Liquidity Score Changelog: v3 archive
Historical release evidence. Return to current versions and the complete history index.
What this controls
Liquidity Score Changelog: v3 archive version history
Use this changelog to trace how Pharos changed the related scoring, data-source, threshold, or interpretation rules over time. Entries record the contract active at each release; archived rules are retained so historical charts and citations can be read against the rules active at the time.
Primary references
Latest version
Retained-pool score recomputation and trusted staged-price hardening
Liquidity scoring now rebuilds every aggregate from the retained pool set after filtering/caps, while staged discovery preserves pool-quality metadata and stricter DEX-price trust rules.
Impact snapshot
- Filtered or TVL-capped pools can no longer keep influencing score inputs through stale aggregate fields
- HHI now uses the full retained pool set before display truncation; global 7d volume is pool-deduped
- Staged pool merge now dedups against token-pair fingerprints and preserves raw DEX metadata/quality multipliers
- DEX price observations require a consistent $50K post-confidence TVL floor across source families
v3.4Mar 9, 2026Retained-pool score recomputation and trusted staged-price hardening
Liquidity scoring now rebuilds every aggregate from the retained pool set after filtering/caps, while staged discovery preserves pool-quality metadata and stricter DEX-price trust rules.
- Filtered or TVL-capped pools can no longer keep influencing score inputs through stale aggregate fields
- HHI now uses the full retained pool set before display truncation; global 7d volume is pool-deduped
- Staged pool merge now dedups against token-pair fingerprints and preserves raw DEX metadata/quality multipliers
- DEX price observations require a consistent $50K post-confidence TVL floor across source families
DetailsHide details
Retained-pool score recomputation and trusted staged-price hardening
Liquidity scoring now rebuilds every aggregate from the retained pool set after filtering/caps, while staged discovery preserves pool-quality metadata and stricter DEX-price trust rules.
- Filtered or TVL-capped pools can no longer keep influencing score inputs through stale aggregate fields
- HHI now uses the full retained pool set before display truncation; global 7d volume is pool-deduped
- Staged pool merge now dedups against token-pair fingerprints and preserves raw DEX metadata/quality multipliers
- DEX price observations require a consistent $50K post-confidence TVL floor across source families
Impact notes
- Filtered or TVL-capped pools can no longer keep influencing score inputs through stale aggregate fields
- HHI now uses the full retained pool set before display truncation; global 7d volume is pool-deduped
- Staged pool merge now dedups against token-pair fingerprints and preserves raw DEX metadata/quality multipliers
- DEX price observations require a consistent $50K post-confidence TVL floor across source families
v3.3Mar 9, 2026Separated discovery pipeline with staged pool confidence decay
Discovery sources (CG Onchain, GeckoTerminal, DexScreener, CG Tickers) now run on an independent 20-minute cron with 3x more budget. Staged pools merged into scoring with freshness confidence decay and explicit defaults contract.
- Discovery cron runs independently on 20-min trigger with ~15 min budget (was 5 min shared)
- Staged pools receive confidence decay: max(0.5, 1 - ageHours/48), excluded after 24h
- Chain-aware source routing reduces wasted API calls by skipping irrelevant chains
- Tiered priority with exponential backoff prevents looping on pool-less coins
DetailsHide details
Separated discovery pipeline with staged pool confidence decay
Discovery sources (CG Onchain, GeckoTerminal, DexScreener, CG Tickers) now run on an independent 20-minute cron with 3x more budget. Staged pools merged into scoring with freshness confidence decay and explicit defaults contract.
- Discovery cron runs independently on 20-min trigger with ~15 min budget (was 5 min shared)
- Staged pools receive confidence decay: max(0.5, 1 - ageHours/48), excluded after 24h
- Chain-aware source routing reduces wasted API calls by skipping irrelevant chains
- Tiered priority with exponential backoff prevents looping on pool-less coins
Impact notes
- Discovery cron runs independently on 20-min trigger with ~15 min budget (was 5 min shared)
- Staged pools receive confidence decay: max(0.5, 1 - ageHours/48), excluded after 24h
- Chain-aware source routing reduces wasted API calls by skipping irrelevant chains
- Tiered priority with exponential backoff prevents looping on pool-less coins
v3.2Mar 2, 2026ReconstructedEffective TVL symbol-fallback inflation fix
Corrected effective TVL inflation when symbol fallback matched non-Curve pools to Curve entries.
- Metapool-adjusted TVL now applies only to address-matched Curve pools
- Symbol-fallback pools keep their own TVL in effective TVL calculations
- Removes accidental score inflation from cross-pool symbol collisions
DetailsHide details
Effective TVL symbol-fallback inflation fix
Corrected effective TVL inflation when symbol fallback matched non-Curve pools to Curve entries.
- Metapool-adjusted TVL now applies only to address-matched Curve pools
- Symbol-fallback pools keep their own TVL in effective TVL calculations
- Removes accidental score inflation from cross-pool symbol collisions
Impact notes
- Metapool-adjusted TVL now applies only to address-matched Curve pools
- Symbol-fallback pools keep their own TVL in effective TVL calculations
- Removes accidental score inflation from cross-pool symbol collisions
- 71cc096
v3.1Feb 28, 2026ReconstructedAnti-duplication and protocol TVL cap normalization
Introduced fingerprint-based deduplication and DeFiLlama-anchored cap logic to suppress inflated secondary-source TVLs.
- CG/GT/DS pools deduped against DeFiLlama using token-pair fingerprints
- Secondary-source pool TVL capped and proportionally scaled by protocol-level DeFiLlama ceilings
- Global protocol and chain TVL totals kept consistent after cap reductions
DetailsHide details
Anti-duplication and protocol TVL cap normalization
Introduced fingerprint-based deduplication and DeFiLlama-anchored cap logic to suppress inflated secondary-source TVLs.
- CG/GT/DS pools deduped against DeFiLlama using token-pair fingerprints
- Secondary-source pool TVL capped and proportionally scaled by protocol-level DeFiLlama ceilings
- Global protocol and chain TVL totals kept consistent after cap reductions
Impact notes
- CG/GT/DS pools deduped against DeFiLlama using token-pair fingerprints
- Secondary-source pool TVL capped and proportionally scaled by protocol-level DeFiLlama ceilings
- Global protocol and chain TVL totals kept consistent after cap reductions
- 0b6bfb8
- 617ab25
- 1224015
- 0e54c20
v3.0Feb 28, 2026ReconstructedCoverage expansion with fallback sources
Expanded zero-pool recovery with DexScreener and CoinGecko tickers fallbacks for orderbook-heavy assets.
- DexScreener fallback adds pools for tracked coins still missing after primary crawl
- CoinGecko tickers fallback synthesizes orderbook liquidity where AMM coverage is absent
- Reduces false zero-liquidity outcomes for long-tail and niche assets
DetailsHide details
Coverage expansion with fallback sources
Expanded zero-pool recovery with DexScreener and CoinGecko tickers fallbacks for orderbook-heavy assets.
- DexScreener fallback adds pools for tracked coins still missing after primary crawl
- CoinGecko tickers fallback synthesizes orderbook liquidity where AMM coverage is absent
- Reduces false zero-liquidity outcomes for long-tail and niche assets
Impact notes
- DexScreener fallback adds pools for tracked coins still missing after primary crawl
- CoinGecko tickers fallback synthesizes orderbook liquidity where AMM coverage is absent
- Reduces false zero-liquidity outcomes for long-tail and niche assets
- 6b2e006
- ef9bb2b