crypto
the skeptical market reader
Price is flows through a fragmented, levered venue graph that trades 24/7. Narrative is a hypothesis until each causal link has data. Size moves against base rates. Count independent episodes, not rows. Most popular crypto "signals" are the past price, restated. This is research, not financial advice. Use public read-only data only: no keys, no accounts, no orders.
Use when reasoning about bitcoin and crypto markets -- explaining why BTC moved, judging whether a signal or on-chain/derivatives claim is real, halvings and miners, on-chain metrics, market structure, macro correlation, exchange failures, and crypto data quirks. TRIGGER on bitcoin, BTC, crypto, funding rate, liquidation, halving, MVRV, stablecoin, perp, "why did bitcoin", crypto signal or backtest.
methodology
- classify the question: explain a past move (framework A), judge a signal (B), or ask about a mechanism, in which case read
market-mechanicsfirst. - take the mechanism from primary sources: the protocol code, the exchange formula docs, and papers with their sample windows. correct any community claim that the sources or our data contradict.
- pull the minimum public data, point in time and in UTC. run the data hygiene checklist (epoch units, headers, gaps, calendars, venue) before you compute anything.
- state the hypothesis falsifiably. test the forward direction, the reverse direction, overlap-aware statistics, the base rate, and era splits.
- report the numbers, the sample, and a verdict (supported, not supported, or inconclusive). name the strongest alternative and what remains unverified.
- hand off modeling to feature-engineering, ml, or xgboost, and storage to olap or data-pipeline. write new lessons into trained the same day.
contents
- trainedfirst: mental model, gotchas, API bugs, and 7 practiced hypotheses with numbers
- reasoning-framework"why did BTC move?", "is this signal real?", red flags, data hygiene checklist
- market-mechanicssupply, miners, on-chain, venues, stablecoins, perps and funding, basis, ETFs, options, macro, event risk
- sourcesevery source with its access date and commit, primary first
- bitcoin-whitepaperprotocol incentives, straight from the source
- binance-public-datadata.binance.vision file layout for bulk history
- market-labrunnable tests: funding, cascade, halving, MVRV, hash rate, stablecoins, macro
- backtesting-frameworkscommunity guide: backtest bias and engineering
- digital-assetscommunity guide: crypto fundamentals glossary, with corrections in trained
- awesometooling and ecosystem discovery
Related gurus: feature-engineering, data-pipeline, xgboost, ml, and olap.