In my opinion, using only this pressure policy will introduce a constant unwanted delay. I suggest adding to that a momentum-based approach that recognizes whether the mouse was placed or thrown by analysing its speed and acceleration.
I'm aware that this does not cover the case where the mouse is already close to the side, as the momentum will be low regardless of what the user does. That's why I'm suggesting using this in *addition* to the pressure policy instead of replacing it.
In my opinion, using only this pressure policy will introduce a constant unwanted delay. I suggest adding to that a momentum-based approach that recognizes whether the mouse was placed or thrown by analysing its speed and acceleration.
I'm aware that this does not cover the case where the mouse is already close to the side, as the momentum will be low regardless of what the user does. That's why I'm suggesting using this in *addition* to the pressure policy instead of replacing it.