Research method
How the Turwin 500 was researched and ranked.
The Turwin 500 is an editorial ranking with a fixed research cut, visible sources, and rules you can inspect. The number records our judgment about an idea's force and usefulness.
Cutoff / 2026-07-10
20 fields × 25 ideas
0 post-cutoff evidence
Eligibility before score
An idea has to qualify before we score it.
An idea cannot enter Issue 001 with a post-cutoff event, duplicate mechanism, unresolved reader flag, unsupported observed claim, invalid taxonomy, or fabricated or contradictory source.
Only eligible ideas receive a score. Twenty-five are selected within each editorial field; the chosen 500 are then ranked globally.
Numbers 001–500 form the global order. The PDF groups the same ideas by field, while the web index reads straight through from 001 to 500.
The seven dimensions
Seven dimensions make up the score.
- 0120%
Evidence strength
Unsupported claims are ineligible; reproduced or authoritative support earns the highest score.
- 0215%
Distinctiveness
Mechanism duplicates are removed before scoring. New framing must change the work, not just the wording.
- 0315%
Leverage
The idea should change a stack, cost curve, workflow, or governing capability.
- 0415%
Demonstrated pull
Investment, deployments, experiments, and repeated demand count when the source actually establishes them.
- 0515%
Downstream consequence
The strongest ideas alter more than one product surface, institution, or research path.
- 0610%
Trajectory + timing
Movement is judged against the fixed July 10, 2026 cutoff—not the present moment after publication.
- 0710%
Editorial utility
An idea should open a buildable surface, a testable question, or a legible institutional choice.
How we label the evidence
Three labels tell you what kind of claim you are reading.
The signal is visible now.
The source directly supports the present-state claim.
This is our read of what the evidence means.
The observed premises are sourced and the editorial step stays visible.
This is a forward-looking judgment.
The premises are sourced and the outcome is tied to a stated time horizon.
What the rank means
Editorial order, with the limits stated.
- Ranks express editorial force: evidence, leverage, pull, consequence, timing, and usefulness.
- The intentionally banded rubric leaves 343 entries at the stable candidate-ID fallback after the published tie chain.
- Confidence is evidence posture: high for direct, authoritative support; medium when premises are solid but the mechanism is inferred; low for sparse, emerging, or long-horizon evidence.
- Declared penalties are medium confidence −3, low confidence −8, section source-family concentration −3, and company-directory framing −5. Any score below 45 is ineligible.
- Company-reported traction remains directional unless independently corroborated.
- Every projection is a wager, not a promise.
- Two convergence passes were rejected: 36 overlap clusters were adjudicated, templated copy was rebuilt, and 27 placements were reflowed without altering the hash-bound research corpus.
- Issue 001 remains frozen so later issues can be compared rather than silently rewritten.
Keep the report
Download the print-ready PDF.
Enter your email to get the report and future Turwin Labs field notes. The complete index stays open on the web.