Tristan Miller
German Research Center for Artificial Intelligence^{1}
20 December 1999
This is a question that practically every male has asked himself at one point or another in his life. Unfortunately, there is rarely a hard and fast answer to the query. Many men try to reason their way through the dilemma nonetheless, often reaching a series of ridiculous explanations, each more self-deprecating than the last: "Is it because I'm too shy, and not aggressive enough? Is it my opening lines? Am I a boring person? Am I too fat or too thin? Or am I simply ugly and completely unattractive to women?" When all other plausible explanations have been discounted, most fall back on the time-honoured conclusion that "there must be Something Wrong™ with me" before resigning themselves to lives of perpetual chastity.^{2}
Not the author, though. I, for one, refuse to spend my life brooding over my lack of luck with women. While I'll be the first to admit that my chances of ever entering into a meaningful relationship with someone special are practically non-existent, I staunchly refuse to admit that it has anything to do with some inherent problem with me. Instead, I am convinced that the situation can be readily explained in purely scientific terms, using nothing more than demographics and some elementary statistical calculus.
Lest anyone suspect that my standards for women are too high, let me allay those fears by enumerating in advance my three criteria for the match. First, the potential girlfriend must be approximately my age—let's say 21 plus or minus three or four years. Second, the girl must be beautiful (and I use that term all-encompassingly to refer to both inner and outer beauty). Third, she must also be reasonably intelligent—she doesn't have to be Mensa material, but the ability to carry on a witty, insightful argument would be nice. So there they are—three simple demands, which I'm sure everyone will agree are anything but unreasonable.
That said, I now present my demonstration of why the probability of finding a suitable candidate fulfilling the three above-noted requirements is so small as to be practically impossible—in other words, why I will never have a girlfriend. I shall endeavour to make this proof as rigorous as the available data permits. And I should note, too, that there will be no statistical trickery involved here; I have cited all my sources and provided all relevant calculations^{3} in case anyone wishes to conduct their own independent review. Let's now take a look at the figures.
We start with the largest demographic in which I am interested—namely, the population of this planet. That is not to say I'm against the idea of interstellar romance, of course; I just don't assess the prospect of finding myself a nice Altairian girl as statistically significant. Now anyway, the latest halfway-reliable figures we have for Earth's population come from the United States Census Bureau's 1999 World Population Profile (WP/98). Due presumably to the time involved in compiling and processing census statistics, said report's data is valid only as of 1998, so later on we'll be making some impromptu adjustments to bring the numbers up to date.
I'd've thought that, given the title of this essay, this criterion goes without saying. In case anyone missed it, though, I am looking for exclusively female companionship. Accordingly, roughly half of the Earth's population must be discounted. Sorry, guys.
We now further restrict the geographical area of interest to so-called "first-world countries". My reasons for doing so are not motivated out of contempt for those who are economically disadvantaged, but rather by simple probability. My chances of meeting a babe from Bhutan or a goddess from Ghana, either in person or on the Internet, are understandably low. In fact, I will most likely spend nearly my entire life living and working in North America, Europe, and Australia, so it is to these types of regions that the numbers have been narrowed.
Being neither a pedophile nor a geriatrophile, I would like to restrict my search for love to those whose age is approximately equal to my own. This is where things get a bit tricky, for two reasons: first, the census data is nearly two years old, and second, the "population by age" tables in WP/98 are not separated into individual ages but are instead quantized into "15–19" (of whom there are 39 560 000) and "20–44" (population 215 073 000). Women aged 15 to 19 in 1998 will be aged 17 to 21 in 2000; in this group, I'm interested in dating those 18 or older, so, assuming the "15–19" girls' ages are uniformly distributed, we have \[39\,560\,000 \times \frac{\left| 21 - 18 \right| + 1}{\left| 19 - 15 \right| + 1} = 31\,648\,000.\] Similarly, of 1998's "20–44" category, there are now \[215\,073\,000 \times \frac{\left| 25 - 22 \right| + 1}{\left| 44 - 20 \right| + 1} = 34\,411\,680\] females within my chosen age limit. The sum, 66 059 680, represents the total number of females aged 18 to 25 in developed countries in 2000. Unfortunately, roughly 1% of these girls will have died since the census was taken;^{6} thus, the true number of so-far eligible bachelorettes is 65 399 083.
Personal attraction, both physically and personality-wise, is an important instigator of any relationship. Of course, beauty is a purely subjective trait whose interpretation may vary from person to person. Luckily it is not necessary for me to define beauty in this essay except to state that for any given beholder, it will probably be normally distributed amongst the population.^{7} Without going into the specifics of precisely which traits I admire, I will say that for a girl to be considered really beautiful to me, she should fall at least two standard deviations above the norm. From basic statistics theory, the area to the left of the normal curve at z = 2 is \[\frac{1}{2} - \frac{1}{\sqrt{2 \pi}} \cdot \int_{0}^{2} e^{-\frac{1}{2}z^2} dz~\approx~0.022\,75\] and so it is this number with which we multiply our current population pool.
Again, intelligence can mean different things to different people, yet I am once more relieved of making any explanation by noting that it, like most other characteristics, has a notionally normal distribution across the population. Let's assume that I will settle for someone a mere one standard deviation above the normal; in that case, a further \[\frac{1}{2} + \frac{1}{\sqrt{2 \pi}} \cdot \int_{0}^{1} e^{-\frac{1}{2}z^2} dz~\approx~84.1345\%\] of the population must be discounted.
I could find no hard statistics on the number of above-noted girls who are already married, engaged, or otherwise committed to a significant other, but informal observation and anecdotal evidence leads me to believe that the proportion is somewhere around 50%. (Fellow unattached males will no doubt have also noticed a preponderance of girls legitimately offering, "Sorry, I already have a boyfriend" as an excuse not to go on a date.) For reasons of morality (and perhaps too self-preservation), I'm not about to start hitting on girls who have husbands and boyfriends. Accordingly, that portion of the female population must also be considered off-limits.
Naturally, finding a suitable girl who I really like is no guarantee that she'll like me back. Assuming, as previously mentioned, that personal attractiveness is normally distributed, there is a mere 50% chance that any given female will consider me even marginally attractive. In practice, however, people are unlikely to consider pursuing a relationship with someone whose looks and personality just barely suffice. Let's make the rather conservative assumption, then, that a girl would go out with someone if and only if they were at least one standard deviation above her idea of average. In that case, referring to our previous calculation, only 15.8655% of females would consider someone with my physical characteristics and personality acceptable as a potential romantic partner.
It is here, at a pool of 18 726 acceptable females, that we end our statistical analysis. At first glance, a datable population of 18 726 may not seem like such a low number, but consider this: assuming I were to go on a blind date with a new girl about my age every week, I would have to date for 3493 weeks before I found one of the 18 726. That's very nearly 67 years. As a North American male born in the late 1970s, my life expectancy is probably little more than 70 years, so we can safely say that I will be quite dead before I find the proverbial girl of my dreams. Come to think of it, she'll probably be dead too.
So there you have it, my friends—finally, a cogent, scientific, non-self-deprecating argument for why I will never have a girlfriend. That said, if you happen to be a girl deluded enough to think that you and I have a chance together, feel free to drop me a line, but I warn you, you face odds of 157 060 to 1. I wouldn't bother if I were you.
Update (2000-04-01): My sarcastic pleas for some e-mail have finally been answered. Take a look at this letter from a hysterical female reader, which I think perfectly demonstrates the point of this entire essay. (I think the fact that she's a WebTV user explains a lot—in fact, I was sure this e-mail was an April Fool's joke until I noticed the return address.)
You are free to produce translations of this article as long as you credit me as the original author, and link back to the English original if possible. Please send me a link to your translation and I will add it to this list.
Varför har inte jag en flickvän? [Why I will never have a girlfriend]. In Är tärningen kastad? Sannolikhetslära för vem som helst. Bombadil Publishing, Trollhättan, 2008. ISBN 978-91-85765-02-7. Translated by Andreas Svensson.
Why I will never have a girlfriend. In Laurence Behrens and Leonard Rosen, editors, Writing and Reading Across the Curriculum. Longman, 8th edition, 2002. ISBN 978-0-321-09102-4.
Why I will never have a girlfriend. The Annals of Improbable Research, 8(3):13–17, 2002. ISSN 1079-5146.
Why I will never have a girlfriend. In Laurence Behrens, Leonard Rosen, and Bonnie Beedles, editors, A Sequence for Academic Writing. Longman, 2001. ISBN 978-0-321-08133-9.