The null hypothesis is the hypothesis in statistics on which the burden of disproof is greater than the burden of proof. To put it simply, in order to reject the null hypothesis the analyst/statistician/etc. must provide statistically significant proof that the alternative hypothesis

- contrary to the null hypothesis - should be accepted, thus rejecting the null hypothesis. The general convention is that the null hypothesis treats parameters as being equal to zero, so that the onus is to “prove” that the parameter is significant.

It should be fair to treat god’s existence as such: the null hypothesis is that there is no god, and the alternative hypothesis is that there is at least [a] god. It is technically very easy to reject the null hypothesis, as all one has to do is provide significant evidence that there exists at least [a] god.

I’ve never understood any rationale for asserting that disbelievers must prove that god doesn’t exist. Perhaps there is a valid way of handling that argument by making allowance for individual’s religious experiences; experiences that cannot be denied, but cannot be allowed to prove anything for any third parties.