Wrong results

[laxris]

We've been using the chi-squared distribution function, and it seems to give us wrong answers. We've compared with the libgsl implementation, and it seems that for many cases, pchisq is just wrong.

For example, test_statistic = 10605, extra_parameters = 9

double p_value1 = 1.0 - gsl_cdf_chisq_P (test_statistic, (double)extra_parameters);
double p_value2 = gsl_cdf_chisq_Q (test_statistic, (double)extra_parameters);
double p_value = 1.0 - stats::pchisq(test_statistic, extra_parameters, false);

p_value is 0, but p_value1 = 1, which is the correct answer. Any ideas what's going on? Are we using it wrong?

Edit: It seems like values lower than 0.05 are often wrong.

[kthohr]

Thanks for the bug report. This was due to an overflow in the gcem::incomplete_gamma function. Should be fixed now, but let me know if it isn't.

[laxris]

Amazing, that seems to fix the issue indeed. Thanks a lot!