The idea
Most timing tools draw a flat box around price. The moment a market starts to lean, that flat box lies — price hugs one edge for ages and every “extreme” misfires. SpanCaster instead casts a tilted fan forward: it reads the recent slope of the highs and of the lows, projects every bar ahead along that slope, and forms a leaning upper/lower edge.
The position line is simply where the close sits inside that fan, 0–100. Crossing above 50 means price stepped into the upper half (lean up); crossing below 50 means the lower half (lean down). One line gives both timing and side — and a separate range gauge makes sure we only act when the range is fanning open rather than pinching shut.
The two building blocks
Projection Oscillator (position line)
Regression slopes of recent highs and lows build a forward-projected fan; the close's percentage position inside it crosses 50 to flip the participation side. Timing and direction in one line.
Gopalakrishnan Range Index (GAPO)
A tidy log-ratio of how wide the recent span is versus how many bars it took. Above its own average = the range is opening (moves have room); below = pinching shut. Regime gate only.
Signal at a glance
Signal rules
- 1Position line crosses ABOVE 50 → upper half of the fan → long side
- 2Position line crosses BELOW 50 → lower half of the fan → short side
- 3Range gauge (GAPO) ABOVE its 10-bar average → range is opening (regime gate)
- 4ATR activity gate → skip stone-cold, barely-moving markets
- 5One position at a time · max 10 windows per day · fixed 1% risk each
Parameters (defaults)
| Parameter | Value |
|---|---|
| Projection fan regression | 14 |
| Position-line midline | 50 |
| Range gauge window / average | 5 / 10-bar |
| ATR period | 14 |
| Activity gate | 0.4 × 50-bar ATR average |
| Stop-loss | 1.5 × ATR |
| Target | 1.8 × stop distance (RR ≈ 1.8) |
| Timeframe | H1 (M30–H4) |
| Simulated frequency | about 2.5 signals / day |
| Magic number | 20220817 |
Execution flow
Risk note
This page is a learning-and-sharing note, not investment advice, and it promises no returns. Markets carry pullback risk. Please backtest 1–2 years of H1 data in the strategy tester to verify trigger frequency, then test on a demo account with small size first.