Signal Intelligence
11-Layer Analysis. Every 5 Minutes.
QARI evaluates every watched symbol across 11 independent analysis layers every 5 minutes. Each layer targets a different dimension of market behaviour. All 11 must converge before a signal is scored. Three must confirm before a trade is placed.
Confidence Score
35+ threshold required before entry is considered
The 11 Analysis Layers
Each layer is independently computed. No layer can override the others. They aggregate into a single confidence score that must exceed the entry threshold.
Multi-layer convergence
95+ features analysed every scan cycle
Wave Structure
Detects Break of Structure (BOS) and Change of Character (CHoCH) events across 15M, 4H, and 1D timeframes. Classifies market as BULLISH (HH/HL), BEARISH (LH/LL), or RANGE. Multi-timeframe alignment between daily and 4H sets the directional bias for all other layers.
Liquidity & SFP
Maps buy-side liquidity (swing highs, equal highs, round numbers) and sell-side liquidity (swing lows, equal lows). Detects Swing Failure Patterns, wicks beyond key levels that close back inside, and quality-scores each SFP on wick depth, reaction strength, and volume presence.
Order Blocks, Breakers & FVG
Identifies the last opposing candle before a strong impulse move confirmed by BOS. Tracks OB lifecycle: active, mitigated, negated, breaker. Grades Fair Value Gaps A/B/C. Grade A gets full confidence, Grade B requires SFP trigger, Grade C is filtered out entirely.
Volume & CVD
Analyses volume ratios, breakouts, climax events, and volume dry-ups. Cumulative Volume Delta (CVD) tracks aggressive buying vs selling with divergence detection. Volume Profile provides Point of Control, Value Area High/Low, and High/Low Volume Nodes. ATR regime gate: no entries if ATR(14) on 4H drops below 50% of the 30-day median.
Sentiment & Funding
Reads perpetual futures funding rates and open interest from exchange APIs. Negative funding on a bullish structure setup is a confluence signal, shorts are paying, and a squeeze could accelerate the move. Open interest collapse combined with a price drop is a primary ELE trigger.
Fibonacci
Computes Fibonacci retracements from the most recent significant swing on 4H and daily. The Golden Pocket (0.618–0.650 retracement) is a primary entry zone. Extensions at 1.272, 1.618, 2.000, and 2.618 serve as profit target candidates, the nearest structural or Fibonacci level becomes PT1, the next becomes PT2.
Indicators & RSI Divergence
Evaluates RSI divergence (bullish: price makes lower low while RSI makes higher low), EMA alignment (8/21/50/200 stack), and Bollinger Band squeeze detection. Indicators serve as confirmation layers, not primary signals, price action and structure come first.
Order Book Depth
Reads live exchange order book to detect large bid walls (support clusters), ask walls (resistance), and absorption patterns where large orders are being filled without significant price movement. Thin order book above current price is a plus for long setups.
VWAP
Tracks daily, weekly, and monthly VWAP levels. A bullish reclaim of daily VWAP, where price moves from below to above and holds, scores as a confluence signal. Positioned relative to an order block or Golden Pocket, a VWAP reclaim satisfies one of the three required confluence points.
Session Context
Weights setups based on market session: London open (high institutional activity, setups get full confidence bonus), New York open (volatility peak), Asia session (lower volume, setups scored more conservatively). Avoids flat markets between sessions where fake moves are common.
BTC Dominance
When Bitcoin dominance is rising (above 24h ago and trending up over 7 days), altcoin long scores are reduced by 15 points. Rising BTC.D historically signals capital rotating out of altcoins into Bitcoin. Squeeze setups are exempt from this penalty, they rely on crowded short conditions, not directional alt momentum.
How Signals Are Generated
From raw market data to executed trade, every step is systematic and auditable.
Scan
140+ pairs every 5 min
Score
11 layers aggregate
Confluence
3-of-5 signal types
R/R Gate
≥ 2.0 mandatory
Execute
Order placed
0
Signal Layers
0+
Features Analysed
0+
Min Score Threshold
Hard Gates: Confluence and R/R
Three-Point Confluence Rule
At least 3 of these 5 distinct signal types must be present at the same price level before a trade is placed:
- Order Block, Breaker Block, or Fair Value Gap (Grade A/B)
- Swing Failure Pattern (SFP)
- Golden Pocket Fibonacci retracement (0.618–0.650)
- VWAP reclaim (daily, weekly, or monthly)
- RSI divergence (bullish or bearish)
Meeting 2 of 5 is not enough. This is check #16 of the 16-check pre-trade gate.
R/R ≥ 2.0. Mandatory
Risk/Reward ratio is calculated before every entry. If R/R is below 2.0, the setup is skipped, no exceptions.
Why 2.0 minimum?
At R/R 2.0, you only need a 34% win rate to be profitable. Requiring a minimum 2.0 provides a mathematical edge even in lower-confidence market conditions.
How R/R is calculated
Entry price, stop-loss (below the OB or swing low that invalidates the setup), and PT1 (nearest structural or Fibonacci extension target) are all computed before order placement.
Collective Intelligence
Every trade across every QARI user contributes anonymised outcome data. The more the platform trades, the smarter the scoring becomes for everyone.
500
Adaptive weights activate
Layer weights adjust per market regime (trending, ranging, volatile) based on which signals predicted profitable outcomes most reliably.
2,000
LightGBM model activates
A binary classifier trained on anonymised trade feature vectors activates. Blended 50/50 with the weighted score for final entry decisions.
Premium
Personalised weights
Premium users receive personalised fine-tuning on top of the collective model, shaped by their own trade outcomes and chosen asset classes.
All training data is anonymised. No user IDs, no absolute prices, no personal data. Feature vectors and outcome labels only.
Signal Intelligence. FAQ
11 layers working for you, 24/7
Start your 3-month free trial. No credit card required.