The similarity scores for the extend pairings also exhibit an overall decay as the time difference between viewings increases. Stated differently, the graph 400 indicates that the similarity score is inversely proportional to the number of hours that elapse between respective viewings of video content. In the graph 400 of FIG. 4, similarity scores for six weeks 410a-f of extending pairings of video content are shown. The time difference, ΔT, between 0 and 168 corresponds to the first week 410a; between 168 and 336 hours corresponds to the second week 410b; between 336 hours and 504 hours for the third week 410c; and so forth respectively for the fourth week 410d, the fifth week 410e, and the sixth week 410f. As seen in FIG. 4, the average similarity score gradually decreases between the first week 410a and the sixth week 410f. In particular, the graph 400 illustrates that the similarity score for the weekly peak 406a at the end of the first week 410a is higher than the similarity score for the weekly peak 406b at the end of the second week 410b which is, in turn, higher than the similarity score for the weekly peak 406c at the end of the third week 410c, and so forth through the sixth week 410f. In addition, the graph 400 illustrates that the average similarity score for the daily peaks 408a during the first week 410a is higher than the average similarity score for the daily peaks 408c during the third week 410c which is, in turn, higher than the average similarity score for the daily peaks 408f during the sixth week 410f. Accordingly the graph 400 indicates that the similarity between video content decreases as the time that elapses between viewing the video content increases.