Writing "for (int i=0; i<limit; i+=step) {}" is easy. Debugging the off-by-one error requires that you understand the alignment issue you ignored in the first p

Writing "for (int i=0; i<limit; i+=step) {}" is easy. Debugging the off-by-one error requires that you understand the alignment issue you ignored in the first place. That would make debugging harder, not because debugging is inherently hard, but because that's the place you're forced to deal with the hard bit you skipped earlier. I wonder how often that applies, compared to well-understood-but-implementation-mistake code?