Published Screens - Return vs Volatility Chart
About this page:
- This page illustrates the return / volatility relationship among Published Screens.
- Each dot represents a separate backtest.
- A model can be selected (clicked) to view details.
- Zooming is accomplished by selecting a region on the chart. Un-zoom by using the right mouse button. Sorry, but touch screens are aren't fully suported with this graph.
- All backtests are not current, although most of those with competitive risk:return profiles are relatively current. Stats can be updated by continuing on to the backtest page where the model will be re-run.
- Star ratings are defined by our user-base in one of two methods. The most direct method is to go to the respective backtest page and click the star rating image. Additionally, selecting a screen as a favorite
implicitly rates it a 5-star, and removing it implicitly gives the screen a 1-star rating. The net, obviously, would be a 3-star rating for being considered for a favorite. Direct rating by a user is limited to one vote per month,
and a user must be logged in to vote.
- Why are most models un-rated? Models are only rated by one of the above user actions. If no action is taken, the model will have no rating.
- Should all models be rated? Not really, but those with competitive risk:return profiles probably deserve some rating.
- What should be considered in rating a screen? Return and volatility are objective measures. The star rating should consider more subjective matters such as:
- How smooth is the equity curve?
- Does the model appear to be overly tuned to the historical data?
- In your judgement, will it continue to perform in the future?
- Bottom line, would you trust it?
|
|
