MY 'belief' with the ball entering the left-side media is based on the length of the vertical pipes...

Once the pipe is long enough, the ball will pick up enough momentum during its 'fall' to push it across
the boundary between light and dense media.
Whenever there are two media face-to-face (as at the boundary) there will be an equivalent of 'surface tension'.
The greater the difference in density, the greater the tension.
So if our two media are close to the same density, then the tension the ball needs to break is less, so the slower
it can be moving.
BUT THEN the density of the ball needs to be more accurately calculated in order to sink on the right side but float on the left.

Remember, it has been a LONG time since scool for me, so this is all based on rusty theory...