₿ BTC Brain — Research Library
This is the raw research library behind the site — the markdown files are served exactly as they are stored in the repo, no dressing-up. Eight reference chapters (compiled April 2026) cover Bitcoin's history, structure, and prediction methodology, and the post-mortems section includes our own falsification analysis: where the Brain's forecasts were wrong, why, and what we changed. Read the autopsies before the chapters if you want to know how seriously to take any of it.
Research chapters
- 01Macro History & Market Cycles
Complete 2009–2026 price timeline, all four halving cycles, tops/bottoms, and the diminishing-returns / cycle-compression evidence.
- 02Technical Chart Structure
Major chart patterns in BTC history, support/resistance behavior, moving averages, volume and volatility regimes.
- 03On-Chain, Sentiment & Liquidity
Key on-chain metrics, holder and miner behavior, sentiment indicators, and liquidity dynamics.
- 04Macro Correlations & Institutional
BTC vs traditional assets, macroeconomic influences, institutional adoption timeline, black swans, and regulation.
- 05Trading Frameworks & Strategies
Cycle positioning, volatility regimes, DCA analysis, BTC-specific technical strategies, and risk management.
- 06Derivatives, Open Interest & Liquidations
Perpetual funding mechanics, futures basis and the CME basis trade, and liquidation-cascade intelligence.
- 07Lightning Network
How Lightning works, capacity growth 2018–2026, adoption milestones, and its value as a market signal.
- 08Advanced Prediction Framework
What actually predicts BTC price, multi-timeframe forecasting, confidence calibration, and invalidation logic.
Post-mortems & method
- Edge Autopsy — 2026-07-21
Falsification analysis of 1,505 resolved baseline forecasts: where the claimed edge did and did not survive scrutiny.
- Upgrade Master Plan — 2026-07-21
The seven-phase plan that followed the autopsy: find a real edge, lock it in, and prove it on the ledger.
- Promotion Decision — 2026-08-10
The documented go/no-go decision on promoting a candidate model out of shadow mode.
- Phase 3 Live Training Report — 2026-08-11
Walk-forward ML training on 17,753 hours of real history: NO CANDIDATE, verified by three independent adversarial audits — including the power analysis that bounds exactly what the negative result does and does not prove.
- BTC Intelligence Brain — master research corpus
The full consolidated research document the eight chapters were distilled from.