scripts/rerun_bilbo_gg_page.py conventions, 2000–2025 window. The corrected marquee cells reproduce the audited counts (bear Low+Falling 120/141, bull High+Rising 137/193). Extension rates dropped slightly vs the original (bear full-ATR 66% → 63.8%) but the continuation story holds.
The Bilbo Golden Gate study showed that momentum-confirmed triggers complete to 61.8% far more often. But 61.8% is just the first target. What are the odds of reaching 78.6%, full ATR, or beyond?
After price enters the Golden Gate at 38.2%, we tracked whether it reached each subsequent ATR level — 61.8%, 78.6%, 100% (full ATR), and the extension levels at 123.6%, 161.8%, and 200% — all within the same RTH session. We then split the results by the 1-hour Phase Oscillator state at the moment the 38.2% level was hit.
The bearish Bilbo carries the continuation. When the last completed 1-hour PO bar is Low + Falling at Golden Gate entry, the probability of reaching a full ATR move (63.8%) is nearly the baseline probability of just completing the Golden Gate to 61.8% (65.4%). Strong bearish momentum doesn't just complete the Golden Gate — it usually keeps going. (Pre-fix this read 66% vs 65%; the corrected join trims it slightly.)
Consider letting bearish Bilbo winners run. When the 1-hour PO confirms the Golden Gate direction, the 61.8% target leaves measurable edge on the table. For a bearish Bilbo the corrected odds of reaching the full ATR are 63.8% — and 46.8% for 123.6%. Consider trailing your stop rather than taking profit at the first target.
The mid-zone slope split is NOT a filter. The pre-fix version of this page showed counter-slope mid buckets as much weaker (bull Mid+Falling 52% to 61.8%). That was a look-ahead artifact: corrected, bull Mid+Falling completes 63.0% and reaches full ATR 28.9% — indistinguishable from (slightly above) Mid+Rising. Only the extreme-zone buckets separate from baseline.
The bearish edge is consistently stronger than bullish. Corrected, bearish Bilbo entries reach extension levels at roughly 2–3x the baseline rate (full ATR 63.8% vs 32.4%; 200% 11.3% vs 3.8%), while bullish runs about 1.3–1.9x (full ATR 34.7% vs 26.3%). Selling pressure carries further and more reliably than buying pressure. Extreme-zone samples are small (n=141–193 over 25 years) — expect noise in the deepest extension cells.