Financial Models are derived from historical data:
- Asset Allocation
- Electronic Trading
- Pricing
These models can be **incorrectly trained or biased ** and hence, are back-tested to understand the expected performance / underlying risk
Back testing methods are insufficient as out of sample data is limited and does not cover all market scenarios
### Synthetic Data
- Machine generated but has similar characteristics to it's historical counter parts
- Helps conduct backtesting on unseen events