One-way restrictions on local streets, which tend to have low traffic stress, can create a significant barrier to low-stress cycling. Contraflow, a treatment that undoes one-way restrictions on bike travel, has the potential to improve low-stress connectivity. Although contraflow is applied routinely in the Netherlands and Belgium, it has been sparingly applied in the United States. We propose refined measures of connectivity and accessibility that account for one-way restrictions by requiring a low-stress round trip path between origins and destinations. Different methods of associating origin–destination demand from polygons with a street network were analyzed. These methods are particularly important where there are one-way restrictions and irregular street networks because of the assumptions they entail in relation to first- and last-segment travel. In a case study of Greater Boston, we found that with the current bike network, low-stress connectivity between homes and jobs would increase from 1.2% to 8.7% if one-way restrictions on local streets were eliminated. We also found that even with a dense mesh of low-stress main bike routes, connectivity would still be 16% lower without contraflow on local streets than with. These results suggest that creating a network of main bike routes is not always enough; it is also important to provide contraflow on local streets. The Boston study also found that providing contraflow on selected links representing only 3% of local one-way street mileage delivered 40% of the connectivity impact of universal contraflow. Based on this finding, a method is proposed for prioritizing streets for contraflow conversion.