The life of a mathematical physicist  from earning his PhD, through marriage, fatherhood and into a midlife crisis  presented in the form of homework exercises from a math book.
We first meet our protagonist, A, in problem number 1:
(quoted from Problems for SelfStudy)
1. Time t equals zero
A is on a train traveling due west along the xaxis at a constant velocity of seventy kilometers per hour (70km/h). He stands at the rear of the train, looking back with some fondness at the town of (6,3), his point of departure, the locations of the university and his few friends. He is carrying a suitcase (30kg) and a small bound volume (his thesis; 0.7 kg; 7 years).
Using the information given, calculate A's final position.

Believe it or not, the author is able to produce a coherent and moving story by continuing in this format.
Despite the fact that it seems like an exercise in itself (perhaps an assignment for a creative writing class), this short story works well as art and fiction. Each of the stories in the collection Third Class Superhero where I found this are, in their own way, the sort of experimental writing in which the plotline and the emotional content often get lost in the structure. That Yu is able to avoid this problem in each of the different scenarios is a testament to his writing skill.
Things I love about this story:
 The description of how he felt about his thesis when he first proved the main result reminds me of the comment from a professor that first inspired me to become a mathematician myself. The professor had said that when you make a discovery, just by thinking and proving a theorem that nobody has ever found before, you know that you are the only person in the world who knows it. Perhaps, he said, that is because nobody ever cared about that particular question, but still it is a wonderful feeling to be the one who has found it. In the story, this is presented in the following form (as if it were the multiple choice answers to problem #4):
(quoted from Problems for SelfStudy)
(a) In [his thesis], he discovered a tiny truth.
(b) When he had written the last step in his proof, A smiled.
(c) A's tiny truth is about a tiny part of a tiny sliver of a tiny subset of all possible outcomes of the world.
(d) When A brought it to his advisor and mentor, the esteemed P, P smiled. A's heart leapt.
(e) P said: What it lacks in elegance, it makes up for in rigor.
(f) P also said: What a wonderful minor result.

 I like the analogy between the threebody problem of dynamics and the way his life becomes more complicated after the birth of his first child with B.
 I like the discussion about what is "negligible". It leaves ambiguous the interesting question of who is being unrealistic: A (for thinking that one can ignore certain aspects of the world around us) or B (for not realizing how effectively we can understand the world if we just look at it piece by piece).
Things I don't like so much:
 Although the way the story is told is quite beautiful and original, the character of A and the events described are nearly cliche. Perhaps that is necessary in this format, but as you may know from reading other reviews, I find the stereotype of the socially inept mathematician who hides from reality behind mathematics to be quite tiresome, and Yu's creative style makes it only bearable.
 Another "pet peeve" of mine is that many people do not even realize there is such a thing as a research mathematician, and I like to think that fiction is a good way to address this deficit. So, I am disappointed that the story suggests that A is a "physicist". The fact that he is described as being a theoretician who does not worry about experiments, that he proves theorems, and that his area of research is "nonlinear dynamic equations" all give me confidence in describing A as a mathematician.
 Towards the end, along with problems in his marriage, he also has trouble with the memory of his thesis. This, I think, is realistic. Researchers are often not very proud of their old work (I think it might be the same for artists, composers, performers and writers), and especially in a midlife crisis this could be exaggerated. However, the suggestion of the story that counterexamples to the main theorem are discovered frequently is a bit extreme. If even one counterexample is found, then it is no theorem at all but rather a mistake. Of course, mistakes do happen, but it seems unlikely that such an error would have gone unnoticed by P and the rest of the thesis committee. It would have been more believable, I think, if it was found to have been discovered before by someone else (that seems to happen in science much more than I expected!) or that it was usurped by a bigger and better result by another author.
Originally published in Harvard Review in 2002, this story has recently been released in a collection of stories by Charles Yu, Third Class Superhero. It contains another story which might be considered "mathematical fiction", although I am not giving it a separate entry on this website. The story "32.05864991%" describes the field of "emotional statistics".
