Two sides to every argument:
1. He's done it a million times like that, it's always been fine.
2. It only needs to go wrong once and...
Unless there is some major advantage to doing option 1, putting at least some sort of risk mitigation in place seems more sensible. There's enough risk in aviation without ignoring simple solutions to reduce it (and I should know). We're doing 'very risky' things all the time - but we attempt to reduce that risk as much as practically possible (whilst still doing it!).