Yes, we don't really know. But at some point the gamble is just too big.

Because the costs aren't just numbers. They represent hundreds or thousands of person-years of effort. You're proposing that a large number of people should spend their entire lives supporting this (either directly as scientists, or indirectly through funding it) - and maybe end up with nothing to show for it.

And there's the opportunity costs. You could fund hundreds of smaller, yet still substantial scientific efforts in many different fields for the cost of just one particle accelerator of the size we think is sufficient to yield some new observations.