About the guy behind this.

Scott was a single guy looking for a casual dating relationship with a bright lady with a good disposition. After sending off something like 20 missives and getting one friend, one fizzle and one maybe, he decided that if he read another profile that had a laundry list of adjectives, he'd go nuts. Some time after going nuts, he got back out of the dating game, but that's a whole 'nother story.

So, being a numbers-oriented kind of guy, he did so in his own unique way. The Picky Woman's Date Calculator combines US Census data, information gleaned from surveys about the US population, a little basic statistics and some of Scott's own prejudices, leavened with a few wild-ass guesses, to create a system that allows women to subtract out the "undesirable" guys - the ones who don't meet their high standards - and see just how many (or how few!) are really left.

He hopes that if any real, live women play with this site and see just what happens when they hold men to their standards, at least some of them might reconsider and give those honest, caring, intelligent, funny but somewhat chunky guys like he used to be a second look.

Write to Scott

The method to this madness

The base of this data is the California population, snarfed from the US Census and rounded off for convenience. In order to better estimate the number of available dates in any given geographical area, the 2001 Census estimates of county population were used to give a table that indicates what percentage of the state's total population live in each county.

Underlying the calculation and hidden from view are three key assumptions - that women are looking for single heterosexual men, that about 90% of single men are hetero, and that the Census numbers for the various categories of "not currently married" are correct. These numbers are used to tweak the starting number of men in the calculation.

Additional Census data was used for the age and ethnic breakdowns. Scott freely admits that he may have made arithmetic or interpretative errors in this area.

Overall appearance and height numbers are based on statistical normal distributions. Appearance is just given in terms of one standard deviation - below average, average and above average. Height is computed based on a mean height of 69" and a standard deviation of 2.5", lifted from a statistics exam Scott found on the WWW. The body type calculation is based on CDC numbers for obesity (Large) and overweight (a few extra pounds) in the general US population; Scott admits that these numbers may not be valid for men in California as a group.

Intelligence is again based purely on normal distributions. Extra categories are listed on the upper side because of the number of women's profiles that ask for a man who is "intelligent".

Scott really went out on a limb with sense of humor. He said "you either got it or you don't, and half of us don't". He has no idea if this number has any relation to reality.

The political views are based on a 30/40/30 distribution of liberal/moderate/conservative. Scott apologizes if you don't believe that this adequately describes the American political spectrum.

The religion categories are based on the CUNY study of 50,000 Americans' stated religious affiliations.

Smoking is broken down using the 80/20 approximation.

Educational attainment is based on US Census data, as is income. The income figures are for California and are correct only if Scott read the tables right.

Going back to the 80/20 rule, it comes in really handy for traits. In the absence of any real data whatsoever, Scott decided to be generous to the other members of his sex and assume that 80% of them had any given good trait. While he realizes that this is entirely unsupportable statistically, he'd really like to imagine that 80% of his neighbors are good, honest people. Basically, each trait selected drops 20% of the remaining population, so picking 10 traits (the maximum) drops 90% of the original candidates on the floor.

The calculation for interests, like that for traits, is again based on entirely arbitrary criteria. In this case, Scott said, "You like something or you don't, 50/50, so that's what I'm going with." Pick one common interest, like sports, and the system throws out half the dating pool. Ok, maybe that's a bad example...let's try quilting. Hm... Anyway, absent a really long and boring data entry session with a comprehensive survey of America's favorite hobbies and pastimes, there's just no way to get a perfect answer here. This one is a killer, though, so be careful with the shared interests. Adding in just three can throw out nearly 90% of the dating pool, and maxing it out at 10 gets you to the 1 in 1000 level, and you probably weren't doing too well before this anyway.

Lastly, most variables are being treated as independent variables, with multiplicative probabilities. This is what cuts the numbers down so severely. One exception is made to the independent-variable rule - education and general intelligence are treated as related, and only one (whichever is the more restrictive of the two) is actually used to influence the final number.