Back in college I lived in a townhouse with seven other guys. We had a house website with bio pages for the eight of us, head shots, etc
One day we threw all our headshots into some morphing software and got a snapshot of what our composite roommate would look like. We threw him onto a website called Hot Or Not where vanity- and charity-cases alike upload photos for others to rank on a ten-point scale. “We” pulled an eight and partied continuously for several weeks.

Until last week, I had no idea Hot Or Not was still around. It is and has drawn the attention of researchers from Columbia, Carnegie Mellon, and M.I.T. All these people feeding quantifiable preferences into Hot or Not’s servers, millions on the month, constitutes an ideal data set. Another apparent data set is Hot Or Not’s “Meet Me” service where you can meet someone but only if he or she wants to meet you also.

All that data and analysis, recently released, fascinated me for a 72-hour stretch last week. My job description, as I try to frame it these days, is to make some fraction of what excites me about life (and math, in particular) exciting to my kids. This one was difficult.
Mostly, it’s tough striking up a good conversation over questions which have already been answered. You’ve gotta tease them with clues without frustrating them, drifting just enough information past ’em without giving it all away.
Here’s how we built from nothing to something:
- Discuss the question: “why do we date?” In my classes, answers ranged from the expected (someone’s fun to be around, cute, makes you feel happy) to the really expected (someone’s hot, horny, fertile, has money).
- Focus for a bit on the “fertile” answer and how making babies is a biological imperative for every species.
- Introduce hotornot.com. These people upload photos of themselves, you rank one, another photo appears, and then it’s three hours later.
- Ask: “Who do you think hands out more ‘hot’ ratings? Guys or girls?” In my classes, both genders selected themselves. (Which confounds me still. Anyone know why?) Turns out it’s guys by 240%. (Which seems like a lay-up to me, but there ya go.)
- Introduce the MeetMe service. If you want to meet this guy here, you click “meet,” at which point your photo is sent to him and then if he wants to meet you, your e-mail addresses are exchanged, love is found, babies are made.
- Ask: “Who do you think clicks the ‘meet’ button more? Guys or girls?” Guys again.
- Ask: “Who do you think clicks the ‘meet’ button more? People with low ratings or high ratings?” Low. Turns out, for each ranking you slip (from a 6 to a 5, for example) you become 25% more likely to accept a date.
- You assign the class a hotornot ranking. You arbitrarily assign ’em a 7, setting them up for what happens next.
- You say, “How do you decide which invitations to accept? Do you accept a 3?” Everyone says no. “Do you accept a 6?” Some would only date their level and above. Others recognized that the rankings came from a community that didn’t, e.g., share their affinity for baldness, and would consider a 6.
- And finally, you show this chart-gem.
- You talk about it. The x-axis is tricky. It’s the difference between you and the person asking you out. You point to the extreme left edge and say, “-5. How hot is that person?” Some will say “-5.” Others see that she’s a 2.
- You gesture at the y-axis and say, “Is your probability of going out with this person low or high? Low. Obviously.
- Ask: “What about this graph is expected?” You talk about how the graph rises, as you’d expect.
- Ask: “What about this is unexpected?” And this is, of course, the kicker. Why does the graph take a dip at the end? Why would you decline to meet someone who was five ranks hotter than you? (Your thoughts are welcome in the comments.)
Total cost: ten minutes. In all, a really good way to kick off a period. If you’re feeling brave and you’ve got their trust, you can discuss the question, “how do 2’s find and maintain love?”
20 Comments
fgk
December 9, 2007 - 5:14 pm -heh. you have a favorite backstreet boy.
dan
December 9, 2007 - 5:39 pm -And The Shawshank Redemption was my favorite movie.
*smacks forehead*
Not sure dy/dan’s servers can handle all this honesty.
Jenny
December 9, 2007 - 6:36 pm -You are a mathematician.
“The following six movies” followed by a list of seven.
dan
December 9, 2007 - 6:38 pm -Alright, alright. Anyone have any commentary on the post proper or are we just gonna needle me over what a dork I was a decade ago? Y’know what? No one answer that.
Peter Rock
December 9, 2007 - 7:13 pm -I’m kinda curious as to what this means to you now…
“Where I’ll end is uncertain but wherever I go, there God will be also.”
fgk
December 9, 2007 - 9:17 pm -some things to note about the graph:
1) the probability scale only goes to 0.5. the average accept rate for people of the same hotness is less than 1/3. questions for the class: how important is appearance in potential dates? what are potential sample biases compared to the general population?
2) it is smooth on the low end, and becomes more jagged at the high end. while there is no record of how many samples there are at any point, it indicates that there is a lot more data (and therefore lower error) at the low end than at the high.
dan
December 9, 2007 - 10:54 pm -Yeah! fgk wins Commenter Of The Evening for staying on point. Regretting that link more by the hour.
Peter, my theology has grown messier and more complicated in the intervening years and, as a flawed representative of Christianity, I’ve never brought it up here. (Saying nothing for its relevance to a teaching blog.)
But since you asked: my faith gives me a certain fearlessness which I’ve found essential in the classroom. I’ve established so much of my identity in who my God thinks I am that, e.g., when I blow a slam dunk in front of the student body, or when some kid tells me that me and my mathematics can fuck off, it doesn’t shake me too hard. I can respond with as much love as my position allows.
Chris Lehmann
December 10, 2007 - 6:06 am -I am speechless for so many, many reasons.
The best thing is that I can see myself doing a walkabout and wandering in on your class while you have Hot Or Not up on the screen… listening for a few minutes as you make your mathematical case and the backing slowly out of the room.
“Hi honey, how was school today?”
“Great… Mr. Meyer taught us all about Hot or Not in Math! Want to see the website?”
These are the moments that make administrators proud.
(seriously… sounds like a great relevant hook to get kids to do some data analysis.)
And I’ve got a question for you —
Given all the data analysis you do with your kids, what are your thoughts about the whole “American High School Math –> Calculus Progression vs. Statistical Analysis Progression?”
I’ve seen some interesting stuff saying that we should re-examine the generally agreed upon sequence so that more students were exposed to deep levels of probability and statistics rather than everyone moving toward calc.
Thoughts? Is it a false dichotomy?
dan
December 10, 2007 - 7:00 am -I don’t know if it’s a false dichotomy but it isn’t one I easily recognize. Every high school I’ve attended or worked had equal sections statistics and calculus, with only the rarest students taking them both. *coughs, polishes his bronzed high school transcript*
At my public college, statistics was favored heavily over calculus, though a lot of majors required ’em both.
I dunno. Calculus is the payoff for that year of unmitigated abstraction called Algebra II. Taught well it can be literally and permanently life-altering. If I had my way, though, we’d add a few modern flairs to the statistics syllabus and push it harder.
Chris Lehmann
December 10, 2007 - 7:26 am -O.k. — pushing this idea a little further then.
What about those kids who aren’t going to get to Calc in HS?
We have a lot of kids in America who do:
9th – Alg I
10th – Geometry
11th – Alg II
12th – Pre-Calc or no math
Would it make more sense for those kids to do something like:
9th – Alg I
10th – Geometry
11th / 12th — Probability and Statistics?
It’s not inconceivable that schools could develop two strands of math sequence for kids…
dan
December 10, 2007 - 7:56 am -You’d know better than I would if charges of tracking would fly.
Practically speaking, though, I’ve got no beef with any of that. The second track would be much less challenging, for whatever it’s worth, and I don’t think Probability or Statistics or its hybrid could sustain two years. There just isn’t enough material.
Chris Lehmann
December 10, 2007 - 10:34 am -I think if it was self-selecting, then you could avoid that charge…
Here’s my question… and this is not based on tons of serious scholasticism, but it’s something I’ve thought about a bit and it was trigger by your “Calc is your reward for Alg II” comment.
What is the most important math that kids retain? Alg I — I see it. Makes sense to me. Geo — same. Got it.
What I see, however, is a world where the phrase “data-driven” is more and more used, and very few people seem to really understand how to use data in meaningful ways — and worse, very few people seem to know how to judge / make sense of data. Show too many people a graph, and you get a sense that there’s this thought, “It’s got MATH in it, it MUST be true!”
Does it make more sense to teach kids how to really examine, judge, create statistics, etc… so that we increase the level of math literacy? Would giving kids a one or two year sequence of that as an alternative to the Calc track make sense?
dan
December 10, 2007 - 12:23 pm -Yes to all of the above.
Sarah
December 10, 2007 - 3:47 pm -My high school required the “Probability, Statistics, and Discrete Mathematics” course before Pre-Calc.
I wish my school now offered a statistics course at all. I haven’t integrated it nearly enough this year. Dan, thanks for providing ideas of how to connect the statistics to my courses.
Jackie
December 10, 2007 - 4:29 pm -Weighing in on the stats vs. calc question: in the long run, for most students, I’d have to say that stats would be more useful (although taking both would be nice too).
Rich
December 10, 2007 - 5:00 pm -So I’ve got to ask, is that smaller photo of the guy in the white suit (the screenshot of Hot or Not) the composite pic of you and your seven roommates? C’mon, own up to it.
I am still thinking about that dip at the high end of the curve, but that could FINALLY help me understand why I’m 42 and still single. Something about people being intimidated by the “too perfect?”
And finally, I must reflect back to Jerry Seinfeld saying that 95% of the population are undatable:
Jerry: Have ya been to the Motor Vehicle Bureau? Its a leper colony there.
Elaine: So, basically what you’re saying is 95% of the population is undatable?
Jerry: UNDATABLE.
Elaine: So how are all these people gettin’ together?
Jerry: Alcohol.
Peter Rock
December 11, 2007 - 2:03 am -“as a flawed representative of Christianity”
What do you mean by “flawed”? What does Christianity mean to you?
dan
December 11, 2007 - 3:54 pm -I’m not sure if my comment box puts a ceiling on word count but that second question could go on for awhile. Only time for the first, I guess:
I find “Christian” and “Christianity” to be descriptive of a lot of things and nothing all at once. At the question, “Are you a Christian?” I’m almost at a loss for a response given how many different people will claim “Christian” as a default label.
Give my penchant for occasional profanity and for what many assume is self-aggrandizement, I don’t rep Christianity much from my blog. I just hate the idea that my flaws and sharp edges would turn someone off from the legitimately cool teachings of Christ.
Ben Chun
December 12, 2007 - 6:22 pm -From discrete math, it’s only a few steps over to some boolean logic and an introduction to computer programming, and/or algorithms. Not that you’d want to push large number down that path either. I’d actually love to run semester math classes like “Math for Gambling” and “Math for Screensavers” and “Math for Video Games”. Wow, that sounds terribly un-academic. We’ll have to come up with better course titles.