Glossary
T-statistic
An estimate divided by its standard error. It counts how many standard errors the estimate is away from zero.
The t-statistic of a mean is t = mean / standard error. For large samples it is close to normal, so a value above about 2 in size is conventionally called significant at the 5% level.
For a return series, t = Sharpe ratio x sqrt(years) when the Sharpe ratio is annual. For many tests, a higher bar than 2 is needed.
Example: a mean of 6% a year with a standard error of 4.74% has a t-statistic of 1.26. It is well below 2, so the data do not separate the mean from zero.