if that “guaranteed” base grows, it provides a voting offset that could allow the candidate to worry less about losing the support of less progressive voters.

Sure.

But why would they? If the base that's 'grown' is guaranteed, then why shift at all? Why not have the new larger guaranteed base, and the less progressive voters. After all, the guaranteed base is guaranteed, you don't need to do anything to get their votes.

But let's say they want to risk it for ideological reasons (no evidence at all that this is the case, but for the sake of argument we could assume it).

You've still not addressed the two main questions.

1. How do they know the extra votes came from left-leaning but 'guaranteed' voters, and not from voters who really liked their centrist policies?

2. If they have some way of knowing (polls, focus-groups etc) then why can't they use that way of knowing to ask about voter commitment, and make the move to the left before the election, why do they need us to actually vote first to find out if we're in this 'guaranteed base'?