Is a crypto strategy's weekly edge distinguishable from noise?

Question

A systematic crypto strategy has 100 weekly returns. The average weekly return is 0.20% and the standard deviation of the weekly returns is 1.5%.

Compute the t-statistic for the hypothesis that the true average weekly return is zero, and decide whether it is significant at the 5% level (t of about 2).

Accepted Answer

The formula. The t-statistic is the average return divided by its standard error, which rearranges to the per-period Sharpe times the square root of the number of periods: t = \frac{\bar{r}}{s}\sqrt{n}. In plain terms: how big is the average return compared to its typical weekly wiggle, scaled up by the square root of how many weeks you have. Plug in the numbers. The per-week Sharpe is 0.20\%/1.5\% = 0.1333. With n = 100, \sqrt{100} = 10, so t = 0.1333 	imes 10 = 1.33. Verdict. A t of 1.33 is well below 2, so the edge is not statistically significant. Almost two years of weekly data still is…

Is a crypto strategy's weekly edge distinguishable from noise?

Your answer

More Statistics questions