```mermaid
graph TD
    A[Add bid_history table] --> B[Add watch_count + estimates]
    B --> C[Create market_indices]
    C --> D[Add condition + year fields]
    D --> E[Build comparable matching]
    E --> F[Enrich with auction house data]
    F --> G[Add AI image analysis]
```

| Current Practice      | New Requirement                 | Why                       |
|-----------------------|---------------------------------|---------------------------|
| Scrape once per hour  | **Scrape every bid update**     | Capture velocity & timing |
| Save only current bid | **Save full bid history**       | Detect patterns & sniping |
| Ignore watchers       | **Track watch\_count**          | Predict competition       |
| Skip auction metadata | **Capture house estimates**     | Anchor valuations         |
| No historical data    | **Store sold prices**           | Train prediction models   |
| Basic text scraping   | **Parse condition/serial/year** | Enable comparables        |


```bazaar
Week 1-2: Foundation
Implement bid_history scraping (most critical)
Add watch_count, starting_bid, estimated_min/max fields
Calculate basic bid_velocity
Week 3-4: Valuation
Extract year_manufactured, manufacturer, condition_description
Create market_indices (manually or via external API)
Build comparable lot matching logic
Week 5-6: Intelligence Layer
Add auction house performance tracking
Implement undervaluation detection algorithm
Create price alert system
Week 7-8: Automation
Integrate image analysis API
Add economic indicator tracking
Refine ML-based price predictions
```