Any limited set of knots can not address every possible knotting problem, by definition - because, even if there was such a set, we could always imagine yet another particular complication of the situation, yet another requirement from the application, that would inevitably lead to another knot.
Now, should we compile a set of knots according to the most “usual uses” of the knots in “everyday things” ? I do not think so. In fact, that would be too easy, just make a list of all the knots you use, count the numbers you use each knot, and delete all parts of the so sorted list after 5. The number of times we use some knots does not make the particular applications of those knots more important than the others, for which we use other knots. Tying shoes is not an important thing - I have not tied my shoes for decades now, just because I do not use shoes with shoelaces. Moreover, millions of people had lived without shoes and shoelaces - and perhaps less than people that lived without a knot to tie a necklace around their wife s neck ! On the contrary, I can imagine a situation where the rare knot I would have been forced to use once, and only once, in my life, would be the most important thing of what I would have ever done ! 
http://igkt.net/sm/index.php?topic=4418.msg27862#msg27862
Therefore, I think that the different knots that we should include in pour list should be different kinds of knots, that belong to different “knot categories”, i.e. different knots in relation to the knots themselves, their own functional and structural characteristics, and not different in relation to different practical applications by different knot tyers - even if these application should be addressed more or less often than others in every day life. I do not use a bend very often, but I would nt imagine any set of knots that will not include one, at least, bend.