CrowdCent Points (CC Points)
The CrowdCent Points (CC Points) system rewards consistent participation and high-quality predictive performance.
Summary
- Participation: Earn a base credit that grows with your daily streak.
- Performance: Earn the majority of points from a blended points percentile — each non-experimental slot contributes 50% raw composite + 50% unique composite, averaged across your slots.
- Influence: Your points translate directly to your weight in the Meta-Model.
Daily Base Credit
You earn a guaranteed 0.5 points just for participating, provided you make at least one valid submission during the daily window.
Maintaining a daily streak increases your base credit:
| Streak Length | Multiplier | Daily Base |
|---|---|---|
| Days 1–29 | 1.0x | 0.5 pts |
| Days 30–89 | 1.5x | 0.75 pts |
| Days 90+ | 2.0x | 1.0 pts |
Missing a submission window resets your streak to 0.
Auto-Rollover Helps Maintain Streaks
By default, submissions are automatically queued for the next period (auto-rollover). This means you can submit once and have it count for the next day, helping maintain your streak for infrastructure hiccups.
Performance Adjustment (The Core Driver)
Your primary source of points is the quality of your predictions relative to other participants.
At the end of the day, all your valid non-experimental slots are evaluated. Your performance adjustment applies a cosine curve to an average points percentile across those slots.
For each slot, raw and unique count equally:
For example, a slot at the 80th composite and 40th unique composite yields a 60th slot points percentile.
Your average points percentile is the mean of those slot values across all evaluated non-experimental slots. See composite percentile and unique composite percentile for how those inputs are computed.
Slots flagged as experimental still receive shadow percentiles for your reference but are excluded from this calculation.
Earlier periods
Inference periods released before July 1, 2026 used the raw composite only. That formula is unchanged for those periods; earlier ledger entries are not recalculated.
The Curve
The curve rewards consistency—moving from "average" (50th) to "good" (70th) yields the steepest point gains. Bad days are penalized, but less severely than a linear model. The curve flattens at the extremes to discourage variance.
Formula:
where AvgPointsPercentile is the average points percentile across your evaluated non-experimental slots, as described above.
Key Benchmarks:
- 0th percentile: -10.0 pts
- 50th percentile: 0.0 pts
- 60th percentile: +3.1 pts
- 70th percentile: +5.9 pts
- 80th percentile: +8.1 pts
- 100th percentile: +10.0 pts
Final Daily Score
Your final score for the day cannot be negative.
Example (Assuming >90 day streak = 1.0 base):
| Scenario | Points Percentile | Base (w/ Mult) | Adjustment | Calculation | Final Points |
|---|---|---|---|---|---|
| Top Tier | 80th | 1.0 | +8.1 | 1.0 + 8.1 |
9.1 |
| Consistent | 70th | 1.0 | +5.9 | 1.0 + 5.9 |
6.9 |
| Average | 50th | 1.0 | 0.0 | 1.0 + 0.0 |
1.0 |
| Poor Day | 20th | 1.0 | -8.1 | 1.0 - 8.1 |
0.0 |
Tier System
As you accumulate CC Points, you'll progress through 5 tiers—each with a unique badge displayed on your profile and in the leaderboard:
| Tier | Points Required | Description |
|---|---|---|
| Citizen | >0 | Welcome to CrowdCent |
| Challenger | 100+ | Active participant |
| Contender | 500+ | Rising competitor |
| Centurion | 1,500+ | Commander of predictions |
| Sovereign | 5,000+ | Pinnacle of mastery |
Higher tiers unlock greater recognition in the community and demonstrate your track record of consistent, quality predictions. Reaching Challenger also unlocks real-time meta-model predictions (users below Challenger receive 90-day delayed data).
Your Points = Your Toolkit. Tiers also unlock the Simulator, where the meta-model becomes a long/short Hyperliquid portfolio you can backtest:
| Tier | Simulator unlocks |
|---|---|
| Everyone | Full-realism backtesting on 90-day delayed meta-model data |
| Challenger | Real-time meta-model, Inverse-Vol & HRP weighting, parameter sweeps |
| Contender | Risk & capacity controls, full sweep & blend budgets |
| Centurion | Classified alpha controls |
The Meta-Model: Why Points Matter
The ultimate goal of the CrowdCent Challenge is to build a meta-model—an aggregate of all participants that outperforms any single participant on their own.
Your Points = Your Weight. The influence of your predictions on the meta-model is proportional to the Exponential Moving Average (EMA) of your Total Points history. Recent performance is weighted significantly higher than past performance, so you cannot "rest on your laurels." To maintain high influence, you must perform consistently.
Formula:
Program Terms
CrowdCent reserves the right to modify, suspend, or cancel the points system, scoring rules, and reward structures at any time without prior notice.
This includes, but is not limited to:
- Adjusting point values, multipliers, or calculation formulas retroactively or proactively.
- Removing points from users found to be in violation of the spirit of the competition.
- Changing the weighting mechanisms for the meta-model.
- Resetting point totals for specific challenges or globally.
Participation in the points program does not create any property rights or contractual obligations. Points have no monetary value and are used solely for ranking and weighting purposes within the CrowdCent ecosystem.