Loops HouseFalsifiable NEAR

NEAR HACKATHON
Links
Team
1 member- ZEOwner
Zecvilan
Overview
Problem: most crypto investment theses are written so they can't be proven wrong, so conviction gets judged instead of evidence.
Product: Falsifiable NEAR is a long-term investment thesis for NEAR, built to fail in public. The thesis: NEAR's value now depends less on selling blockspace and more on NEAR Intents, and the case holds only if captured Intents revenue grows faster than token dilution. It is split into six plain claims: Intents volume, captured revenue, base-layer usage, fees per user, buyback and burn vs issuance, and whether Intents activity leads the NEAR price. Each claim has a pass/fail threshold frozen at git tag thresholds-v1 (pinned by a SHA-256 hash test in CI) and is re-graded daily from public data, with no API keys needed. As of Nov 13, 3 of 6 claims hold. Failing claims stay on the page.
How NEAR fits in: all evidence is NEAR onchain and ecosystem data, centered on NEAR Intents volume and the revenue that funds NEAR buybacks. Raw-chain queries on the BigQuery NEAR public dataset are cross-checked against NearBlocks.
Honesty: the thresholds were set after seeing the Nov 12 snapshot, so that data is in-sample. Only gradings on data after the freeze count as out-of-sample, and backfilled history is labeled descriptive.
Try it: open https://falsifiable-near-sand.vercel.app (see the Verdict card and the Falsification scorecard), or run make data && make test in the repo.