Also consider how the opponent has been conditioned based on what you did after you hit their protect earlier in the game. This will strongly influence their play, consciously and unconsciously.I guess what I'm thinking is in situations like this it seems kinda like Rock-Paper-Scissors to a beginner like me. If you assume your opponent is as smart as you, then you can predict their play and outplay them at level 2. But if they aren't smart enough they might not even play into your prediction. Or if they are, they might know you will play at level 2 and outplay you at level 3.
In a lot of other situations in pokemon there is a rather gradual risk reward scale that is easily understood and even if you predict wrong you can play the "safest" or "high risk high reward" etc, but when it comes to protect strategy it seems like it's all-or-nothing RPS. Even if you accurately assess the situation your opponent could just pick randomly and screw it up by not playing "correct". (obviously in my example I know I lost not by chance but because the other opponent knew what I would do but I didn't know what he would do i.e. he outplayed me)
How accurate is that?
Also if you want to practice reads in Pokémon I would recommend playing 1v1 for a little bit. The format is basically built around getting good reads and in a very limited number of turns.