As per the title.
I’m in a bit of a pickle since I know that the equity curve of a profitable system should be exponentially increasing. Is linear regression appropriate for checking the “health” of an equity curve? If it’s good enough, what minimum value is appropriate? I think that R^2 = 0.9 is good enough but am interested in how other people approach this topic.
Cmon peeps. Does NO-ONE graph an equity curve after a backtest?
Well it could have been looked up on the internet. Lets approach the 1st assuming your making, that a profitable system’s equity curve should grow exponentially. If your trading the same size from start to finish the answer is NO. it should not. It should represent linear growth. Compounding of returns, in this case increasing position sizing as the capital grows, is how a profitable systems equity curve should become exponential.
R^2 is the coefficient of determination of a regressional model, in this case a linear regression. If you have an intercept in an ordinary least squares model, then you have yourself a Pearson correlation coefficient. a linear models R^2 should be a number between 0 - 1. The closer to 1 the better the model fits the data.
Yeah, what I’m wondering is what do other traders use as a benchmark, if they indeed use one? I did google it but found no actual numbers, just definitions.