Rewind

PHASE 2

REWIND is position insurance. A short-dated put option on prediction market positions. If the market resolves against you, claim back 70% of your entry. Premium is priced by probability, time to resolution, pool utilization, and your Chrono Score.

graph LR
    User -->|pays premium| RewindPool
    User -->|receives| PolicyNFT
    MarketResolves -->|user loses| ClaimsContract
    ClaimsContract -->|pays 70% USDC| User
    ClaimsContract -->|burns| PolicyNFT
    MarketResolves -->|user wins| PolicyExpires[Policy expires worthless]

How It Works

Select a position — the Hub shows your open Polymarket positions with quote cards
See your premium — priced by current probability, time to resolution, your CS tier
Pay in USDC — a Rewind policy NFT (ERC-721) mints to your wallet
If you lose — return within 24h of resolution, claim 70% payout; policy burns
If you win — policy expires worthless; the premium was the cost of certainty

⏪REWINDPosition insurance

Position size (USDC)

$500

Recovery tier

Scenario: market resolves against you. You lost $500.

Pricing

MVP Formula (v0.1)

Premium = PositionSize × 0.70 × (1 − p) × timeDecayMultiplier × utilizationMultiplier

Variable	Definition
`p`	Current market probability for your side
`timeDecayMultiplier`	`1 / √T` where T = days to resolution (floor 1.0)
`utilizationMultiplier`	`1 + max(0, (util − 0.70) × 10)` — Aave-style kink at 70%

Worked example: p=0.72, $500 position

Coverage: $500 × 0.70 = $350
Base: $350 × (1 − 0.72) = $98
With favorable time/utilization: ~$58.80 (≈12% premium)

Advanced 5-Layer Formula (v0.2+)

Premium = Position × 0.70 × (1−p)
        × θ(T, σ)          // time decay via Dalen logit jump-diffusion
        × λ(U)             // utilization kink (Aave/Compound model)
        × γ(demand, dp/dt) // surge: Nexus Mutual + momentum
        × (1 + ρ × E_cluster/Pool)  // correlation charge

The Dalen model (arXiv:2510.15205) captures how near-expiry positions have dramatically higher effective volatility — insurance on a 3-day market costs far more than on an equivalent 30-day market.

CS Band Discounts (v0.2+)

Band	CS Tier	Premium Modifier	Claims Frequency
A	Premium	0.70–0.80× (20–30% discount)	18–22%
B	Standard	0.85–0.90×	25–30%
C	Basic	1.00×	30–35%
D	Unscored	1.10–1.20× (surcharge)	40–50%

High-CS users claim less often — the actuarial reality. They pay less.

Adverse Selection Defense

A common objection: “Past accuracy doesn’t predict any single future bet.” This is correct — and irrelevant. Rewind doesn’t price individual bets. It prices risk classes across portfolios, exactly like auto insurance. Your driving record doesn’t predict your next accident, but it statistically predicts your claims frequency over the next year. CS works the same way: a Band A forecaster claims 18-22% of the time vs. 40-50% for Band D. The actuarial spread is the profit margin.

The deeper defense: users who buy Rewind on their worst positions (adverse selection) are priced for it. The 5-layer premium formula incorporates market probability, time decay, and utilization surge — a last-minute insurance purchase on a collapsing position is priced accordingly (high premium, often uneconomical to buy). Adverse selection is a pricing problem, not an existence problem, and the pricing engine is built for it.

”Isn’t This a CDO?” — Why It’s Not 2008

The Junior/Senior tranche structure invites comparison to CDOs. The critical difference: CDOs failed because their underlying assets were all correlated (mortgages → housing market). When housing fell, every tranche failed simultaneously.

Chronomancy’s pools are categorically segregated: Elections, Sports, Crypto, Geopolitical. A political black swan (unexpected election result) drains the Elections pool — it has zero impact on Sports or Crypto pools. The 10% per-event and 25% correlated-cluster concentration limits further prevent single-event contagion. This is the structural lesson of InsurAce’s Terra collapse, and it’s why Nexus Mutual V2 (which Rewind models after) has survived $200M+ in cover without a solvency crisis.

Pool Architecture

Modeled on Nexus Mutual V2 ($200M capital, $5.1B cover written, 4.8× gearing at 99.5% solvency).

Category Segregation

Pool	Markets Covered
Elections	Political prediction markets
Sports	Sports betting markets
Crypto	Crypto/DeFi markets
Geopolitical	War, diplomacy, macroeconomic

Segregation prevents a single correlated event (e.g., US election) from draining the entire protocol.

Concentration Limits

Limit	Cap
Per single event	Max 10% of relevant pool
Correlated cluster	Max 25% of pool on correlated events

The InsurAce Cautionary Tale

InsurAce suffered catastrophic losses on the Terra/LUNA collapse in 2022 — $11.4M in claims, which wiped out their claims reserve. Post-mortem: insufficient concentration limits on correlated positions (everything Terra-adjacent triggered simultaneously). Chronomancy’s 10% per-event and 25% cluster caps are designed specifically to avoid this failure mode.

Junior/Senior Tranching (v0.2+)

Tranche	Funded By	Risk Profile	Target Yield
Junior	$CHRONO stakers	First-loss — absorbs drawdowns	25–40% APY
Senior	External USDC LPs or raise	Protected — last to take losses	3–8% APY

The community-funded junior tranche reduces the external capital requirement by 30–50%.

Smart Contracts

Four contracts on Polygon. See Smart Contracts for full detail.

Contract	Role
`RewindPool.sol`	ERC-4626 capital vault
`RewindPolicy.sol`	ERC-721 policy NFTs
`RewindPricingEngine.sol`	Premium calculation
`RewindClaims.sol`	Claims verification + payout

Token Integration Roadmap

Version	$TM Integration
v0.1	USDC-only. $TM not required.
v0.2	$TM stakers provide junior tranche
v0.3	CS-linked $TM burn for premium discounts
v0.4	$TM accepted as premium payment token

Design Alternatives Considered

Pari-Mutuel Insurance (Rejected as Primary, Possible Hybrid)

An alternative to fixed-payout insurance: users pay premiums into an isolated pool for a specific market. If the event triggers, the pool is divided pro-rata among claimants. The protocol takes zero actuarial risk — the pool can never be insolvent.

Why it’s structurally elegant: Mathematical insolvency is impossible. Payouts are just whatever premiums were collected, divided among claimants. No junior tranche needed, no concentration limits needed, no correlation charge needed.

Why it was rejected as the primary model: Pari-mutuel fails exactly when users need insurance most. In a correlated event (election upset), many users claim simultaneously — each gets a tiny fraction of the pool. The insurance promise evaporates at the moment of maximum need. Fixed-payout with proper pool architecture (segregation, concentration limits, actuarial pricing) offers a stronger user promise.

Why dual-product doesn’t work: Offering both creates an adverse selection engine. Sophisticated users buy fixed-payout for correlated risks (guaranteed 70%, protocol absorbs the hit) and pari-mutuel for idiosyncratic risks (where the pool won’t be diluted). The toxic correlated risk concentrates in the fixed-payout product — exactly what it’s designed to avoid.

Possible hybrid (to explore): Fixed-payout Rewind reserved for high-CS users (who have demonstrated actuarially favorable claims frequency), with pari-mutuel as the default for lower tiers. This uses CS as the selection mechanism — the protocol takes actuarial risk only on the user segments where the math works. Lower-CS users still get some recovery through pari-mutuel, just without a guaranteed amount. This requires further analysis of whether the CS tier boundary creates its own adverse selection problem.

Related:

Cross-Product Economics — Rewind + FF profitability across CS bands
Retention Flywheel — how Rewind turns losses into re-engagement
Smart Contracts — full contract architecture
Chrono Score — the CS system that drives pricing