Requesting new stats: Correlation and M/E

The theory of statistics holds that doubling the number of independent members of an ensemble reduces the standard deviation by the square root of two. By doubling again, the standard deviation is thus halved. Therefore, trading four strategies in a portfolio will theoretically cut the deviation in half.

A great addition (in my opinion) would be to select a number of strategies to display a matrix of their correlation with each other as well as a benchmark such as the S&P 500. This has the potential to offer much more stable returns for the investor over time and dampen otherwise white-knuckle drawdowns with an aggressive strategy in the mix. Can this be put into development?

Together with the correlation, it would be really helpful to show what a strategy’s typical margin to equity ratio is to calculate a rate of return adjusted for margin efficiency and to compare trading system performance more accurately. Can this also be put into development for use in the Grid?



Great points! I think this is the next step for the Grid here. In my book, The Global Macro Edge, the two most important aspects of a strategy are as you just articulated, return over max drawdown and capital efficiency. It’s my proposal that not only should someone screen for this variables but compensation in the hedge fund industry should be based on them as well. In essence an incremental compensation system which allows investors to pay for alpha while not overpaying for beta.

1 Like

A short term feature add would be the ability to download the (daily) equity curve. It is already plotted, so clearly available.

Allows users to run correlations between Strategies. But also against a number of other factors: market volatility. Test performance in upmkt vs downmkt. Etc

1 Like

You can download any data for particular systems.
Check out the Data Service API here