re: the last "wisdom of crowds" bit Assuming a random sampling of customers, would you not get the same "highly accurate" result if you assigned 1 to those who

re: the last "wisdom of crowds" bit Assuming a random sampling of customers, would you not get the same "highly accurate" result if you assigned 1 to those who bought the item and 0 to those who didn't and averaged the results? --- Though I suppose one thing you could say about it is that the positive and negative bias just about cancel each other out.