foxalgoGraveyard
Die meisten getesteten Ideen halten einer kritischen Überprüfung nicht stand – sei es aufgrund von Lookahead-Bias, verzerrter Schichtung oder einer Placebo-Kontrolle, die still und leise versagt. Wir veröffentlichen hier die Ideen, die gescheitert sind, anstatt sie einfach stillschweigend fallen zu lassen.
TF timeframe agreement DEAD
Behauptung: Cross-timeframe candle agreement predicts next-move direction.
Agreement rate came out at ~49% across 242k states — a coin flip. It predicts move size, not direction. Strong-looking hits were window-boundary artifacts (the TF candle sat inside the London/Asia session window).
Quelle: Research archive #01
Cross-sector / intermarket cues DEAD
Behauptung: Rates move stocks, USD moves metals/energy — cross-sector signals carry information.
Raw deltas of 15-17 bps shrank to 1-5 bps once stratified by each market's own position (pos930_atr). Daily-level intermarket analysis adds nothing beyond a market's own state. Only the peer-Asia-sweep pattern inside the same sector survives.
Quelle: Research archive #02
Event days (FOMC / CPI / NFP) DEAD
Behauptung: Scheduled macro events shift conditional touch/breakout probabilities.
Conditional probabilities did not move on event-flagged days in any tested pair.
Quelle: Research archive #03
VWAP-side pairs as new information DEAD (as a combo)
Behauptung: print8 & vwap_side930 combined carries information beyond either signal alone.
Raw delta looked like +50 bps versus +27 bps for pos930 alone, but the increment over stratified singles with shared weights was only ~+1.5 bps (52/54 draws positive, 0/54 above 5 bps). Lesson: stratified deltas must be subtracted using shared weights, or three different band-weight distributions produce a compositional artifact.
Quelle: Research archive #04
Sweep-behind-back setup FALSIFIED — lookahead
Behauptung: A liquidity sweep behind price plus room ahead gave +20.6 bps edge.
The +20.6 bps result used information not yet known at decision time (lookahead). A placebo control did not catch it — the key lesson was that a naive placebo does not reveal lookahead bias by itself.
Quelle: Research archive #05
Expansion as a directional signal FALSIFIED — lookahead
Behauptung: Range expansion predicts next-bar direction, +18.4 bps.
+18.4 bps was intrabar lookahead (the signal used the fill bar's own outcome). After the fix, the effect dropped to +0.2 bps. A vendor-direction agreement of 87.5% remains as a descriptive fact, not a tradeable edge.
Quelle: Research archive #06
HTF anti-persistence timing DEAD — lookahead
Behauptung: Anti-persistence against the higher timeframe improves selection timing.
The construction required future information to compute; retired as lookahead.
Quelle: Research archive #07
Calendar-day flag FALSIFIED
Behauptung: Day-of-week / calendar position explains part of the observed ratio.
The apparent calendar effect was volatility repricing, not a calendar effect.
Quelle: Research archive #08