Headline edges
- Trend-regime is the strongest single filter. Above 20-DMA: +0.26% mean C2C, 62% win, n=1056 (p<10⁻⁸). Below 20-DMA: −0.35% mean, 39% win, n=532.
- Big gap-up days (>0.5%) continue. 64% close above open, +0.32% mean (n=200, p<0.001). Big gap-downs continue too: 42% win, −0.33% mean.
- Strong-trend × DOW. Friday in StrongUp regime: 63% win, +0.20% O2C (n=158). Monday in StrongUp: 67% win, +0.165% (n=153).
- Long-side ORB has an edge; short-side does not. 15-min ORB long: 57.2% win, +0.05% per trade, n=797 (p<10⁻⁴). Short ORB returns ≈ zero — consistent with bull-biased period.
- First-hour direction predicts rest of day. When FH closes up, ROD mean +0.07%, win 53%. When FH closes down, ROD mean −0.09%, win 49%.
- Prior day high/low are magnetic. 91% of days touch either PDH or PDL; only 8.8% form a true inside-of-prior-range day; 10% reach both extremes.
- Asymmetric mean reversion at the tails. After single-day rallies >2%, next-day mean −0.69%, 41% win (n=56, p=0.04). After single-day drops >2%, mean +0.44%, 54% win (n=66, p=0.21 — directional but not significant). Big up-days reverse more reliably than big down-days in this sample.
Top 30 edges by quality score
Composite of statistical significance, sample size, and effect magnitude.
Mean O2C return by Day of Week
Mean C2C return by Month
Win rate by SMA regime
Mean O2C by Gap class
Open-to-Close return by DOW
Intraday-only edge by day of week.
Close-to-Close return by DOW
Full-day edge including overnight gap.
Overnight return by DOW
Overnight = prior close to today's open.
Range / abs(return) by DOW
DOW × Regime — intraday edge heatmap
Win rate by month
Mean C2C by month
Turn-of-Month effect
First 3 + last 3 trading days of each month vs middle of month.
O2C return by gap class
Gap fill rate
Probability of touching PDH / PDL by gap
Time of high/low by gap class
Mean C2C by SMA regime
Mean Next C2C by Combined Regime
Mean C2C by RV20 quartile
Open vs Prior-day Range → O2C
Streak length → next-day mean C2C
Negative streak = N down days, positive = N up days.
Day-after-move (decile)
3-day pattern → next day
Close strength bucket → next day
Opening Range Breakout — first-break strategy (long vs short)
Place buy-stop at OR high and sell-stop at OR low; first to trigger is the trade. Hold to EOD close.
OR class → rest-of-day return
Where OR closes within OR range (top third = bullish, bottom third = bearish, middle = neutral).
ORB strategy details
First hour direction → rest-of-day
First-hour magnitude (quintile) → ROD
Mean return by 30-min bucket (intraday)
Time of HOD vs LOD
Distribution of when daily high and daily low form.
FH × LH cross-tab
Midday return given first-hour direction
Post-midday return given midday direction
Prior-day-range level touch probability
If you split prior day's range PDL → 25% → mid → 75% → PDH, how often does today's intraday range hit each level?
Touch rates conditional on gap
Range expansion distribution (today vs prior)
Prior-range decile → today's range ratio
After tiny prior days, today is usually larger; after huge days, today contracts.
NR7 / WR7 / Inside / Outside → next day
Prior-range extension touch rates
Camarilla H3/L3/H4/L4 touches
Static % bands — % of days fully inside the band vs breached
Useful for sizing 0DTE strangles / iron condors at expiry. Only 4% of days fully stay inside ±0.3% of prior close; 18% inside ±0.5%; 53% inside ±1%; 74% inside ±1.5%. (Path-touch risk during the day is higher.)
Composite-score → next-day return
Score sums {gap dir, prior C2C dir, regime, close-strength, DOW heuristic} → range −5..+5. Reading: even with a 5-factor score, the day-after relationship is mildly mean-reverting.
DOW × Gap class — intraday O2C
DOW × prior-day close strength → C2C
Gap × Regime → O2C
DOW × prior-day direction → C2C
All edges — searchable
Filter by family, condition, or numerical range. Sortable. n ≥ 50 edges shown by default; click "Show all" to include small samples.