Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
risk measures vs. performance metrics

Just a high level thought about the layout and content of the results that users are supplied with once an algorithm has completed.

Sharpe, Sortino etc. are more traditionally thought of as ways to describe the performance of an investment strategy (with some measure of 'volatility' masquerading as 'risk', being an input to their calculation) and might therefore be better placed under a heading 'Performance Metrics' in the results section. Volatility, Max drawdown etc. would rightly stay under the 'Risk Metrics' heading.

What do others think? And what other metrics would people like to see? %Winning trades? CAGR?

2 responses

%winning trades is interesting because ?typically an algorithm will have a large percentage of trades that are not successful or just breaking even on average, and then a smaller percentage of trades which are the big winners where the theory has transformed into real cash. So it would be nice to get a better summary of the winning trades (perhaps they can be ranked by the amount gained). Snapshot information would also be nice: detailed graphs for where each trade was placed, preferably with the ability to label data points in the graph from the code.

https://www.quantopian.com/posts/run-summary provides some individual stock info at the end of the run, just that time has gone by, needs Q2 updates if I remember right, it was never perfect, needs provision for shorting added for example.