Skip to content

Gallery

One-liner: Measured sentences grouped by phenomenon, including rejects. Each card states: what the sentence is, the three-window five KPIs, and the conclusion.

Trend

MA golden cross (BTC · 2h)

  • Sentence: go long when the close is above the 50-day line and just crossed above the 200-day; exit on break of the 50-day.
  • Three-window five KPIs: bear +3.7% / bull +3.0% / recent −3.1%.
  • Conclusion: recent-window CAGR is negative; against the falsification line, this sentence does not hold in the recent window.

Funding

Funding fade

  • Sentence: fade extreme funding rates.
  • Three-window five KPIs: bear +2.7% / bull +3.0% / recent −3.4%.
  • Conclusion: recent-window CAGR is negative; this sentence does not hold in the recent window.

Calendar

Monday rebound

  • Sentence: buy after Monday open, sell before Friday close.
  • Three-window five KPIs: bear +0.61% / bull +0.46% / recent +0.06%.
  • Conclusion: all three windows are near zero; this sentence does not hold.

Cross-symbol

BTC leads AI alts

  • Sentence: when BTC leads, AI alts follow.
  • Three-window five KPIs: bear no sample / bull +2.81% / recent −1.52%.
  • Conclusion: recent-window CAGR is negative; this sentence does not hold in the recent window.

Fat-tail

P99 big-order chase

  • Sentence: chase P99 big orders.
  • Three-window five KPIs: bear +0.24% / bull −0.25% / recent +0.28%.
  • Conclusion: all three windows are near zero; this sentence does not hold.

Cross-section

CS momentum + amount

  • Sentence: score a basket daily by momentum + amount, buy the top N.
  • Conclusion: loses to the equal-weight benchmark in all three windows; reject.

Hot sectors

  • Sentence: buy the top N hot sectors.
  • Conclusion: loses to the equal-weight benchmark in all three windows; reject.

Tenbagger cohort

  • Sentence: buy the tenbagger cohort.
  • Conclusion: needs a measurable rewrite; reject.

Beta

US SPY / QQQ buy-and-hold

  • Sentence: buying the ETF misses the US bull; you must pick stocks or wait for a washout.
  • Local five KPIs: SPY CAGR 13.7% / MaxDD −34.1%; QQQ CAGR 20.4% / MaxDD −35.6%. Washout timing CAGR below same-window buy-and-hold.
  • Conclusion: class is beta. The popular sentence is hit. Declared reject.

AI chip-sales QoQ vs BTC regime

  • Sentence: when AI compute spend QoQ is high, BTC closed bars sit below MA200 more often (not next-few-days return, not the segment labels).
  • Phase 1: high-QoQ below MA200 13.4% vs low-QoQ 50.5% (−37pp). Same sign in bull and recent.
  • Conclusion: opposite of the paper claim. Not --declared. Walkthrough: docs/examples/20260914_ai_chip_spend_btc_regime.en.md.

Fine print