The response to your question is a little more complicated than just calculating areas of watershed.
First, the effectiveness of the BMP must be determined and the percentage of annual rainfall. LEED provides some guidance concerning the average annual rainfall depth, where LEED suggests that the depth of treatment for a humid area is 1-inch, semi-arid is 0.75 inch and arid is 0.5 inch. The IMP (Integrated Management Practice) treatment train must be designed to remove 80% of TSS for this run-off volume (rainfall depth x area x run-off coefficient x BMP effectiveness). Table 2, page 104 of the LEED Reference Guide provides some guidance on the average TS removal for different BMP’s. This process will provide you with the level of conformance that you can reasonably expect to meet for your project for the BMP’s that you have selected.
If a portion of the project site is not providing any treatment, then the reduction in TSS that you are achieving is reduced by the portion of the site for which you have no treatment. If the portion of the site that is being treated is being treated to a higher level than 80% TSS removal, you can claim some credit against the portion of the site that is not being treated.
So, you are correct that if 5.0 acres of runoff is draining to a bioretention area and the total area of the LEED boundary is 10.0 acres, then the % of Annual Rainfall treated by the bioretention area is 50% assuming a similar run-off coefficient for all of the site boundary. However, it is possible that you are not meeting the requirement for 80% TSS reduction for 50% of the site, depending on the size and effectiveness of your bioretention area, or it is possible that you are achieving more than 80% reduction for 50% of the site.