Grandmaster chess and unique games

Given a random game between 2 Grandmaster level humans, how likely is it that the game will not at any time reach a position that has never been reached before?

How many moves on average does it generally take before a game is the first one to ever use that exact moveset? (Whether or not it creates an actual position that occured before by a different moveset)

I haven’t done any research, but certainly there have been Grandmaster games that were identical to previous ones. (Sorry, I hesitate to give a %.)
Mainly this is because both players are satisfied with a draw, but very rarely a Grandmaster will repeat a previous blunder and soon resign. (No doubt this happens far less since computer databases became available.)

The length of time before a new move is played will vary by opening.
For example, the Ruy Lopez (1. e4 e5 2. Nf3 Nc6 3. Bb5) has been heavily analysed and it wouldn’t surprise me to see the same 20 moves being played - before the game really starts.
An off-beat opening like the Monkey’s Bum (1. e4 g6 2. Bc4 Bg7 3. Qf3 e6 4. d4) could easily see a new position by move 10.