July 24, 2006

  • Speed Dating Odds

     
    There are 17 men and 17 women at a speed dating event.  At the end
    of the night, each gets to write down their top two dates.  What is the
    probability of a match being made?


    To simplify the problem, assume each person only has one choice, and
    that guy #1 chooses girl #1, etc…  What are the chances that a girl
    has picked the guy who chose her?  The prob that she hasn’t picked him
    is 16/17.  So the odds that all 17 girls miss is (16/17)^17 = .357.  So
    the odds of at least one match is pretty good already: 1-.357 = 64.3%.


    Of course, it improves when you allow choice #2.  In that
    case the odds of a single girl missing her pick drops to 15/17.  So 17
    girls all missing = (15/17)^17 = .119.  Odds of at least one match:
    1-.119 = 88.1%.

    Corrected by altoz: the above paragraph holds true if you allow either gender 2 choices, but not both genders.  Allowing both genders increases the chances to 98.6%, as per his comment below.  Hey, 98.6 is the temperature of the human body.  Coincidence?  I think not…

    If you assume the guys and girls are at least somewhat socially savvy
    and able to sense a mutual connection, then the odds go up
    dramatically.  Start booking your weekends at the church now…

    The big question is, if you eventually marry your choice #2, do you ever tell your spouse they were #2???  ;)

Comments (19)

  • Not to quibble, but I don’t know if your model’s very fair to the individual — a participant isn’t concerned about whether the group as a whole results as a match, they want to know if they will walk home with a match. 

    So, their specific odds are much smaller. 

    The chance of her missing one of the two guys that picked her is just straight 1 – (15/17), right?  Granted, I think a 12% chance is better than most people would fare at any bars or parties now that I think about it…

  • You are right about the odds of an individual, whose odds are only 2/17 = 12%. Those are better odds than my experience in bars! ;)

  • This problem was studied by Gale and Shapley in the American Mathematical Monthly (1962). A brief synopsis is available here. The takeaway is that if you ask each man to rank-list the 17 women, and you ask each woman to rank-list the 17 men, then there are a number of ways you can create “stable” matches. By stable, I mean that there is no man and woman that could simultaneously defect from their assigned matches, pair off together, and both be happier than they were in the algorithm’s output. A variant of this algorithm is used to match residency programs with first-year medical residents.

  • Wow, changed, does that mean you’ve hit on 17 wymin or more at bars at one time?!?! Dang, my hero!

  • throw in some game theory and this, too

  • you know, i just redid this and the probabilities aren’t correct.

    the probability that the guy chosen by the girl at #1 has her down at #1 or #2 is 2/17. the probability that the guy chosen by the girl at #2 has her down at #1 or #2 is also 2/17.

    Thus, the probability of a match (at #1 or #2) for a given girl is 1-(15/17)*(15/17) = 22%

    The probability of 0 matches is
    ((15/17)^2)^17 = 1.4%

    Thus, the probability of a match is 98.6%

    The expected value (where if person A chose B and C and both B and C had A down would count as 2 matches) is something like:

    (1-(15/17)^2)*17 = 3.76

    So we should have expected 3.76 couples assuming a completely random distribution. BTW, I got to this conclusion because I was trying to run some simulations on this.

  • Thanks, I knew I could count on you guys to figure this out!

  • you guys are NERDS!!  :)   i’m so impressed.

  • “husband!? what was all that one in a million talk?”
    -Lloyd, Dumb and Dumber

  • No matter what simulations you run or probability statistics or theoretical framework…..

    I WAS THE TOTAL MAC-DADDY THAT NIGHT!!!! MUHahahhahahahaha!!!!!

    Your simulations will have to figure in only 16 guys and 15 wymin (-2 for the two wymin who matched with me). Oh and then, you’d have to take away 1 choice from all the wymin, since it’ll be a constant (meaning all 17 wymin put my name as #1)….

  • that sounds pretty competitive of you, jwkcheng. a little testosterone-heavy from working out recently?
    ! b

  • James, you forgot to figure in:

    1) The wymin will put you down for both #1 and #2. After a taste of the JWK-Master, who could settle?

    2) The men will put you down for #1 and #2 also.

    So 2 of the wymin will luck out (although for both to really luck out, you’d all have to move to Utah). And that leaves 31 broken-hearted singles!

  • ybrfdy: It’s not a competition when you know what the outcome is going to be before it actually happens!

    Zoiks!!! >;^P

    changed: So true so true! My heart goes out to the 31 broken-hearted singles, and even more so to Wymin #2…. so close yet so far. =( sadness.

    Maybe I should retire now from speed-dating as to keep the hope alive for all of you Kraft cheese slices out there!

  • hmmm…altoz beat me to it with regards to the probabilities, and he said it much more elegantly than i ever could. i look forward to jwkcheng’s post on this.

  • hey changed, i can’t believe you sat and thought through/calculated all of this. you nerd! =D please say hello to the the missus for me.

  • I’m clueless about the probability stuff, but the answer to that last “big question” is “no.” 

  • (hanna)
    oh geez…this post started off so fun for me, then i came upon the word problem-esque scenario/probability math stuff, and skimming others’ responses…i got all panicky, head-rush, sweaty palms all over again…

    just like the last math class i ever took: algebra II/trig.

    but, kudos to all who get it….

  • LoL. You should write a little book of all of this quantitative and graphed-out nerdy wisdom on dating. i’d buy it

  • hello – remember to take into account personalities…

Post a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *