What this is
A fixed-weight basket of 15 assets across five sleeves — bedrock networks, exchange rails, AI & data infrastructure, one gold anchor, and a tiny new-generation slice — chosen by a published scoring model during maximum market fear.
How it's graded
Every day, each component is marked at its end-of-day price and the index level recomputed with the exact published formula — next to an identical-method Bitcoin benchmark. The ledger is append-only: yesterday's points are never edited.
Why it's public
Picks are cheap; a scored record is not. Anyone can copy a portfolio — nobody can fake 365 days of daily grading done by an automated referee. This page exists so you can watch the methodology earn (or lose) its credibility in real time.
The Index vs Bitcoin — Since Day One
Both lines start at 100 on August 14, 2026. Simulated marks, identical method, zero hindsight edits.
| # | Component | Sleeve | Weight | Baseline px | Latest px | Change | Contribution |
|---|---|---|---|---|---|---|---|
| Loading live component marks… | |||||||
Fixed weights between pre-registered rebalance events · end-of-day marks via CoinGecko · no trading costs modeled · simulated paper research only.
Six Criteria. One Formula. Zero Vibes.
Every component was rated 1–10 on the same six public questions before earning its sleeve weight. You can re-run the whole ranking yourself.
| Rank | Component | Weighted score | Stars | Verdict |
|---|
Five Sleeves, One Machine
Why “Utility”, Why “Fear”
- Bitcoin fell about −49% over the twelve months before publication — a full bear market. Most altcoins fell −60% to −85%.
- Fear & Greed read ~29 (Fear) around publication. Historically, the strongest multi-year entries started in exactly this zone.
- Exchange-rail tokens were the exception — TRX −8%, HYPE +20%, BNB −28% vs BTC −49% — because their networks collect real fees in every regime.
- AI tokens sat at maximum fear: TAO −49%, LINK −63%, RENDER −70%, FET −82%. The thesis didn't die; it got cheap.
- Tokenized gold was already in a bull market — PAXG +29% over that same year — which is why the index carries a 6% anchor of it.
A Machine, Not a Mood
- Rebalance review each quarter (Jan / Apr / Jul / Oct), or when any component drifts ±30% from target weight.
- Every rebalance is pre-registered: new weights and rationale publish BEFORE they take effect. No silent edits, ever.
- Fixed weights between events. Nothing is bought or sold — this is a simulation; only the numbers move.
- A component leaves only when its thesis breaks — never merely because the price dropped.
- The watchlist is re-checked at every review; candidates must qualify on the same six criteria to enter.
- All of it stays simulated. The index exists to be graded in public — not followed.
First scheduled review: October 2026 — diff and rationale will publish on this page before anything changes.
Every Component, Its Thesis, Its Risks
Expand any row. Theses are research summaries written at publication (Aug 14, 2026) — kept verbatim as the historical record, misses included.
Watchlist — Under Review, Not Included
Honest limits — on the record
- This record is young. Days prove discipline, not destiny — the referee publishes daily precisely so you never have to trust a claim.
- No trading costs are modeled between rebalances; real-world friction would reduce results.
- CoinGecko end-of-day marks can differ slightly from any given exchange's tape.
- A bear-market-start bias means the index began from beaten-down prices; regimes change and results will too.
- If a data source fails, the affected day is flagged stale rather than silently backfilled.