If I'm not mistaken, the Netflix recommendation algorithm corrects for this effect by normalizing the scores. For example, suppose you're a movie cynic. The m

If I'm not mistaken, the Netflix recommendation algorithm corrects for this effect by normalizing the scores. For example, suppose you're a movie cynic. The most you'll ever rate a movie is 3; suppose your average rating is a 2. Someone else might rate movies from 3. to 5 with an average rating of 4. From the recommendation engine's point of view, the cynic's 2 rating is equivalent to the other person's rating of 4. Thus the algorithm corrects for different notions of "average."