The WebVTT parser does handle position, line, and align directives and try to use them in the painter code that draws those subtitles. However, I think you've uncovered a bug in the interaction between the alignment and position values. It looks like we're using the line and position values to set the top-left corner of the text area, and then centering the text in the box that's left there on the right side of the screen.
I'll see if we can get this fixed for a future release. In the meantime, try removing the position tags and see if that gives you centered text like you want.