CSE Historical Analysis & Event Prediction

Step 2B: Event Impact Prediction

Run:

python main.py predict-event "description of the event"

This:

1. Searches vector database for similar historical events
2. Matches against known event patterns
3. Retrieves relevant global and local news
4. Generates AI prediction using Claude with historical context

Step 3: Present Analysis

For Historical Review:

• Timeline of major events with ASPI impact
• Recovery patterns (V-shape, U-shape, L-shape recoveries)
• Sector-specific impacts (which sectors were hit hardest/recovered fastest)
• Key lessons from each event type

For Event Prediction:

• Most similar historical events (ranked by similarity)
• Expected ASPI impact (% range based on historical precedent)
• Expected recovery timeline (weeks/months based on patterns)
• Sector impact matrix (which sectors will be most affected)
• 3 scenarios: Best case, base case, worst case
• Recommended actions: Buy/sell/hold with specific sectors/stocks

Historical Events Database (Key Events)

Pattern Recognition Guidelines

1. War/Peace: CSE shows extreme reactions. Peace = massive rally (50-125%). War threat = 10-20% drop.
2. Terror attacks: Short-lived impact (2-8 weeks). Tourism/aviation hit hardest. Recovery is rapid.
3. Sovereign debt: Deepest and longest impact. Banking/finance most affected. Recovery takes years.
4. Pandemic: Initial sharp drop (20-30%), then gradual recovery. Pharma/telecom outperform.
5. Natural disasters: Localized impact (2-5%). Insurance/construction affected. Quick recovery.
6. Elections: Pre-election uncertainty (-3-5%), post-election rally if pro-business winner.
7. Global events: CSE has 0.3-0.5 correlation with global markets. Impact is delayed and muted.

Important Notes

• CSE is a frontier market — reactions can be amplified compared to developed markets
• Liquidity dries up during crises — consider this in position sizing
• Foreign investor flows are a key indicator (net selling = bearish signal)
• CBSL policy response (rate cuts/hikes) heavily influences recovery speed
• Always compare current valuations to pre-crisis levels for context

Examples

Example 1: Historical Review

User says: "How did CSE react during COVID?" Actions:

1. Retrieve COVID-19 event data from historical database
2. Show ASPI chart with pre/during/post COVID levels
3. Explain the -30% drop, government response, and 9-month recovery
4. Highlight which sectors recovered fastest (banking, telecom)

Example 2: Event Prediction

User says: "If there's a global recession, what happens to CSE?" Actions:

1. Search for similar events: 2008 GFC, 2020 COVID economic impact
2. Analyze patterns: CSE typically drops 15-30% during global recessions
3. Predict: Banking sector hit hardest, exporters may benefit from LKR weakness
4. Recommend: Defensive stocks (utilities, consumer staples)

Example 3: Specific Event

User says: "What if oil prices spike to $150?" Actions:

1. Search historical: 2011 commodity spike, 2022 post-Ukraine oil surge
2. Impact: Sri Lanka is oil importer — LKR weakens, inflation rises, CSE drops 5-15%
3. Winners: Energy stocks (if any). Losers: Transport, manufacturing
4. Recommend: Reduce exposure, hedge with export-oriented stocks