There's nothing wrong with 1-based indexing. The only reason it seems wrong to you is because you're familiar with 0-based, not because it's inherently worse.
That's simply untrue. 1-based indexing is inherently worse because it leads to code that is less elegant and harder to understand. And slightly less efficient but that's a minor factor.
If a language has a well-designed collections library (think Smalltalk and derivatives, not C++'s STL), the difference between 1 and 0-based indexing is hard even to notice, much less lead to "less elegant" code. Between Stream and Collection subclasses, the API is rich enough that using indexes is reserved for very low-level operations that you seldom use in non-FFI, non-VM code.
What you say is true for languages that don't have collections, real arrays, or vectors, only memory ranges. This is the case of C, but not Fortran, Pascal, or Ada. So yeah, if all you have is a hammer, you'd better use 0-based nails; hopefully, though, we'll allow for non-hammer tools in popular toolboxes sometime this century.
