by Aimee Bender
DoubledayBuy at Powell’s »
Welcome to the second year of Bracketology!
In 2010, I was asked to add some statistical order to the seeming chaos of the always unpredictable Tournament of Books. Like the N.C.A.A. March Madness from which our tournament takes its structure, I looked for trends, tendencies, and temptations at play when it comes to determining who advances to the next round. Some trends that I found were stronger than others; others may in fact be completely factitiousbut hey, that’s the nature of expert opinion.
Of course, how books advance in our tournament is different from how basketball teams advance in March Madness, although I imagine both contests involve a slightly queasy amount of sweat. Books advance by convincing judges that they are better than their opponents; what counts is not whether the judge likes the book, but that she likes it better than the other book. This is obviously also quite different from how we usually readalthough some of the most frequent comments about a book are it’s better than this book or it’s not as good as that one, we rarely structure those comparisons quite so rigorously against one specific other book. What kinds of attributes, then, might affect that process of comparison?
I am personally a big believer in the idea that the physical properties of a book have a substantial effect on our interaction with it.1 I’m not so much talking about that gut-level thrill you get when you touch a certain book and everything about just feels right (or just wrong). Well, actually I am talking about that, but rightness is a little difficult to define and plug into Excel. In discussing Lionel Shriver’s So Much for That, John Warner half-jokingly pointed to its disaster of a cover as a contributing factor in its loss, but again, I don’t feel confident in my abilities to create a reliable metric for Cover Awfulness.
One thing that is a little easier to attach numbers to is bulkor page count. It’s difficult not to think about size when you’re comparing two books, especially if the difference in page count is significant. A longer book is also probably one that the judge has spent more time on, and so it’s plausible that some sense of personal investment will attach itself to the novel, if not of accomplishmentwho doesn’t feel pretty good about herself after getting to the end of a 500-page baggy monster?
Here are the year-by-year average page lengths of both winners and losers for the opening round:
|Total Last 3||370.83||348.04||22.79|
2005, the first year of the tournament, was obviously a strange year, but there also hasn’t been a great deal of consistency in terms of judges’ preferences for longer or shorter books. What is notable is that, after 2005, the losers’ average page length changes very littlethough the message should hardly be don’t write books between 330 and 360 pages or you’ll never win the Rooster.
Of course, these kinds of averages can hide quite a bit, so let’s look at individual matchups where the difference between the two books amounts to 300 pages or more:
|Cloud Atlas||Finishing School||337||Longer||2005|
|I Am Charlotte Simmons||Wake Up, Sir||354||Longer||2005|
|Jonathan Strange||The Rope Eater||496||Longer||2005|
|Birds Without Wings||Harbor||336||Longer||2005|
|On Beauty||Beasts of No Nation||322||Shorter||2006|
|The Historian||Home Land||336||Shorter||2006|
|Against the Day||Pride of Baghdad||984||Longer||2007|
|Tree of Smoke||Ovenman||372||Longer||2008|
|Savage Detectives||Let the Northern Lights||351||Shorter||2008|
|New England White||You Don’t Love Me Yet||332||Shorter||2008|
|2666||Steer Toward Rock||643||Longer||2009|
|Shadow Country||Disreputable History||547||Longer||2009|
Now, including this year’s matchup of Skippy Dies (661 pages) and A Visit From the Goon Squad (273 pages), the overall record looks like this: eight wins for longer books, six wins for the shorter ones. That’s not quite as strong a trend as one might expect, however, to justify what we last year called a gravitas gap. On the other hand, if we break out the individual records of different page ranges, maybe we’ll get a better idea of what’s happening. The following graph also gives you a preview of what happens in the quarterfinals:2
|Page Count||Opening Round
|>200, <300 Pages||37.14%||18.18%|
|>300, <400 Pages||45.00%||57.14%|
|>400, <500 Pages||64.29%||55.56%|
The outstanding opening-round performance of very short books (like this year’s Nox, by Anne Carson) is somewhat surprising, but as you can see, that success doesn’t carry over to the quarterfinals, while the very solid performances by longer books do continue. And overall, the last two rows show that, whether the longer book is longer by 3 or 300 pages, fortune doesn’t favor the briefthe longer of the two books takes the bracket at a quite healthy clip, opening round or quarterfinals.
What else do judges compare when they’re thinking about which book to promote to the next round? One important possibility is genre, although I would argue that genre is not very useful to us, in part because so many books make legitimate claims to being two or three or more genres at once. Comparison, therefore, is often multivariable, not reducible to a straightforward statisticsay, dystopian novels beat historical fiction 75 percent of the time.
On the other hand, what we often mean when we say genre is something like presumptions about where a book fits into the larger ecology of literature. These presumptions can certainly produce elaborate classifications and taxonomies, but they also often operate more impressionisticallyfor instance, it seems that many people think that a historical epic that required a great deal of research is a more serious or more challenging work than a domestic novel. One (very inadequate) shorthand for collecting broad similarities in these types of fiction has been the author’s genderwho a writer is often activates routines in the reader’s mind of how to place him or her within the literary ecology, and the book itself is sort of assigned a genderand often an ageto whom it is supposed to appeal, thereby determining its place more securely in the ecology.
Gender discrimination has recently been a hot topic of conversation among literary types, but I worry a little that by focusing primarily on numberswith an implicit goal of a simple parity between gendersthe question becomes less how a woman’s experience in the literary marketplace often differs from a man’s, and more about how we can avoid embarrassing ourselves by appearing to favor men too much. Parity isn’t always experientially the same as equalityjust because there is an equal number of women and men in a room doesn’t mean that men might not still benefit from some forms of gender privilege.
Numbers do have their place, however, and here’s a quick table of the gender breakdown in the Tournament of Books:
|Male||Female||M %||F %|
|Total Books, Last 3 Years||28||20||58.33%||41.67%|
|Total No. 1 Seeds||19||9||67.86%||32.14%|
|No. 1 Seeds, Last 3 Years||7||5||58.33%||41.67%|
|Opening Round Wins
M v. F Matchups, Total
|Opening Round Wins
M v. F Matchups, Last 3 Years
As you can see, men have quite an edge in the opening roundthey’ve made up almost 60 percent of the books that make the tournament, more than two-thirds of the no. 1 seeds, and they also win almost three-fifths of the time. Again, my point isn’t to focus too much on the numbers, but rather to suggest thatparticularly this year, when such good conversations have been had about how gender affects the reception of books that might, under the covers, actually look a lot alikeparity or something like it might facilitate those discussions, if for no other reason than that it’s difficult to have these discussions without having read a lot of fiction by womenand a lot of fiction by men. And on the other hand, as we may be about to see, women have historically owned the quarterfinals, winning 67 percent of the time in head-to-heads against men.
The other obvious peculiarity of the way a judge reads books in the tournament versus how she might read them in real life is in dealing with the seeding of books. The way that seeding has worked through seven years of the tournament, however, has been simple: No. 1 seeds do have a significant advantage, but the no. 2 seed historically has basically no effect whatsoeverdespite no. 2 seeds having swept their opening round matches this year. Here are the numbers for the whole seven years of the opening round:
And here are the numbers for six years (2005-2010) of quarterfinals:
|No. 1 v. No. 2||7||6||1||85.71%|
|No. 1 v. No. 3||11||8||3||72.73%|
|No. 2 v. No. 4||3||3||0||100.00%|
|No. 3 v. No. 4||3||1||2||33.33%|
Yet one wonders if seeding here is really just acting as a stand-in for that nebulous concept known as hype. Seeding is easy to keep track of, but could we even begin to quantify hype in the Tournament of Books? And if we can, can we do so in a helpful way? That is, if we make a reasonable stab at quantifying the Hype Quotient of this year’s books, does it have any correlation with the opening round’s results? Is there a preference for underdogs, or a submission to the juggernauts? Is there a level at which too high a hype becomes perilous for a book, leading to backlash?
These are some questions I want to take up in the next Bracketology, where I’ll actually make an attempt to put some numbers behind the idea of hype. For now, I hope you’ve enjoyed the opening round and are looking forward to the quarterfinals.
1Although I’m not sure what to make of the fact that judges seem increasingly to be doing their reading on iPads, Kindles, or iPods. Comments are very welcome here. ↩
2Also, 2011 results are factored into the opening round data. ↩
Andrew Seal is a second-year grad student in American Studies. Fortunately for his career, he’s not called on to work with numbers very often. His research interests include Midwestern literature, the history of consumer culture, and frontier/Western history.