HivePredict Fairness Updates: Stronger Anti-Snipe Rules

We shipped a fairness update focused on one goal: make prediction markets reward conviction and early participation, not last-minute sniping. This is a learn as you go type deal where your feedback helps shape how HivePredict works. It's exciting, but also difficult to get right for version 1.0 or even version 2.0, but we are receptive to improving this.

If you think things should work differently, please let @beggars know and don't be shy, a lot of this platform has been improved from community feedback.

What Changed

1. Crypto trading cutoff moved earlier

   • Crypto markets now close for new predictions at 20% of market duration before betting close (previously 10%).
   • This removes the final stretch where traders could wait for near-final information and jump in at the last moment.

2. Late-entry weighting became much stricter

   • Profit-share weighting now uses:
     • weight = stake * (1 - lateEntryPenaltyRate * timeFraction)^lateEntryPenaltyPower
   • Current parameters:
     • lateEntryPenaltyRate = 1.0
     • lateEntryPenaltyPower = 2
   • At the cutoff, late entries have effectively no profit-share weight.

3. Transparency in market UI

   • The Predictions tab now shows time fraction and effective weight per prediction.
   • The resolved Distribution & Proof tab shows weighted stake evidence so users can audit how profit-share was determined.

Why This Is Fairer

• Encourages early participation: entering earlier meaningfully improves profit-share weight (something we already had but have made stricter)
• Reduces snipe upside: late entries can still participate, but with sharply reduced edge.
• Removes predictability gaming: earlier cutoff windows make outcomes harder to exploit with near-resolution information.

Rollout Safety: No Retroactive Distortion

This update is forward-applied.

• Existing markets already resolved before this deployment are not recalculated.
• Historical payout records remain exactly as they were settled.
• Weighting visibility on resolved markets is now driven by the weighting snapshot stored at resolution time.
  • If a resolved market does not include that snapshot (older markets), weighting breakdown is not shown to avoid misrepresenting historical settlement logic.

In short: old resolved markets keep their original economics; new and not-yet-resolved markets use the new fairness rules.