Advanced strategies
Capabilities beyond a single condition on a single instrument. Each is expressible in plain language; the difficulty is knowing they exist.
Higher-timeframe gates
Trade one timeframe, gate on another.
buy on the 5-minute RSI cross, but only when the daily trend is up
enter on the 15-minute breakout only if the 4-hour ADX is above 25
Every gate's condition must pass on its most recently closed higher-timeframe bar before a lower-timeframe entry is allowed to fill. Not the in-progress bar, the closed one.
That is the correct semantic and it is worth knowing: a daily gate evaluated on an in-progress daily bar would flip mid-session and cause entries the strategy would not have taken with complete information.
| Use it for | Why |
|---|---|
| Trend alignment | Take fast signals only in the direction of the slow trend |
| Regime filtering | Only trade when higher-timeframe volatility is in a usable range |
| Reducing false signals | Most of a fast strategy's losses are signals against the bigger picture |
Multiple gates are allowed; all must pass.
Multi-target ladders with risk actions
Scale out, and move the stop as you do.
exit 40% at 1R and move the stop to break-even,
exit 30% at 2R and move the stop to the first target,
let the rest trail 0.8% behind the peak
| Risk action | Effect when the rung hits |
|---|---|
move_sl_to_breakeven | Stop to entry |
move_sl_to_previous_tp | Stop to the previous rung's level |
none | Leave the stop |
Rung prices resolve at entry time from the entry price and the initial stop, so R-multiple rungs work with any stop type: ATR-based, structure-based or percent.
See Exit conditions.
Trailing mechanics
Four distinct ratcheting behaviours, and they are not interchangeable:
| Mechanism | Side | Fires when |
|---|---|---|
| Trailing stop | Protective | Price pulls back to the trailing protective line |
| Trailing take-profit | Profit | Price pulls back to the trailing profit line after arming |
| Break-even | Protective | Stop moves to entry once a threshold gain is reached |
| Lock profit | Protective | Stop jumps to a locked profit level at a threshold |
Trailing stop and trailing take-profit are mutually exclusive. One protects, one follows, and two ratchets on the same level would fight.
When several fire together, the exit engine resolves by most conservative (the tightest level) by default. See Exit conditions
Short and both-direction strategies
A long_only strategy cannot short whatever its exit says, direction and legs must agree, and
validation enforces it.
For a two-sided strategy, state both legs:
long when RSI crosses above 55, short when RSI crosses below 45
The builder can derive a short leg from a long one when the request implies symmetry, and it marks the leg as auto-derived.
A mirrored condition is not always the rule you would have written for the short side. Markets are not symmetric, volatility behaves differently in declines, and a stop distance that suits a long is often too tight for a short.
Pattern and market-structure strategies
Beyond indicators: computed chart and structure patterns.
buy on a break of structure to the upside after a higher low
enter on a bullish fair value gap fill
buy the cup and handle breakout with volume confirmation
short a double top confirmed on the second peak
Available: swing highs and lows, higher highs, lower lows, break of structure (both directions), bullish and bearish fair value gaps, double top, double bottom, head and shoulders and its inverse, bull and bear flags, rising and falling wedges, ascending and descending triangles, cup and handle, inside bar, plus the full TA-Lib candlestick set.
Pattern signals emit on confirmed pivots, not provisional ones. That is why a pattern condition does not flicker as a bar develops. And it is also why a pattern confirms later than the eye sees it.
See Indicators.
Relative-strength and reference-series strategies
Compare an instrument against a reference:
buy RELIANCE when it is outperforming NIFTY over 20 days and both are trending up
Otherwise the engine loads no reference series, those terms are NaN on every bar, and you get a silent zero-trade strategy. Validation catches it. See Validation system.
Alternative-data strategies
Gate a price rule on a fundamental, event, macro or on-chain metric:
buy the breakout only when the current ratio is above 1 and revenue is growing
avoid entries within 5 days of an earnings date
A quarterly fundamental gives about ten observations over the standard backtest window. A rule gated on it has been tested against ten states of that gate, which is not enough to say much about the gate. Useful as a filter; weak as evidence. See Indicators
Dynamic universe strategies
Trade a selected set rather than a fixed instrument, reselected each bar by a ranking metric. Genuinely different from a multi-symbol backtest. See Dynamic & universe strategies
Derivatives and multi-leg structures
Options and futures, including named multi-leg structures, straddles, strangles, vertical spreads, iron condors, butterflies, covered calls.
Build, backtest and paper only. Live entry is refused. See Derivatives.
Remote / external-signal strategies
Generate signals outside Stretus and send them in over an authenticated endpoint, then execute through the platform's risk and capital controls. Three ingress paths: push, pull, and a Telegram webhook. See Remote strategies
Chart-guided improvement (Target Moves)
Annotate historical moves on a chart and have the platform generate strategy hypotheses from them, using deterministic pattern matching. No model involved. It is the third mode of the Improve modal. See Improving a strategy
What is still not expressible
Honest boundaries:
| Not available | Nearest alternative |
|---|---|
| A cash-reserve percentage | Allocate less capital to the strategy |
| A build-time mark-to-market breaker in currency | Configure it as an execution gate |
| Forex strategies | Out of scope |
| Live derivative execution | Backtest and paper |
| Cross-strategy portfolio construction | Each strategy has its own isolated allocation |
| Intra-candle execution | Conditions evaluate at candle close |
| An "unless" / exception scope | Restate the rule positively |
| Timeframes outside 1m-1d | Not accepted |
See AI limitations.