are no stakes, cash prizes, or payouts.