- Joined
- 5/22/19
- Messages
- 18
- Points
- 13
I'm currently taking a quantitative risk-management class in my MFE and learning about VaR, expected shortfall etc.
From all the papers we have read in class, it'll discuss a previous model, i.e. VaR and show why it doesn't work well to predict loss on past crashes. It will then say why this new idea is better at predicting loss. But then obviously a future paper shows why that model does not work well either
Is it all bogus, as in; it's going in circles. Or are the models actually getting better? Because to me it seems like a more complicated way of overfitting.
From all the papers we have read in class, it'll discuss a previous model, i.e. VaR and show why it doesn't work well to predict loss on past crashes. It will then say why this new idea is better at predicting loss. But then obviously a future paper shows why that model does not work well either
Is it all bogus, as in; it's going in circles. Or are the models actually getting better? Because to me it seems like a more complicated way of overfitting.