Why I Will Never Have A Girlfriend

Tristan MillerGerman Research Center for Artificial Intelligence∗

Erwin-Schrodinger-Straße 5767663 Kaiserslautern, Germany

tristan.miller@dfki.de

20 December 1999

Abstract

Informal empirical and anecdotal evidencefrom the (male) scientific community has longpointed to the difficulty in securing decent,long-term female companionship. To date,however, no one has published a rigorous studyof the matter. In this essay, the author inves-tigates himself as a case study and presents aproof, using simple statistical calculus, of whyit is impossible to find a girlfriend.

Why don’t I have a girlfriend?

This is a question that practically every malehas asked himself at one point or another inhis life. Unfortunately, there is rarely a hardand fast answer to the query. Many men try toreason their way through the dilemma nonethe-less, often reaching a series of ridiculous expla-nations, each more self-deprecating than thelast: “Is it because I’m too shy, and not ag-gressive enough? Is it my opening lines? Am Ia boring person? Am I too fat or too thin? Oram I simply ugly and completely unattractiveto women?” When all other plausible expla-nations have been discounted, most fall backon the time-honoured conclusion that “theremust be Something WrongTM with me” be-fore resigning themselves to lives of perpetualchastity.1

∗This paper was written when the author was atGriffith University, Australia.

1After a short period of brooding, of course, thesemales will eventually come to the realization that thereal reason they were never able to get a girlfriend is

Not the author, though. I, for one, refuse tospend my life brooding over my lack of luckwith women. While I’ll be the first to ad-mit that my chances of ever entering into ameaningful relationship with someone specialare practically non-existent, I staunchly refuseto admit that it has anything to do with someinherent problem with me. Instead, I am con-vinced that the situation can be readily ex-plained in purely scientific terms, using nothingmore than demographics and some elementarystatistical calculus.

Lest anyone suspect that my standards forwomen are too high, let me allay those fearsby enumerating in advance my three criteriafor the match. First, the potential girlfriendmust be approximately my age—let’s say 21plus or minus three or four years. Second, thegirl must be beautiful (and I use that term all-encompassingly to refer to both inner and outerbeauty). Third, she must also be reasonablyintelligent—she doesn’t have to be Mensa ma-terial, but the ability to carry on a witty, in-sightful argument would be nice. So there theyare—three simple demands, which I’m sure ev-eryone will agree are anything but unreason-able.

That said, I now present my demonstra-tion of why the probability of finding a suit-

that they were too discriminating with their attentions.They will consequently return to the dating scene, en-tering a sequence of blase relationships with mediocregirls for whom they don’t really care, until they finallymarry one out of fear of spending the rest of their livesalone. I am convinced that this behaviour is the realreason for today’s alarmingly high divorce rate.

1

able candidate fulfilling the three above-notedrequirements is so small as to be practicallyimpossible—in other words, why I will neverhave a girlfriend. I shall endeavour to makethis proof as rigorous as the available data per-mits. And I should note, too, that there willbe no statistical trickery involved here; I havecited all my sources and provided all relevantcalculations2 in case anyone wishes to conducttheir own independent review. Let’s now takea look at the figures.

Number of people on Earth (in1998): 5 592830000[WP98, Table A–3]

We start with the largest demographic in whichI am interested—namely, the population of thisplanet. That is not to say I’m against theidea of interstellar romance, of course; I justdon’t assess the prospect of finding myself anice Altairian girl as statistically significant.Now anyway, the latest halfway-reliable fig-ures we have for Earth’s population come fromthe United States Census Bureau’s 1999 WorldPopulation Profile [WP98]. Due presumably tothe time involved in compiling and processingcensus statistics, said report’s data is valid onlyas of 1998, so later on we’ll be making some im-promptu adjustments to bring the numbers upto date.

. . . who are female:2 941118000

[WP98, Table A–7]

I’d’ve thought that, given the title of this es-say, this criterion goes without saying. In caseanyone missed it, though, I am looking for ex-clusively female companionship. Accordingly,roughly half of the Earth’s population must bediscounted. Sorry, guys.

2Due to rounding, figures cited may not add up ex-actly.

. . . in “developed” countries:605 601000

[WP98, Table A–7]

We now further restrict the geographical areaof interest to so-called “first-world countries”.My reasons for doing so are not motivated outof contempt for those who are economically dis-advantaged, but rather by simple probability.My chances of meeting a babe from Bhutan ora goddess from Ghana, either in person or onthe Internet, are understandably low. In fact,I will most likely spend nearly my entire lifeliving and working in North America, Europe,and Australia, so it is to these types of regionsthat the numbers have been narrowed.

. . . currently (in 2000) aged 18to 25: 65 399083[WP98, Tables A–3, A–7]

Being neither a pedophile nor a geriatrophile,I would like to restrict my search for love tothose whose age is approximately equal to myown. This is where things get a bit tricky, fortwo reasons: first, the census data is nearly twoyears old, and second, the “population by age”tables in [WP98] are not separated into indi-vidual ages but are instead quantized into “15–19” (of whom there are 39 560 000) and “20–44” (population 215 073 000). Women aged 15to 19 in 1998 will be aged 17 to 21 in 2000; inthis group, I’m interested in dating those 18 orolder, so, assuming the “15–19” girls’ ages areuniformly distributed, we have

39 560 000× |21− 18|+ 1|19− 15|+ 1

= 31 648 000.

Similarly, of 1998’s “20–44” category, there arenow

215 073 000× |25− 22|+ 1|44− 20|+ 1

= 34 411 680.

females within my chosen age limit. The sum,66 059 680, represents the total number of fe-males aged 18 to 25 in developed countriesin 2000. Unfortunately, roughly 1% of these

2

girls will have died since the census was taken;3

thus, the true number of so-far eligible bache-lorettes is 65 399 083.

. . . who are beautiful: 1 487 838

Personal attraction, both physically andpersonality-wise, is an important instigator ofany relationship. Of course, beauty is a purelysubjective trait whose interpretation may varyfrom person to person. Luckily it is not neces-sary for me to define beauty in this essay ex-cept to state that for any given beholder, it willprobably be normally distributed amongst thepopulation.4 Without going into the specificsof precisely which traits I admire, I will saythat for a girl to be considered really beautifulto me, she should fall at least two standard de-viations above the norm. From basic statisticstheory, the area to the left of the normal curveat z = 2 is

12− 1√

2π·∫ 2

0e−

12z2

dz ≈ 0.022 75

and so it is this number with which we multiplyour current population pool.

. . . and intelligent: 236 053

Again, intelligence can mean different things todifferent people, yet I am once more relievedof making any explanation by noting that it,like most other characteristics, has a notion-ally normal distribution across the population.

3[WP98] gives the annual death rate for developedcountries as 10 per 1000, but does not list death ratesper age group. Presumably, the death rate graphs as abathtub curve, but in absence of any numbers support-ing this hypothesis, and for the sake of simplicity, I willconservatively estimate the death rate among this agegroup to be 1% biennially.

4Despite my efforts to research the matter, I couldfind no data on the distribution of beauty, either outeror inner, amongst the population. Perhaps attractive-ness, being a largely subjective trait, does not lend itselfto quantification. It is not unreasonable, however, toassume that like most other traits, it has a normal dis-tribution. Indeed, this assumption seems to be backedup by informal observation and judgment—in any rea-sonably large group of people, most of them will beaverage-looking, and a tiny minority either exceedinglybeautiful or exceedingly ugly.

Let’s assume that I will settle for someone amere one standard deviation above the normal;in that case, a further

12

+1√2π

·∫ 1

0e−

12z2

dz ≈ 84.1345%

of the population must be discounted.

. . . and not already committed:118027

I could find no hard statistics on the numberof above-noted girls who are already married,engaged, or otherwise committed to a signifi-cant other, but informal observation and anec-dotal evidence leads me to believe that theproportion is somewhere around 50%. (Fellowunattached males will no doubt have also no-ticed a preponderance of girls legitimately of-fering, “Sorry, I already have a boyfriend” asan excuse not to go on a date.) For reasonsof morality (and perhaps too self-preservation),I’m not about to start hitting on girls who havehusbands and boyfriends. Accordingly, thatportion of the female population must also beconsidered off-limits.

. . . and also might like me:18 726

Naturally, finding a suitable girl who I reallylike is no guarantee that she’ll like me back.Assuming, as previously mentioned, that per-sonal attractiveness is normally distributed,there is a mere 50% chance that any given fe-male will consider me even marginally attrac-tive. In practice, however, people are unlikelyto consider pursuing a relationship with some-one whose looks and personality just barelysuffice. Let’s make the rather conservative as-sumption, then, that a girl would go out withsomeone if and only if they were at least onestandard deviation above her idea of average.In that case, referring to our previous calcula-tion, only 15.8655% of females would considersomeone with my physical characteristics andpersonality acceptable as a potential romanticpartner.

3

Conclusion

It is here, at a pool of 18 726 acceptable fe-males, that we end our statistical analysis. Atfirst glance, a datable population of 18 726 maynot seem like such a low number, but considerthis: assuming I were to go on a blind date witha new girl about my age every week, I wouldhave to date for 3493 weeks before I found oneof the 18 726. That’s very nearly 67 years. Asa North American male born in the late 1970s,my life expectancy is probably little more than70 years, so we can safely say that I will bequite dead before I find the proverbial girl ofmy dreams. Come to think of it, she’ll proba-bly be dead too.

Reference

[WP98] U.S. Bureau of the Census, ReportWP/98, World Population Profile: 1998.Washington, DC: U.S. Government Print-ing Office, 1999.

4