Skip to content

Chrono Score

Chrono Score (CS) is a continuous, non-transferable forecasting reputation score. It is the foundational infrastructure of the Chronomancy protocol — every module’s pricing, access, and terms are gated by it.

CHRONO SCOREOn-chain reputation

A soulbound (non-transferable) on-chain record of your forecasting skill. It accumulates across every prediction you make — and unlocks better terms across all Chronomancy products.

Accuracy
Did your predictions resolve correctly?
Calibration
Were your confidence levels well-tuned?
Market volume
How much did you put at stake?
Time in system
Consistent track record over epochs
What it unlocks
CS 1–3
Developing
  • Fast-Forward: +5% exit premium
  • Rewind: Silver tier (70% recovery)
CS 4–6
Established
  • Fast-Forward: +10% exit premium
  • Rewind: Gold tier (100% recovery)
CS 7–10
Elite
  • Fast-Forward: +18% exit premium
  • Rewind: Gold tier + priority access

A soulbound score that measures how well you predict outcomes. It answers: “Given this wallet’s track record, what is the probability-adjusted edge on their next prediction?”

CS is not just a leaderboard score. It is a financial credit score — high CS translates directly to:

  • Better Fast-Forward exit prices (the vault pays more for your positions)
  • Cheaper Rewind insurance premiums (you claim less often)
  • Priority allocation when capacity is constrained

CS is computed internally using the capped logarithmic scoring rule:

score = max(-ln(p), 3.0)

Where p is your stated probability for the outcome that occurred. The cap at 3.0 prevents catastrophic losses from overconfident predictions — a 1% prediction on a correct outcome scores the same as a 5% prediction, not infinitely better.

The publicly displayed number is the Brier score (mean squared error between your predictions and outcomes). Brier score is easier to communicate: 0 = perfect, 0.25 = uninformed guessing. The Metaculus community aggregate is ~0.107. GJP superforecasters outperform intelligence analysts with classified information by ~30%.

CS outputs two signals:

SignalWhat It MeasuresMethod
CS LevelCumulative prediction accuracyCapped log scoring, running total
CS ConfidenceRating uncertaintyGlicko-2 Rating Deviation (RD)

Glicko-2 was chosen over Elo because it explicitly tracks uncertainty — a forecaster who just returned from a long break has high RD (uncertain CS), and the vault prices accordingly.


Near-certain markets (85%+ probability) contribute less to CS:

difficulty_weight = 1 - 2 * |p - 0.5|

A market at p=0.50 has weight 1.0. A market at p=0.85 has weight 0.30. A market at p=0.95 has weight 0.10. Combined with Bayesian shrinkage (κ=150 predictions before raw score dominates), farming a fake CS requires hundreds of real-money predictions across difficult markets over months — at which point you’ve either proven genuine skill or lost enough money that the attack is uneconomical.

New forecasters with few predictions are shrunk toward the population mean:

CS_adjusted = (N × CS_raw + κ × CS_prior) / (N + κ)

κ = 150 — only after 150 predictions does your raw score dominate. This prevents lucky streaks from dominating early CS.

Recent predictions count more:

decay_factor = λ^months_ago, λ = 0.95

A prediction from 12 months ago retains ~54% weight. Keeps CS responsive to improving or deteriorating performance.

CS requires ≥2 predictions per month to maintain active status. Inactive wallets see CS freeze (no decay, but no access to Premium tier until activity resumes).


Global CS is supplemented by domain subscores for: Politics, Sports, Crypto, Geopolitical, Science/Tech.

Domain subscores enable precision pricing:

  • A user with high crypto CS but average global CS gets better FF terms on crypto markets specifically
  • A user with high politics CS gets Rewind discounts on election markets

TierPredictionsCS PercentileFF Max PositionStaking Required
Unscored0–29No access
Basic30–49Top 50%$1K
Standard50–99Top 25%$10K1,000 $TM
Premium100+Top 10%$50K10,000 $TM

Staking requirements create a cost to holding Premium tier — aligning incentive for high-CS users to maintain their position.

The Staking Tension: Meritocracy vs. Sybil Resistance

Section titled “The Staking Tension: Meritocracy vs. Sybil Resistance”

An honest acknowledgment: staking gates create a capital barrier that tensions with the meritocratic premise. A brilliant forecaster with no capital gets worse financial terms than a mediocre forecaster who can afford 10,000 $TM.

This is a known design trade-off, not an oversight:

  • Without staking gates: Sybil attacks are trivially cheap. Create 100 wallets, farm CS on each, extract maximum value. The cost of being wrong is zero.
  • With staking gates: There is a real financial cost to holding Premium tier. This aligns skin-in-the-game with access.
  • The mitigation path: As stronger Sybil resistance matures (World ID biometric verification, Humanode, cross-platform import verification via TLSNotary), staking thresholds can decrease or be replaced. The goal is to make staking gates unnecessary — but launching without them is launching with an open Sybil vulnerability.

Basic tier (top 50% CS) has no staking requirement. The gate only applies to Standard and Premium — the tiers with significant financial privileges.


Users with verified track records from other platforms can import them:

PlatformCredit AppliedVerification Method
Metaculus50%TLSNotary proof of public profile
Good Judgment Project50%TLSNotary
Polymarket (past trades)30%On-chain via subgraph, auto-detected
Manual superforecaster100%Reviewed by team

Import credit is capped at top-50% CS tier — you cannot skip directly to Premium without on-platform track record. Imported scores decay faster unless confirmed by subsequent on-platform performance.


Do Financial Modules Distort Forecasting Incentives?

Section titled “Do Financial Modules Distort Forecasting Incentives?”

If you can insure losses (Rewind) and exit early (FF), does that change how you predict? The short answer: CS scores prediction accuracy at time of entry, not financial outcomes.

  • Your CS updates based on whether your prediction was correct — regardless of whether you later activated Rewind or Fast-Forward.
  • A user who buys Rewind on every position and a user who never touches Rewind generate identical CS trajectories if their predictions are equally accurate.
  • FF doesn’t affect CS either: you exit the financial position, but the prediction outcome is still attributed to your record when the market resolves.

The modules are financial tools layered on top of the reputation system, not inputs to it. Your Chrono Score measures how well you predict, not how well you manage money.


A manual tier override for verified, publicly known superforecasters who have extensive track records on other platforms but haven’t used Chronomancy yet.

Named after Scott Alexander (Astral Codex Ten, Mantic Monday), who has years of published forecasting track record but would start at CS zero on a new platform. The clause allows the team to verify a public identity, confirm their external track record, and grant Premium tier access without requiring them to re-prove their skill from scratch.

This is an onboarding shortcut for known quantities — not a bonus for surprising predictions or a subjective override on scoring. It applies only to publicly verifiable individuals and is expected to be used rarely (tens of people, not thousands).


The vault tracks whether its CS-based pricing actually earns alpha:

  • Meta-score: Rolling 90-day comparison of vault returns from CS-gated FF exits vs. a naive market-maker baseline
  • Auto-adjustment: If meta-score shows CS alpha is above target, the vault widens FF premium ranges. If below, it tightens.

This closes the feedback loop between CS accuracy and vault economics.


AI agents can participate in prediction markets and accumulate CS. The protocol does not gate by “human” identity — it gates by track record. An AI agent with a strong CS gets the same FF terms as a human.

This is intentional: Chronomancy is building a meritocratic financial layer. The value proposition is your accuracy, not your biology.

Related: