idea
A 20-day high, only if the target is twice the stop
A close through the 20-day high counts only when the target is at least twice the stop.
The buy rule is plain. Today’s close has to clear the last 20-day high, and price should sit above the 200-day line. If the target is not at least twice the distance to the stop, I throw the name out.
Rule name: Setup name breakout_rr, minimum reward/risk min_rr = 2.0.
Entry idea
- Close breaks above the prior 20-day high (Donchian channel)
- Price remains above SMA200 (intermediate uptrend assumption)
- Optional quality gates: relative strength vs SPY, sector vs SPY, volume ratio, sector top-N
Stop and target
| Leg | Rule |
|---|---|
| Stop | 2× ATR(20) below entry (or structure-based tighter stop) |
| Target | Entry + min_rr × risk (2R when min_rr=2.0) |
Risk model: each trade risks risk_pct of equity (default 1% = 1R).
Gate variants (research matrix)
| Gates | Intent |
|---|---|
base |
Setup only |
rs_sector |
RS>SPY + sector>SPY (default paper) |
flow |
Volume + RS + sector top-5 |
full / full_qqq |
Heavier liquidity / breadth filters |
SIT_FLAT |
Explicit no-trade control |
Bear and sideways windows show rs_sector often beats heavier gates — not because more filters always help, but because weak regimes punish naive breakouts.
Operating cadence
- Timeframe: daily bars — no intraday chart required
- Scan: broad universe (~150 names), not top-10 mega only
- Loop: 30-minute paper cycles with optional random subsample for alert efficiency
Backtest basis on the liquid book (2019→now, min_rr=2.0): high stop count, fewer targets, payoff clustering around 1.8–2.2R when targets hit.