I don't understand craziness, so craziness will explain all claims of miracles equally, and if there are more miracles that can easily be explained by a god than ones that could not (and gods doing miracles doesn't explain everything) then that's evidence towards a god. Also, it's evidence that crazy people just see miracles similar to what we'd expect from a god, which is admittedly a pretty good possibility since they're both supposed to be people.

Perhaps you understand craziness better, and it doesn't explain everything equally. In which case, there'd be some miracles that it doesn't explain. Those miracles would be evidence against the theory that all claims of miracles are from crazy people.

If you're discovering for the first time that there are crazy people and you suddenly have an alternative explanation, then it's evidence against gods. But it will just bring you to the same posterior probability of someone who knew about crazy people from the beginning and updated their beliefs in miracles appropriately.

It's like you're saying "lying explains everything". No it doesn't and it's not meant to do so.

Imagine someone tells you that the winning lottery numbers were 18-24-27-42-43 / 34. There's only a one in 175,223,510 chance that those are actually the lottery numbers that won. Since there's more than a one in 175,223,510 that they're lying, it might seem at first that you should conclude that they're lying. But lying explains all 175,223,510 possibilities equally. If they lie, there's still only a one in 175,223,510 chance that they'd say those numbers. So it's not evidence one way or the other.

Granted, some psychologist might be able to tell you that people are terrible random number generators, and that some combinations are more likely if they're lying. But someone who doesn't know about that has no evidence either way, and someone who does would need to look at the specific set of numbers to tell whether or not they were lying.