Live
CuratorWatch
The annualized probability that a dollar entrusted to a curator suffers a loss, decomposed into the channels they control. Loss axis only — yield and fees are separate.
Updated Jul 6, 2026 · pinned snapshot, refreshed manually
Fit on everything before November 2025 with zero loss events in the training data — structure and priors only — then asked to rank curators going into the Stream-collapse window. It put the 8 curators that actually took losses in the riskiest quartile 2.5× more often than chance. This is the honest test: rate the losers low in advance.
Pre-event temporal backtest, cutoff 2025-11. Converged clean (0 divergences, r̂ 1.003). Honest caveat: structural features are current-book, so this tests model form + cross-sectional discrimination, not a full point-in-time replay. v1 (curatorwatch-risk-methodology-v1.md).
| Curator | Grade | Expected loss / yr | Confidence | Signals | |
|---|---|---|---|---|---|
| 1 | NuWealth | NR | 0.1 bps | Tight | — |
| 2 | Avantgarde | A+ | 0.3 bps | Tight | — |
| 3 | SigmaLabs | NR | 0.6 bps | Moderate | — |
| 4 | K3 Capital | NR | 0.8 bps | Tight | Declined risk |
| 5 | Felix | A+ | 0.9 bps | Tight | — |
| 6 | Pangolins | A+ | 1.0 bps | Tight | — |
| 7 | Unified Labs | NR | 1.0 bps | Tight | — |
| 8 | GOATed | NR | 1.1 bps | Moderate | — |
| 9 | Sky | A+ | 1.1 bps | Moderate | — |
| 10 | Apostro | NR | 1.2 bps | Tight | — |
| 11 | My USDS | NR | 1.3 bps | Moderate | — |
| 12 | KPK | A+ | 1.4 bps | Moderate | Declined risk |
| 13 | Galaxy Curation | A+ | 1.7 bps | Moderate | — |
| 14 | SparkDAO | A+ | 2.4 bps | Tight | — |
| 15 | Hakutora | A+ | 2.6 bps | Tight | — |
| 16 | Api3 | A+ | 3.6 bps | Moderate | — |
| 17 | UltraYield | NR | 3.7 bps | Moderate | — |
| 18 | 3Jane Ecosystem | A+ | 4.4 bps | Moderate | — |
| 19 | AlphaPing | A+ | 6.1 bps | Moderate | — |
| 20 | SingularV | A+ | 6.9 bps | Wide | — |
| 21 | Etherealm | NR | 8.3 bps | Wide | — |
| 22 | Steakhouse Financial | A+ | 8.3 bps | Wide | Declined risk |
| 23 | Rini Oak | NR | 9.0 bps | Wide | — |
| 24 | Hackarrot | NR | 11 bps | Wide | — |
| 25 | Flowmark | NR | 11 bps | Wide | — |
| 26 | Yearn | A | 16 bps | Wide | — |
| 27 | Mainstreet | A | 17 bps | Wide | — |
| 28 | Clearstar | A | 18 bps | Wide | Declined risk |
| 29 | DeTrade | NR | 23 bps | Wide | — |
| 30 | Anthias Labs | A | 25 bps | Wide | Repeat offender |
| 31 | RockawayX | Flagged | 27 bps | Wide | — |
| 32 | Wintermute | A | 33 bps | Wide | — |
| 33 | Gauntlet | A | 33 bps | Wide | Declined riskRepeat offender |
| 34 | Flowdesk | B+ | 37 bps | Wide | — |
| 35 | Keyrock | B+ | 67 bps | Wide | — |
| 36 | CCM | NR | 68 bps | Wide | — |
| 37 | Sentora | B+ | 70 bps | Wide | — |
| 38 | Origin Protocol | B | 80 bps | Wide | — |
| 39 | August Digital | B | 85 bps | Wide | — |
| 40 | Hyperithm | B | 86 bps | Wide | Declined risk |
| 41 | B.Protocol | NR | 101 bps | Wide | — |
| 42 | Trezor Steakhouse | B | 123 bps | Wide | — |
| 43 | OV | NR | 125 bps | Wide | — |
| 44 | Gtsy | NR | 128 bps | Wide | — |
| 45 | Waterline RUSD | B | 150 bps | Wide | — |
| 46 | Cross Collat | NR | 201 bps | Wide | — |
| 47 | RE7 Labs | C+ | 221 bps | Wide | Repeat offenderDiligence conflict |
| 48 | Bloom | NR | 235 bps | Wide | — |
| 49 | Chip | NR | 244 bps | Wide | — |
| 50 | Utorg | NR | 247 bps | Wide | — |
| 51 | Brrrstr | NR | 247 bps | Wide | — |
| 52 | Fiber | NR | 248 bps | Wide | — |
| 53 | HavenFi | NR | 249 bps | Wide | — |
| 54 | MEV Capital | C | 329 bps | Wide | Repeat offenderDiligence conflict |
| 55 | Stake DAO | C | 360 bps | Wide | — |
Sorted safest first by expected loss. Every band is currently Wide-confidence (early data); grades are durable and TVL-weighted across each curator’s live book. Open a curator for the channel breakdown.