Composite Score
Each module returns a score in [-1, +1]. Weighted blend gives the composite score. Higher absolute values mean stronger directional conviction.
Education Center
A practical guide to the exact mechanics inside Q Signals. Use this before changing thresholds, horizons, or risk overlays.
Module Inputs (What Each Signal Consumes)
This table maps each module cluster to its primary input, so users know exactly what is driving each score.
| Module | Primary Input |
|---|---|
| Fibonacci | Retracement and extension structure from recent price swings |
| Elliott Wave | Wave-pattern structure and trend phase |
| Swing Strategy | SMA crossover, RSI, Bollinger Band context |
| Sentiment / Social | News and social tone direction |
| Insider / Institutional / Congress | Ownership and disclosed transaction behavior |
| Supply Chain / Competitor | Peer stress and dependency signals |
| Volume + Multi-Timeframe | Volume confirmation and cross-timeframe trend alignment |
| Sector / Correlation / VIX Regime | Macro regime and cross-asset flow context |
| Economic / Analyst | Macro indicator direction and analyst consensus |
| Price Trend / Volatility / 52w / Moving Avg | Core technical direction and overextension |
| RL Agent (DQN) | Learned BUY/HOLD/SELL Q-values from historical outcomes |
Agreement Gate
Even if score crosses threshold, direction can be downgraded if bullish/bearish vote share is too split. This reduces one-factor false positives.
Confidence
Confidence grows as score exceeds threshold and module agreement strengthens. Low confidence directional signals should be sized smaller.
Horizon Logic
Short/Mid/Long horizons change ATR target distance, ATR stop width, and hold-cap windows. Same ticker can produce different plans by horizon.
Targets and Stop
T1/T2/T3 and stop-loss are ATR-driven and refined by structure checks. Moderate target is usually the base for risk/reward evaluation.
Regime Stop Multiplier
Stop width adapts to volatility regime: tighter in calm conditions, wider in high/extreme volatility to reduce noise stopouts.
Reward-to-Risk Floor
If R:R for a directional setup is below your configured floor, action is downgraded to HOLD. This filters weak setups early.
Time Stop
Backtests can cap evaluation by hold-days target. If thesis does not resolve by that window, the system exits instead of waiting indefinitely.
Universe Portfolio Overlay
Universe backtests now report inverse-vol weights, correlation cap impact, and volatility-targeted leverage estimate at the portfolio level.
News + Social Trend Pulse
Single Ticker page includes a lightweight trend card that loads separately, so context still appears when full analysis is slower.
Provider Failover
Market data can route across Yahoo, TwelveData, and Alpha Vantage where keys are configured. News can fall back to Finnhub.
What This Is Not
This is a research decision system, not financial advice. Always use position sizing, portfolio limits, and event-risk awareness.
Detailed Learning: Each Module Explained
Dive deeper into each analysis module. Use this reference before adjusting module weights or relying heavily on a particular signal path.
Fibonacci Retracement
Fibonacci retracement levels are a powerful tool that helps identify potential support and resistance zones based on the Fibonacci sequence. The key levels are 0%, 23.6%, 38.2%, 50%, 61.8%, and 100%. These levels are plotted between significant highs and lows to mark zones where price may find support or face resistance during pullbacks.
The theory behind Fibonacci retracements is rooted in the golden ratio (0.618), which appears frequently in nature and financial markets. When a strong trend reverses, prices typically retrace between 38.2% and 61.8% of the prior move before continuing. Q Signals plots these levels on recent swing highs and lows, then scores bullish when price finds support at a level during a downtrend, and bearish when price faces resistance during an uptrend.
Treat Fibonacci levels as probability zones, not exact predicted prices. Combine them with other indicators like RSI, volume, and moving averages to confirm retracement signals. Higher conviction occurs when multiple indicators align at the same Fibonacci level.
Elliott Wave
Elliott Wave theory describes market movement in recurring 5-3 wave patterns. An impulse wave consists of five waves: three in the direction of the trend (1, 3, 5) and two against it (2, 4). After the impulse, a corrective wave unfolds in three waves (A, B, C) against the prior trend.
Q Signals identifies Elliott Wave structures using price action and Fibonacci relationships. Wave 3 is typically the longest and most powerful. Wave 1 breakouts signal the start of a potential impulse; wave 3 extensions confirm trend strength; wave 5 represents exhaustion if it is shorter than wave 3. Corrections to wave 2, 4, or the entire 5-wave sequence often complete at Fibonacci levels (38.2%, 61.8%).
Elliott Wave is most effective on timeframes with clear trending structure (daily and weekly). Intra-day waves are noisier and harder to validate. Always confirm wave counts with other indicators and market context before acting on them.
Swing Strategy
The swing strategy module combines Simple Moving Average (SMA) crossovers, RSI (Relative Strength Index), and Bollinger Bands to identify short-to-medium-term directional setups. A bullish swing setup occurs when price crosses above the SMA, RSI is above 50, and price is near or above the lower Bollinger Band, signaling potential continuation momentum.
Bearish swings align when price is below the SMA, RSI is below 50, and price is near the upper Bollinger Band, suggesting resistance. The Bollinger Bands (20-period SMA ±2 standard deviations) mark overbought/oversold extremes; price tends to mean-revert when it touches the outer bands.
Swing signals are strongest when all three indicators agree. False signals increase when RSI diverges from price (price makes a higher high but RSI makes a lower high), indicating fading momentum. Use the swing module as a confirmation layer, not a standalone timing tool.
Sentiment & Social Media
Sentiment analysis combines news headline NLP processing and social media mention tracking from Reddit and StockTwits. News sentiment detects hawkish/dovish language (bullish keywords like "surge," "breakthrough"; bearish keywords like "collapse," "risk"). Social sentiment measures post count velocity, bullish/bearish tone split, and mention frequency spikes.
High positive sentiment can indicate strong retail enthusiasm, which often precedes pullbacks when momentum traders exit. Conversely, high negative sentiment during long-term trends can represent capitulation—a potential reversal signal. The key is understanding context: sentiment is most powerful as a contrarian fade during extremes, not as a directional follower.
Be aware that sentiment lags real data and can be manipulated by coordinated social groups. Always cross-check sentiment signals with technical breakdowns or fundamental catalysts. Use sentiment to confirm, not initiate.
Insider & Institutional Ownership
Insider trading (SEC Form 4) and institutional holdings changes (13F filings) reveal smart-money positioning. Insider purchases—especially by executives—signal confidence and often precede price appreciation. Insider sales can be tax-driven but sustained heavy sales may indicate insiders are less bullish on long-term prospects.
Institutional ownership trends matter even more. Large funds adding positions across multiple filings suggest building case; exits could presage downturns. Congressional STOCK Act trades track elected officials' holdings, which sometimes lead retail awareness of regulatory shifts or sector tailwinds.
Insider and institutional signals are slow-moving and best used for longer-term thesis confirmation (weeks to months), not short-term entry/exit timing. An insider buy is strongest when combined with technical breakouts or earnings catalysts.
Supply Chain & Competitor Strength
Supply chain analysis monitors SEC filing relationships to identify companies dependent on stressed suppliers or facing disruptive competitors. A company that relies heavily on a collapsing supplier risks production delays. A competitor gaining market share threatens pricing power and margins.
Competitor relative-strength scoring compares a stock's momentum vs. its peer group. A leader breaking out while peers lag suggests structural advantage. A laggard amid peer strength suggests sector headwinds are picking specific losers. Sector rotation also matters: a strong competitor in a deteriorating sector can still underperform.
Supply chain and competitor signals are best used alongside fundamental catalysts (M&A, new partnerships, quarterly guidance). They identify structural risks or tailwinds, not immediate price moves.
Volume & Multi-Timeframe Analysis
Volume profile analysis examines order flow and on-balance volume (OBV). High volume during uptrends suggests conviction; high volume on downtrends indicates liquidation. OBV accumulation (volume rising on up days, falling on down days) signals insider strength; distribution (volume rising on down days) warns of institutional unloading.
Multi-timeframe analysis ensures daily signals align with weekly trends. A daily breakout is stronger if the weekly chart also shows an established uptrend. A daily bounce in a weekly downtrend is likely a bear-trap. Horizon-based alignment reduces false signals caused by short-term noise conflicting with longer-term structure.
Volume confirms price moves. Rising volume on breakouts suggests institutional participation; declining volume on reversals suggests shorts covering rather than new demand. Always examine volume as a check on price action strength.
Sector Rotation & VIX Regime
Sector rotation tracks momentum shifts across 15 market sectors (Tech, Healthcare, Energy, etc.). During risk-on periods (low VIX, strong GDP), growth sectors (Tech, Semiconductors) lead. During risk-off periods (high VIX, recession concerns), defensive sectors (Utilities, Staples) outperform. A stock breaking out into a trailing sector risks underperformance.
VIX regime classification splits market conditions into four states: Calm (VIX < 12), Elevated (12–20), High (20–30), Extreme (> 30). Calm regimes favor growth; extreme regimes favor defense. Stop-loss width scales with VIX regime to reduce noise-driven exits in extreme volatility while tightening stops in calm conditions.
Sector and VIX signals provide macro context. A strong technical signal in a lagging sector may still struggle. Align individual stock signals with sector and regimes for higher-conviction setups.
Economic Indicators & Fed Policy
Economic indicators (CPI, unemployment, yield curve) and Federal Reserve minutes sentiment drive long-term trend direction. A steepening yield curve (long rates rising faster than short rates) signals economic growth; an inverting curve warns of recession. Strong employment data supports growth narratives; rising jobless claims trigger flight-to-safety demand.
Fed hawks (raising rates, tightening) weigh on growth stocks; Fed doves (cutting rates, stimulating) support equity bulls. Fed minutes language reveals policy trajectory months in advance. Markets often rally in anticipation of rate cuts and fall ahead of rate hikes.
Economic and Fed signals work best on longer timeframes (weeks to months). They establish the regime; technical signals and volume confirm entries within that regime. A bullish technical setup during Fed tightening is riskier than one during easing.
Analyst Ratings & Earnings
Wall Street consensus ratings (Buy, Hold, Sell) and price targets reflect collective professional opinion. A stock with 80% Buy ratings faces higher odds of support on dips; 80% Sell suggests shorts are crowded, risking short squeezes. Price target upside (target price minus current price) reveals expected return potential within analyst forecast windows (usually 12 months).
Analyst recommendation changes (upgrades/downgrades) often trigger sharp moves. Earnings dates are critical inflection points: surprises (beats/misses) can override technical technicals, so avoid holding through earnings if your setup is purely technical without fundamental conviction.
Analyst signals lag retail awareness and can represent consensus that is already priced in. Look for outliers (contrarian analysts) and changes in recommendation direction rather than absolute ratings. Use analyst buy ratings to confirm, not initiate.
Price Trend, Volatility & 52-Week Range
Price trend analysis checks if price is above or below key moving averages (50-day, 200-day). An uptrend is confirmed when price is above the 200-day SMA; downtrends below. Volatility (ATR—Average True Range) marks the stock's true daily price swing distance and is used to compute targets, stops, and hold duration.
52-week range (52w high/low) marks yearly extremes. Price near 52w highs implies strength and reduced downside risk (if on rising volume); price near 52w lows signals weakness or opportunity (if fundamental catalysts suggest recovery). Mean-reversion bias is strongest when price reaches 52w extremes without fundamental deterioration.
These technical metrics are foundational. High volatility stocks need wider stops and larger average targets to achieve attractive risk/reward. Trending markets (price well away from 52w range) suit directional strategies; range-bound markets (price in middle of 52w range) favor mean reversion.
Reinforcement Learning (DQN) Agent
Q Signals includes a Deep Q-Network (DQN) reinforcement learning agent trained on historical signal outcomes. The agent ingests 10 offline-calculated module scores (Fibonacci, Elliott, Swing, etc.) and outputs learned BUY/HOLD/SELL Q-values—essentially probability weights that this combination of signals is likely to succeed.
The agent trains continuously by tracking outcomes: did a BUY signal lead to price appreciation? Did a SELL signal predict decline? The agent's historical win rate feeds back into its decision weight. Early in training (Infant stage, <500 episodes), the agent is exploratory and less reliable. As it learns (Learning→Developing→Experienced, 500→100K+ episodes), its recommendations carry more weight in the composite score.
The RL agent is not a crystal ball. It systematizes patterns the 10-input combination has historically shown in real outcomes. In novel market conditions or after regime shifts, it must re-learn. Do not over-rely on the RL score alone; it is one of 23 votes in the full consensus system.