This article covers:
- Creating models
- Back testing models (Single Period and Deep Analytics)
- Comparing Models
- Understanding Model Performance
Related articles:
- Setting Target Return and Target Volatility for models
- Setting up model set to be used in the Risk Tolerance Test and portfolio visualizations.
Creating Models
The Andes Risk platform supports strategic models, tactical models, and model of models. Click on the respective link below.
- Creating strategic models
- Creating tactical models
- Creating Model of models. We call it Strategies in Andes platform for the lack of a better word.
Adding New Security When Entering Models and Portfolios Manually
Click here for instruction.
Back Testing Models (Single Period and Deep Analytics)
Calculate Single Period
To back test models, use "Calculate Time Period" to see how the model performed for the longest possible range or for a given time period.
The historical return and volatility are one of the data points to help you set the target return and volatility. For details, see Setting Target Return and Target Volatility for models.
Calculating Deep Analytics
For a deeper analysis, click on Calculate Deep Analytics at the bottom of the Model Positions tab.
To view the deep analytics, use the analytics tab of the model page and the Model Market Place.
Here are the analytics tabs:
The deep analytics offer additional data points to help you set the target return and volatility. For details, see Setting Target Return and Target Volatility for models.
Comparing Models
You can use Model Comparison to compare your models. To Access the Model Comparison, select Model Management -> Model Comparison from the top right menu.
The classic models are useful reference points for comparison. Change the time period and end date to specify any time period for comparison.
If your models are comparable with the classic models, you are doing fine.
If your models consistently do better than the classic models (few can do it), you can use it with clients.
If your models are below the classic models, use it internally first. Often times it is the market trends going against the factors, industries or regions that you overweight.
Understanding Model Performance
You can use Model Portfolio Attribution to find out which securities or funds drive the risk and return for any period of time.
To access Model Portfolio Attribution, select Model Management -> Model Portfolio Attribution from the top right menu.
If you specify a model (see the orange box), it shows all the stocks/ETFs/mutual funds in the model. You can specify multiple models.
If you create a fake model called "Index Model", you can add stocks and ETFs of interest to it, so you can compare them against the securities you already have in your models.
Use the brown box to specify any time period and specific end date. For any model/portfolio, we calculate the risk/return for any period of time on a daily basis.
