What happens is that all of your answers to your profile questions are translated into numbers - data points. The algorithm then matches your data points with others' data points and comes up with answers to your profile based on percent match. However, what if someone said they are 5'5" instead of 5'4"? What if you answer that you are Christian but your denomination is Catholic? What if you say that you like to run but you like to run 10 miles a day and someone else says they like to run but only 1 mile a day? What if my hair is black but I put dark brown? The odds of mismatching people become astronomically high because the algorithm can't calculate variations of degree or intent. And the fewer data points they have you enter, the more simplistic the matching methodology and the higher risk of a mismatch. Phew!