Your animations seem to happen in fundamentally different scenarios than the ones I have shared. Your mini slides are basically forehands forced into backhand due to sudden direction change while those of CPU which I shared happen after he has hit the shot. The CPU's mini slide seem like they are a consequence of changing direction running a long distance. The takeaway seems to be mini slides are more of a consequence of direction change rather and inertia of motion(not direction) has little to do with them.
I guess the main difference with the CPU is that the human player always makes fast direction changes, while the CPU can do them a bit more slowly, which might help trigger that animation. In that video, you can see that the end of that animation is eaten away several times.
In that case, it seems to me that the requirements for perfectly animated mini slide are extremely stringent, which only CPU can do. It would be great if these requirements are lessened for the human player and we see human player do minislide after running a long distance to hit the ball and immediately change the direction. The triggering mechanism should take into account the way human players change the direction in analog stick, which will obviously not be as gradual and smooth as CPU.
On another note, how are the inside out animations working on your end? Is this an experimental build of the game? They never seem to work for the human player for me.