Overview

Prediction markets are becoming real-time belief infrastructure — the Federal Reserve now benchmarks Kalshi against Bloomberg consensus, Goldman Sachs has opened an event desk, ICE committed up to $2B at an $8B Polymarket valuation, and notional volume has grown 130× to $13B per month in 18 months. But the ecosystem has a structural trust gap: nobody can tell who’s credible, resolution fails catastrophically, and 72% of participants lose money and leave.

Chronomancy is the trust-and-sustainability layer. Three products, sequenced by capital intensity:

• Chrono Score — calibration-based forecaster reputation
• Fast-Forward — structured early-exit vault
• Rewind — resolution-failure insurance

Zero of 264+ tracked PM projects build any of these. We wrap existing venues (Polymarket, Kalshi, Azuro) rather than competing with them. Revenue is fees and spreads — no token in the base case.

Narrative hierarchy

The framing pyramid everything else descends from.

LEVEL 1 — WHY (epistemics)
Prediction markets are becoming the infrastructure for how institutions
and societies form beliefs about the future.

LEVEL 2 — WHAT (trust layer)
That infrastructure has a trust gap: no one knows who's credible,
resolution fails catastrophically, 72% of participants lose and leave.
Chronomancy fills the gap.

LEVEL 3 — HOW (three products)
Chrono Score: calibration-based reputation, portable across platforms.
Fast-Forward: structured early exit at reputation-adjusted prices.
Rewind: insurance against resolution failure.

LEVEL 4 — EVIDENCE (52 academic papers, industry data)
Skill is real and persistent (89.4% out-of-sample accuracy, top 5%).
Selection outperforms aggregation by 21% (Atanasov 2024).
Resolution risk is inherent — Oracle Trilemma formally proven.
Institutional demand is validated: Fed paper, ICE, Goldman, $35M fund.

The three structural failures

1. No reputation layer

Track records die with the platform. Nobody knows who’s skilled. Yet the literature is unambiguous that skill is real, persistent, and concentrated in a tiny minority.

• Top 1% capture 84% of gains (Akey 2026)
• 3.1% of wallets take all $228M net profit (Yang 2026)
• Only 7.3% of traders are “price-sensitive” (Bossaerts 2022)
• Skill persists across seasons (Atanasov 2024: r=0.38); top-5% wallets sustain 89.4% accuracy in hold-out samples (Yang 2026)
• Capital size is anti-correlated with edge (Deleep 2026) — whales are not the elite

2. Resolution is the weakest link

Every major scandal is a resolution failure. The Oracle Trilemma formally proves no blockchain oracle can simultaneously achieve decentralization, truthfulness, and scalability (Cong 2025). Oracle attacks cost $403M in 2022 alone (Duley). MakerDAO Black Thursday: $8.32M liquidated at zero DAI. Compound: $89M bad debt. Pyth: 90% BTC/USD misreport.

Resolution risk is not priced — it is responded to with liquidity flight. A 0.1 increase in Semantic Risk Score reduces contract volume by ~60% (Sanjabi 2026). This is an Akerlof lemons dynamic and an Akerlof opportunity.

3. 72% of participants lose and leave

The optimal strategy (full Kelly) produces a 75% chance of a 25% drawdown and a 25% chance of a 75% drawdown (Osband 2025). The math is psychologically terrifying even for the skilled. For everyone else, it is a retention crisis: 72% lose money, the share is increasing as new entrants arrive (Reichenbach 2025), 63% of average-user trades happen at extreme prices (<10¢ or >90¢) (Akey 2026).

The white space

Zero of 264+ tracked prediction market projects build insurance, derivatives, or protocol-level reputation (PredictionIndex 2025–2026). The closest analogues — Gondor (lending), Causal Markets / Sozu (information layers), functionSPACE (resolution primitives) — are adjacent, not competitive. The financial-infrastructure layer is empty.

Build order

Phase	Product	Capital	Duration	Status
0	Chrono Score (read-only, cross-platform)	$0 seed	4–8 weeks	In progress
1	Fast-Forward vault	$50–200K seed + $30–50K audit + $60–100K Solidity	10–14 weeks	Blocked on Solidity hire + legal entity
2	Rewind insurance	$200–500K pool + $50–80K audit + $15–25K quant + legal	14–20 weeks	Blocked on Phase 1 + $420–985K raise

Non-negotiable: ship Chrono Score before raising. The free, public beta is the credibility-building artifact.

Why this reads as honest

We have a bear case, and the bear case is a business.

• If Chrono Score shows no out-of-sample alpha after 6 months, Fast-Forward degrades to a standard market-maker capturing bid-ask spread plus lifetime-decay premium (both quantified in Yang 2026). Still a business.
• If PM volume collapses, the build pauses; Chrono Score retains signal value for Metaculus- and GJP-style deployments.
• If resolution risk turns out to be well-priced (premiums consistently exceed claims), Rewind becomes a Nexus-style flat-fee infrastructure subsidy rather than a profit center.

Full conceptual model — products, two-score architecture, economics, risks, team — is in the canonical doc and research hub.