In N bits, an integer can represent 2^N indexes starting at zero. If you start at index 1, you'll only be able to access 2^N-1 indexes. Using 1-based indexes therefore requires indexes twice as big and requires more code to support the same array sizes. This was not a luxury that could be afforded.

This may not be the case anymore, but you can't ignore how it's been done for ages, and there are many other good reasons as you've seen throughout this thread.