Warning: Big Post.
Except it's bullshit. People who use that argument make a leap from 'There's a 50% chance a god exists' to 'There's a 50% chance THE God exists'.
It also makes absolutely no sense to go from the original statements even to a 50/50 split. Just because there are two "options" says absolutely nothing bout their percentage chances of being true. For example if I had a perfect random number generator that generated a single random number from 1-10. I could then hypothetically give these two statements:
A. The number will be greater than 9.
B. The number will be less than or equal to 9.
No where does it say that just because I have divided the outcomes into two groups that those groups have anything near the same possibility of being true, and in fact this example is a perfect showcase of how that can not be true. In that example B is 9x more likely to be true than A is.
So it isn't just a division amongst many different gods or possibilities, it's also a total unknown just on what the base chances are in the first place. For all I know the statement "C. God Doesn't exist" is infinitely times more likely to be true than "A. God Exists", or maybe it's the other way around. The original statements show absolutely nothing about their relative probabilities of being true.
Because without those things all inquiry is impossible and we may as well be writing gibberish. Science doesn't work if you can just say "well that or maybe a wizard did it" to every hypothesis (See Ancient Aliens for this in action), math doesn't work if numbers don't equal themselves, etc.
It might be slightly more correct to say "math wouldn't be as commonly useful if numbers didn't equal themselves". You can actually rewrite pretty much any basic principle of any field and still have it "work" as long as the new laws are consistent (since all of our labels and symbols are really just arbitrary codified conventions anyways); it just often wouldn't be near as useful for actual purposes. For example I could redefine "=" to mean "the first digit of the number to the left is the same as the last digit of the number on the right" and still be able to perform potentially useful calculations despite the fact that this: 31 = 31 would no longer be true. (31 = 43 would be though). As such "Logic" actually can be rewritten in that way and still count as logic (albeit with a different system of basic axioms and operators).
The word "Science", on the other hand, is an excellent example of imprecise language. Science as a particular bit of knowledge such as the "science of math" or even the "science of physics" can have it's base axioms rewritten and still function as a working system. We could, for example, redefine all negative charges as positive and vice versa, and weigh things in units of "wizards" with a logarithmic scaled number system and you could still use said system to calculate valid scientific knowledge (though many things might not be quite as easy to calculate as in the current system
). And even if we take "Science" to mean "the scientific method" it can still "work" if I'm allowed to say "well that or maybe a wizard did it" to every hypothesis, it just wouldn't work near as well (and even if the laws said that "a wizard did it, period", that still wouldn't mean they were totally nonfunctional, because they would be a valid method in the extremely unlikely case that a wizard
did do it!).
As long as the rules that underwrite any system are consistent, then it doesn't matter how close to gibberish anything looks, because it has actual laws that it follows, and thus allows for inquiry. So you could give someone a statement that was founded on totally new basic principles, it would just be all but impossible to understand until the person was able to figure out said basic principles (at which point it would be functionally equivalent to giving them the translated statement in an already known system).