Example: Forex Strategy with Correlation Sweep

Scenario: Quantifying Sentiment's Predictive Power for EUR/USD

You trade EUR/USD and want to know whether news sentiment has measurable predictive value for this currency pair. Forex markets behave differently from crypto — they are more institutional, less retail-driven, and often respond to sentiment on longer time horizons.

Step 1: Run Correlation Sweep with Extended Lag

Go to SentiLab and configure a Correlation Sweep:

• Asset: EUR/USD
• Time Range: 90 days
• Min Quality: 0.8 (higher threshold — forex news is more institutional, so quality filtering matters more)
• Lag Range: ±72h (Extended) — forex markets move slower than crypto; the standard ±24h may miss the optimal lag

Step 2: Analyze Initial Results

The Correlation Sweep results show:

• Best lag: +24h
• Pearson r: 0.28
• FDR-confirmed: Yes
• Confidence: Suggestive

This tells you: sentiment today has a moderate (but confirmed) correlation with EUR/USD price 24 hours later. The signal is real but not strong. Can we improve it?

Step 3: Apply Predictive Only Filter

Enable the Predictive Only toggle and re-run the analysis. This filters the dataset to include only articles that the AI classified as forward-looking (forecasts, predictions, analysis) rather than backward-looking (event reports, summaries).

Updated results:

• Best lag: +24h (unchanged)
• Pearson r: 0.35 (improved from 0.28)
• FDR-confirmed: Yes

By filtering to forward-looking news only, the correlation improved from 0.28 to 0.35. This makes intuitive sense — backward-looking articles describe what already happened and add noise to the predictive signal.

Step 4: Daily Workflow

Based on these results, establish a daily routine:

1. Each evening, check EUR/USD sentiment from quality sources (minQuality 0.8+)
2. Focus on forward-looking content: analyst predictions, central bank commentary, economic forecasts
3. If sentiment is strongly directional (bullish or bearish with multiple sources agreeing), expect the EUR/USD price to follow by the next evening (+24h)

Step 5: Macro Context Integration

Forex markets are heavily influenced by macroeconomic events. Sentiment shifts around these dates carry significantly more weight:

• ECB meetings: Rate decisions, press conferences, forward guidance
• Fed meetings: FOMC decisions, dot plots, Chair press conferences
• Major economic releases: NFP (Non-Farm Payrolls), CPI (inflation), GDP

When a central bank meeting is scheduled, sentiment shifts in the 24–48 hours beforehand often reflect institutional positioning and informed speculation. These are the highest-signal periods for EUR/USD sentiment analysis.

Step 6: Narrative Clustering for Forex

Run Narrative Clustering for EUR/USD to check for media convergence patterns:

• Pre-ECB convergence: When many sources converge on the same ECB narrative (e.g., "ECB will cut rates"), the market has likely priced it in. A surprise (no cut) would cause outsized movement.
• Divergent narratives: When sources disagree about the direction (some bullish EUR, some bearish), uncertainty is high — expect increased volatility rather than a clear directional move.

Key Forex Insight

FX markets are institutional-dominated. Unlike crypto, where retail social media can move prices, forex is driven by central banks, hedge funds, and corporate treasury flows. This means:

• Sentiment from high-credibility sources (minCredibility 0.8) matters far more than volume of low-quality articles
• The quality of analysis matters more than quantity — one Reuters analysis piece may carry more predictive weight than 50 blog posts
• Lag times tend to be longer than crypto (24–72h vs 4–12h) because institutional decision-making is slower

Disclaimer

This is an educational example, not financial advice. Forex trading carries significant risk, including the possibility of losing more than your initial investment when using leverage. The correlation values shown are illustrative. Always conduct independent analysis and use appropriate risk management. See Risk Management and Sentiment Analysis for essential risk guidelines.

Why This Matters

This example demonstrates that sentiment analysis is not crypto-only — it applies to forex, commodities, and any asset class SentiSignal covers. The key is adapting your parameters: longer lags for slower markets, higher quality thresholds for institutional assets, and integration with macro context. Finally, read Risk Management and Sentiment Analysis to complete the course.