Ad 1) Just for the record, my stance is not completely objective with the first turns matter thinking.
Ad 2) I see, yeah that is a third option. Plays differently.
a) Decision and repositioning without consequences: I thought this worked well even as a part of every battle. The only problem there was the nonsencical yes/no. With % chance that's similar to the roll for a map you get. However, as you can influence the chance it puts a strong incentive to specialise one character if want to use re-positioning. Perhaps the chance can be based on the party sum instead of the highest? (so the player can decide if specialize one or make more allies good)
b) Consequence of failed roll: So the encounter boils into one percentage number, and you just weight it against the consequence of failed roll. People enjoy low chance gambling esp. if immediate reward (eg. slot machines), but I personally still see only one real decision, the same strategical one as in (a) - whether invest or not into Bypass on Sam (to get close-to-100% more often). Btw. -1 to initiative is low, not having any impact on fast guys...
c) Re-positioning itself has consequences: Tactical in comparison to (a) & (b). Deciding if trying to undertake the action, resp. how much, is worth based on where units are. I've mentioned a drawback of 'yes' decision. But can easily be more refined than that. For example: each tile moved => -1 to initiative on the first turn for this unit. So how many steps you want to go depends on the base initiative. Can be combined: eg. with (a) roll happens in the background and you weight the steps, or with (b) the failed roll has different consequence than steps (eg. distracted status), etc.
Ad 3) Listened to the video, the correlation sounds great as such % estimation adds to the difficulty of the encounter. Btw. Survivalists again? They are gradually becoming even more difficult than already were compare to others.