Manufacturers are testing a die to make sure that it is fair (has a uniform distribution). They roll the die 102 times and record the outcomes in the table below. They conduct a chi-square Goodness-of-Fit hypothesis test at the 5% significance level.
The null and alternative hypotheses are:

H0: The die has the uniform distribution.
Ha: The die does not have the uniform distribution.

Outcome 1 2 3 4 5 6
Expected 17 17 17 17 17 17
Observed 5 11 21 17 7 41

Compute the test statistic χ20, rounded to three decimal places.