Interesting problem that we face everyday. Knuth assumes that people are either big or little choosers. I don't think this is the case, e.g. my strategy is dependent on the size of the rolls.
When the roll is really big (i.e almost brand new), the low quality toilet paper (that my company thinks we're worth, if you're working in a better place this is a moot point) tends to tear, so that I can't get a piece of adequate length. This is because the tension on the paper is proportional to the moment of inertia of the roll and larger rolls have higher inertia. So, given a choice, I avoid these rolls.
On the other hand, really small rolls are annoying, too, since it's hard to estimate if the amount left is enough, i.e. while pulling a piece of paper they may finish unexpectedly, which is annoying.
So, my strategy is: avoid the very big and very small rolls. Around medium size, go for the smaller roll. I wonder how Knuth's analysis could accommodate such dynamic strategies based on roll size.
Wait - what kind of person would draw from the big roll?
The goal isn't to use them both up simultaneously. You use all of one and then the stall still has the other one during the period it takes the janitorial staff to notice and replace the empty.
Or perhaps we should be asking: What are all the other NSF-funded scientists doing on their bathroom breaks? Knuth's output has clearly been accounted for.
This paper was published in the American Mathematical Monthly, a journal which is mainly known for its expository articles, not necessarily for original research.
For that purpose (exposition) this paper does its job admirably. Whether it's useful research is not really important in evaluating whether it was worthy of its funding.
When the roll is really big (i.e almost brand new), the low quality toilet paper (that my company thinks we're worth, if you're working in a better place this is a moot point) tends to tear, so that I can't get a piece of adequate length. This is because the tension on the paper is proportional to the moment of inertia of the roll and larger rolls have higher inertia. So, given a choice, I avoid these rolls.
On the other hand, really small rolls are annoying, too, since it's hard to estimate if the amount left is enough, i.e. while pulling a piece of paper they may finish unexpectedly, which is annoying.
So, my strategy is: avoid the very big and very small rolls. Around medium size, go for the smaller roll. I wonder how Knuth's analysis could accommodate such dynamic strategies based on roll size.