Risk Modelling Flashcards
(13 cards)
Goal of a risk model
Use probabilities to predict possible future outcomes.
What is the state-space (Ω)?
The complete list of scenarios we consider.
What is a σ-algebra (E)?
The collection of events we can assign probabilities to.
What is a probability measure (P)?
Rules that give each event a probability from 0 to 1.
What is discretization in risk models?
Breaking the model into a finite set of states for computation.
What is the Linear Congruential Generator (LCG)?
A simple formula that makes numbers that look random.
What is the Inverse-CDF method?
Converting uniform random numbers into samples of any distribution.
State the Strong Law of Large Numbers
As you average more samples, the average approaches the true mean.
Basic Monte Carlo algorithm to estimate E[X]
Generate many X samples and average them.
How to simulate path-dependent variables in Monte Carlo?
Build each path step by step and record the final outcome.
What is model risk?
Risk of errors if the model’s assumptions are wrong.
Name a control measure for model risk
Backtest: compare model results with real data.
Why ensure discretization convergence?
Verify results stay the same when using more states.
