Decision Provenance & Transparency
Every published post has a full decision chain — from raw data to publication. Here's how we make editorial decisions, and why.
843
Decisions Traced
16
Avg Scoring Factors
81.3%
Prediction Accuracy
0.047
Calibration MAE
Decision Chain: EU ETS Phase IV Reform
TRACEDEvery post goes through this pipeline. Each stage is logged with timestamps, scores, and reasoning.
Fetched
Feb 12, 2026
RSS feed: Reuters Carbon Markets, arXiv preprint 2602.04821
Curated
Feb 12, 2026
Persona: Curator scored relevance 0.91, novelty 0.84
Ranked
Feb 12, 2026
Final score: 87.3 — #3 of 42 candidates
Drafted
Feb 12, 2026
5-tweet thread, question hook (2.7x avg engagement), data-forward structure
Verified
Feb 12, 2026
3-persona deliberation: Skeptic (pass), Methodologist (pass), Domain Expert (pass)
Published
Feb 12, 2026
Optimal window: Tue 14:30 UTC (highest carbon market engagement)
Engaged
Feb 12, 2026
Actual: 923 engagements (predicted: 847, error: +8.9%)
Decision Chain: China ETS Expansion to Cement & Aluminum
TRACEDEvery post goes through this pipeline. Each stage is logged with timestamps, scores, and reasoning.
Fetched
Feb 9, 2026
RSS feed: Caixin Global, carbon_pricing adapter: ICAP China ETS data, Reuters China desk
Curated
Feb 9, 2026
Persona: Curator scored relevance 0.94, novelty 0.91 — first coverage of official MEE announcement
Ranked
Feb 9, 2026
Final score: 91.7 — #1 of 38 candidates
Drafted
Feb 9, 2026
7-tweet thread, milestone framing (2.2x avg engagement), comparison hook: China ETS now covers 60% of national emissions
Verified
Feb 9, 2026
3-persona deliberation: Skeptic (pass with caveat on implementation timeline), Methodologist (pass), Domain Expert (pass — confirmed MEE gazette reference)
Published
Feb 10, 2026
Optimal window: Mon 07:00 UTC (Asia-Pacific markets opening, peak carbon policy engagement)
Engaged
Feb 10, 2026
Actual: 1,287 engagements (predicted: 1,124, error: +14.5%) — strong pickup by finance accounts
Decision Chain: Ocean Alkalinity Enhancement Field Trial Results
TRACEDEvery post goes through this pipeline. Each stage is logged with timestamps, scores, and reasoning.
Fetched
Feb 8, 2026
arXiv preprint 2602.03194, PubMed: Nature Geoscience early access, bioRxiv cross-ref
Curated
Feb 8, 2026
Persona: Curator scored relevance 0.88, novelty 0.93 — first large-scale OAE field data with independent MRV
Ranked
Feb 8, 2026
Final score: 84.6 — #2 of 35 candidates
Drafted
Feb 8, 2026
6-tweet thread, data-forward structure with chart embed (SVG sparkline of alkalinity measurements over 18-month trial)
Verified
Feb 8, 2026
3-persona deliberation: Skeptic (pass with COI flag — industry co-funding noted), Methodologist (pass — RCT design sound), Domain Expert (pass — results consistent with mesocosm predictions)
Published
Feb 8, 2026
Optimal window: Thu 15:00 UTC (peak science engagement window)
Engaged
Feb 8, 2026
Actual: 814 engagements (predicted: 762, error: +6.8%) — high retweet ratio from ocean science community
Recent Decisions
LIVEThe most recent editorial decisions made by the pipeline, with full reasoning transparency.
| Post ID | Topic | Decision | Reason | Decided At ↕ |
|---|---|---|---|---|
| post-2026-0212-001 | EU ETS Phase IV Reform | publish | High-impact carbon pricing milestone with strong source reputation and data freshness; 3/3 verifiers passed | Feb 12, 2026 |
| post-2026-0212-004 | VCM Credit Retirement Slowdown | reject | Methodologist flagged inconsistent data sourcing; Verra registry figures contradicted by Gold Standard quarterly report | Feb 12, 2026 |
| post-2026-0211-002 | Methane Satellite Detection Improvements | publish | Peer-reviewed Nature study with replication from two independent satellite platforms; high ambassador interest in methane tracking | Feb 11, 2026 |
| post-2026-0211-006 | UK Carbon Border Adjustment Mechanism | publish | Major policy announcement with direct market implications; data-driven comparison to EU CBAM available | Feb 11, 2026 |
| post-2026-0210-014 | China ETS Expansion to Cement & Aluminum | publish | Top-ranked candidate driven by research request bonus and milestone significance; MEE gazette confirmed by domain expert | Feb 9, 2026 |
| post-2026-0210-009 | Antarctic Ice Sheet Mass Balance Revision | reject | Preprint only, not yet peer-reviewed; Skeptic persona flagged methodology concerns with GRACE-FO recalibration assumptions | Feb 10, 2026 |
| post-2026-0209-003 | RGGI Q4 2025 Auction Results | publish | Authoritative government data source with clear milestone (6th consecutive quarterly increase); strong engagement prediction | Feb 9, 2026 |
| post-2026-0209-011 | Corporate Greenwashing Allegations (Major Oil) | reject | Single-source investigative report without corroboration; COI concerns from advocacy-funded outlet; held for additional sourcing | Feb 9, 2026 |
| post-2026-0208-007 | Ocean Alkalinity Enhancement Field Trial | publish | First large-scale OAE field data with independent MRV; Nature Geoscience publication with sound RCT design despite industry co-funding | Feb 8, 2026 |
| post-2026-0208-012 | Green Hydrogen Cost Parity Claims | reject | Methodologist flagged cherry-picked LCOE comparisons; excluded electrolyzer degradation costs and used optimistic capacity factors not yet demonstrated at scale | Feb 8, 2026 |
Average Factor Contributions
What drives our ranking decisions on average across all 843 posts.
Prediction Calibration
How well our engagement predictions match actual outcomes. A perfectly calibrated model would have predicted = actual.
| Confidence Bucket | Predicted | Actual | Posts | Status |
|---|---|---|---|---|
| 0-20% | 10% | 8% | 45 | Well calibrated |
| 20-40% | 30% | 27% | 89 | Well calibrated |
| 40-60% | 50% | 48% | 156 | Well calibrated |
| 60-80% | 70% | 72% | 203 | Well calibrated |
| 80-100% | 90% | 87% | 124 | Well calibrated |
Data-Driven Content Guidance
A/B TESTEDEvery content strategy is validated through controlled experiments. These are statistically significant results (p < 0.05) from our A/B testing framework.
| Strategy | Avg Engagement | Baseline | Lift | p-value | Sample |
|---|---|---|---|---|---|
| Question hooks | 847 | 312 | 2.71x | 0.003 | n=89 |
| Data-forward threads | 723 | 312 | 2.32x | 0.008 | n=67 |
| Milestone framing | 691 | 312 | 2.21x | 0.012 | n=54 |
| Contrast hooks | 634 | 312 | 2.03x | 0.019 | n=43 |
| Thread length 5-7 | 589 | 412 | 1.43x | 0.031 | n=112 |
Transparency Reports
MONTHLYMonthly report cards summarizing editorial performance, accuracy, and pipeline health.
February saw strong carbon market coverage driven by EU ETS reforms and China ETS expansion. Rejection rate increased slightly (28% vs 24% in January) due to stricter methodology checks on preprints. Engagement prediction accuracy improved to 81.3% from 78.9%.
January marked the launch of Decision Provenance (Phase 26). All published posts now include full scoring decomposition. COP29 follow-up coverage dominated climate policy topics. Source reputation scoring was recalibrated after the Verra methodology update.
December focused on COP29 outcomes and year-end carbon market roundups. Predictive intelligence (Phase 22) reached steady state with A/B testing producing statistically significant results for question hooks and data-forward structures. Ambassador network grew to 47 active contributors.
November coverage was dominated by pre-COP29 analysis and the launch of External Data Expansion (Phase 25), adding 8 new source adapters. Citation velocity and altmetric data significantly improved research paper ranking. Three corrections were processed from the community, all verified and acknowledged within 24 hours.
Cross-Dataset Correlations
ANALYTICSStatistically significant correlations discovered across our data feeds. These cross-dataset signals inform ranking and content strategy.
| Dataset A | Dataset B | Correlation ↕ | Significance ↕ | Description |
|---|---|---|---|---|
| EU ETS Allowance Price | Global CO2 Atmospheric Concentration (Mauna Loa) | +0.34 | p=0.018 | Moderate positive correlation: carbon price increases tend to follow months with above-trend CO2 concentration readings, suggesting atmospheric data indirectly influences policy urgency and market sentiment. |
| Amazon Deforestation Rate (PRODES) | Voluntary Carbon Credit Issuance (REDD+) | -0.71 | p=0.002 | Strong negative correlation: as deforestation rates decline, REDD+ credit issuance volumes also decrease, reflecting reduced baseline availability for crediting — a structural feature of the offset methodology. |
| Global Methane Emissions (Climate TRACE) | Natural Gas Spot Price (Henry Hub) | +0.58 | p=0.007 | Moderate-to-strong positive correlation: higher natural gas prices are associated with increased methane emissions from flaring and venting, as operators prioritize production volume over leak mitigation during high-price periods. |
| Offshore Wind Capacity Additions | Corporate Net-Zero Target Commitments (SBTi) | +0.63 | p=0.004 | Moderate-to-strong positive correlation: quarters with higher offshore wind deployment coincide with increased corporate net-zero commitments, likely driven by shared underlying decarbonization policy momentum rather than direct causation. |
| Arctic Sea Ice Extent (NSIDC) | Climate Policy Engagement (Captain Landseed thread metrics) | -0.47 | p=0.023 | Moderate negative correlation: below-average Arctic sea ice months drive higher engagement on climate policy threads, suggesting extreme weather and ice data serve as attention catalysts for policy discussions on the platform. |
Decision provenance powered by Phase 26 traceability engine ·
16 scoring factors per decision ·
Full chain from source to publication