Prediction Standards & Participation Guide

• A clear outcome, a clear deadline, and a reliable verification source.
• Neutral wording with no hidden information advantage.
• No ambiguity about what counts as “Yes” or “No.”
• No outcome that is already known before the prediction is created.

A good prediction names the subject, the measurable event, the deadline, the source, and the timezone — especially for “today,” “tomorrow,” or price predictions.

Avoid vague words unless they are precisely defined — “affect”, “soon”, “safe”, “popular”, “best”, “strong”, “successful”, “increase” without a baseline, or “by end of today” without a timezone.

A prediction must not be created or left open after the result is already available.

• Sports: close before match start.
• Price: close before the final measurement time.
• Elections / announcements: close before official results are expected.
• “Today” predictions: specify exact timezone and end time.

The creator or moderator must not reinterpret a prediction after the fact. If no rules are defined, the ordinary public meaning applies — but ambiguity may result in voiding. Settlement must never be based on personal preference, majority staking direction, or a private explanation offered after close.

• It passed its close time without proper settlement.
• The outcome cannot be reliably verified, or has no objective Yes/No standard.
• The wording is too unclear, or the outcome was already known before creation.
• The event was cancelled or materially changed, or the wrong close time was used.
• There is evidence of manipulation, or it depends on private / insider information.

Creators are responsible for the quality of their wording. Creators must not:

• Create intentionally vague predictions to settle in their favor.
• Create predictions where they already know the outcome.
• Encourage coordinated staking, or farm fees with low-quality predictions.
• Set wrong deadlines to allow staking after the outcome is known.
• Settle incorrectly, use multiple accounts, or rely on private information.
• Recreate the same malformed prediction after warnings.

Before staking, check what exact event is predicted, when it closes, what source verifies it, whether the wording is neutral, and whether the outcome could already be known. Avoid predictions that are ambiguous, rumor-based, already resolved, subjective, or missing a timeframe or source.

Earnings may be deducted, withheld, or cancelled if a prediction was wrongly settled, voided, misleadingly worded, based on an already-known outcome, or showed signs of coordinated or abusive activity.

Creators may participate, but self-staking alone becomes a concern when combined with other signals:

• Unusually large creator stakes on their own prediction.
• The same accounts repeatedly staking on the creator’s predictions.
• Most stake concentrated on one side.
• Related or newly created accounts participating in similar patterns.
• Vague or hard-to-verify predictions.

A prediction may be reviewed before withdrawal approval when more than 90% of stake sits on one side, the same users repeatedly stake together or win from the same creator, the creator’s own stake is unusually large, or earnings are generated quickly from a small group. This does not automatically prove misconduct — but it may trigger review.

Will [specific event] happen by [specific date and time, timezone], according to [verification source]?

Record the prediction ID, exact wording, close time, outcome source, result, and winning option — and whether the wording was clear enough to settle or a refund/void is needed.

• Is the outcome clear, and is it still unknown?
• Can anyone verify it, and did I define the source, deadline, and timezone?
• Is the wording neutral?
• Would I be comfortable with a moderator settling this strictly as written?

If not, rewrite the prediction before posting.

• The event has not already happened.
• The close time is before the outcome is known.
• The result can be verified publicly.
• The wording is neutral and the winning condition is measurable.
• The data source and timezone are clear.
• The options are mutually exclusive.
• The prediction does not rely on private information.
• The creator can explain how to settle it using only the written text.