Use of/effect of compression markers in Compressor 4...

I have a question about the effect of compression markers in Compressor 4. I know they are used to mark areas where you want compression lessened so that the video looks better (like action scenes or fast camera movement).

I just worked through a Compressor 4 tutorials. The tutorial says using them forces the encoding engine to add more video data for the location vs. other locations in the video stream, thus improving the video quality (at the expense of greater file size, of course).

What I don't understand is how the period for which lesser compression is applied is determined. In other words, if you place a single compression marker, which marks a single frame, how long before and after that marker in the timeline is compression reduced? For instance, if you have a 5 second scene of sports play vs. a 2-minutes of the same action, how does Compressor know how long you want to reduce the video compression around the marker?

I know it can't apply only to the frame the marker marks. But does it apply for a second before and after? Ten seconds? Or does the marker cause compressor to analyze the action around the marker and somehow determine a period for which compression should be lessened?

Thanks!