Well, the blinding is supposed to allow the juniors to speak up when the seniors are making mistakes, so it makes sense its only one way.

And while double blinding sounds nice in theory I'm not so convinced it is that useful in practice, because it requires the reviewer to play along and pretend they can't figure it out from the text alone: if they are able to do that reliably they can probably be trusted to keep an open mind anyway.

Reliable blinding of the author would mean having them consciously copy the style of others and avoiding citing their own previous work, which would be very hard in a small sub-field since they are by definition a sizeable fraction of it!