Keeping score in public, misses included.
A rating you cannot check is not diligence. So we write down the calls that can be settled, say what would settle them, and record how they turn out. The wrong ones stay on the page.
A claim with a date attached.
Some claims settle in time. “Direct air capture will cost between $600 and $1,000 a tonne by 2030” is one. When a claim like that enters a report, it also enters the ledger: the figure, who made it, and the date the world will tell us whether it held.
An honest range must never score worse than a lucky narrow guess. A source that says “between X and Y, I am not sure” and turns out right is not marked down against one that named a single number and happened to land on it. Calibration is rewarded. False precision is not.
Four outcomes, set on the date.
- CorrectThe stated figure or range held.
- Directionally correctThe direction was right; the magnitude was off.
- IncorrectReality fell outside the claim.
- UnresolvableNo adjudicating evidence by the resolve date.
Dated calls, and when each one resolves.
| Ref | Claim | Attributed to | Resolves by | Outcome |
|---|---|---|---|---|
| RL1 | Technical mitigation potential is up to ~5.3 GtCO2/yr by 2030, but saturation and reversibility mean the realistic figure is far lower. | IPCC (AR6, synthesised) | 2030-12-31 | Pending |
| RL2 | A realistic cost for direct air capture by 2030 sits between $600 and $1,000 per tonne of CO2. | Howard Herzog | 2030-12-31 | Pending |
| RL3 | DAC cost by 2030 can reach roughly $300–425 per tonne, based on developer analysis of liquid-solvent plants at scale. | David Keith (work commercialised via Carbon Engineering / Occidental) | 2030-12-31 | Pending |
| RL4 | Component-based learning curves put DAC cost at $230–540 per tonne by 2050 — falling, but well above $100. | Katrin Sievert, Bjarne Steffen, Tobias Schmidt | 2050-12-31 | Pending |
| RL5 | Stratos is a $1.3bn plant designed for 500,000 t/yr at a reported ~$400–500/t, backed by a $550m BlackRock JV, ramping to full capacity by mid-2026. | Occidental Petroleum / 1PointFive (reported via Axios, freethink) | 2026-06-30 | Pending |
| RL6 | Durable carbon-removal demand is projected at 40–200 Mt/yr by 2030 ($10–40bn) and 80–870 Mt/yr by 2040 ($20–135bn). | Frontier (Stripe, Alphabet, Shopify, Meta, McKinsey) | 2040-12-31 | Pending |
| RL7 | The IPCC treats large-scale carbon removal as unavoidable for net zero; DAC scenarios reach 5–15 Gt/yr by 2050. | IPCC (synthesised via RMI, Nature Communications) | 2050-12-31 | Pending |
Every row here is pending — these calls resolve between 2026 and 2050. The point is not the score today. It is that the score exists, is dated, and will not be quietly edited when a number goes the wrong way. A flag that moves in our favour while the field has not moved is something you will be able to see.
The ledger is the other half of the methodology. One publishes how we work; this one publishes how the work turned out.