Forgot your password?
Science News

Math Whiz Breaks Calculation Record 391

Posted by samzenpus
from the show-off dept.
keyshawn632 writes "The Associated Press reports that Gert Mittring, 38, needed only 11.8 seconds to calculate the 13th root of a 100-digit number in his head at a math museum in Giessen, a small town, located in western Germany. It's worth noting though that his feat will not be recognized by The Guinness Book Of World Records because of the difficulty of standardizing such mathematical challenges."
This discussion has been archived. No new comments can be posted.

Math Whiz Breaks Calculation Record

Comments Filter:
  • What? (Score:5, Insightful)

    by HardJeans (793993) <`moc.notsnhojivel' `ta' `todhsals'> on Wednesday November 24, 2004 @09:34PM (#10914857) Homepage
    I can't even read 100 digits in 30 seconds.
    • Re:What? (Score:5, Interesting)

      by ricotest (807136) on Wednesday November 24, 2004 @09:47PM (#10914955)
      The dude can memorize a 22 digit number in four seconds (according to the article) so I'm sure he can take a similar time to juggle the numbers around in his head. Perhaps his mental algorithm focuses on certain numbers at a time so that he can handle it.
      • Re:What? (Score:4, Informative)

        by jsprat (442568) on Wednesday November 24, 2004 @10:36PM (#10915236)
        Here [] is a list of two other records he holds. It hasn't been updated for the one mentioned in the myway article.

        Notice it took him 44.7 seconds to calculate the square root of a six digit number, but only 11.8 seconds to calculate the 13th root of a 100 digit number!!!!

        He also calculated the 23rd root of a 200 digit number in 40.83 seconds.
        • Re:What? (Score:3, Interesting)

          by henrygb (668225)
          The square root involves the first five places after the decimal point of an irrational number, while the thirteenth root results in an integer - this allows some tricks.

          Show me any 13th power of an integer and I can immediately tell you the final digit of the root. Similarly with 5th powers and 9th powers. But square roots of non squares don't give so many tricks.

          • Re:What? (Score:3, Interesting)

            by op00to (219949)
            Why don't people teach this in schools? Obviously this kind of stuff is pretty tricky, but there must be other interesting little tricks and relationships that when taught correctly, could have interested a whole lot more kids in math. Where'd you learn this?
      • Re:What? (Score:5, Funny)

        by hackstraw (262471) * on Wednesday November 24, 2004 @10:54PM (#10915334)
        Let me try some rough math with the help of a calculator.

        To memorize 22 digits, this guy takes ~4 seconds. So for 100 digits that would take about 18 seconds.

        Now I forgot, he did what in 11.8 seconds

      • Re:What? (Score:3, Funny)

        by iamhassi (659463)
        remind me to not let him see my credit card...
    • Roomie in College (Score:5, Interesting)

      by AlexTheBeast (809587) on Wednesday November 24, 2004 @11:47PM (#10915662)
      I roomed with a guy in college who would calculate a 10 digit by 10 digit multiplication in his head throughout the day on weekends. He would be grilling or watching TV and you would see him get him and write down 1 digit of his answer.

      In grade school he had memorized 52 decks of shuffled cards in some insane short period of time. The teacher would ask him what the 12th card of the 17 deck was... and he would start listing them forward and backward from there.

      We often went to the casinos with him. He would card count and we just would bet whatever he would bet. We would all make a $100 or so and leave. He was always afraid of getting caught.

      Some government agency approached him for running sets of numbers from point a to point b. They liked the fact that he could just put all those digits in his head without a papertrail.

      Last I heard of him, he was avoiding math as much as possible... he enrolled in some DO program in a medical school somewhere. Numbers came too easy for this guy... and he knew he would go crazy if he went into a math field.

      So now he's a doc somewhere. Probably calculating 10 by 10 digit numbers in his head as he examines you...

      • by SamSim (630795) on Thursday November 25, 2004 @12:08PM (#10918303) Homepage Journal

        Memorizing and regurgitating and manipulating numbers is a very different skill from mathematics. These are things which computers are very good at - things which we DESIGN computers to be good at so we waste minimal time on such trivialities and work on the stuff which a computer can't do, the stuff which maths is really about: logical reasoning with abstracts.

        My point is that just because he's good with numbers doesn't mean he'd enjoy (or be any good at) mathematics. Also, mathematicians can't count.

  • by Attar81 (574867) on Wednesday November 24, 2004 @09:34PM (#10914863)
    I can say all fifty states in a quarter of a second!
  • by bmzf (731840) on Wednesday November 24, 2004 @09:35PM (#10914866)
    I can do that with my eyes closed. It'll just take me a bit longer.
  • by Anonymous Coward
    Now what if a black cat crossed his path!!!!! He would then like factor the matrix code and the world would hang in an infinite loop from the resulting glitch!!!!!111

  • by rackhamh (217889) on Wednesday November 24, 2004 @09:37PM (#10914883)
    Just memorize the 13th root of every 100-digit number in existence. Sheesh.
    • Re:That's easy. (Score:5, Informative)

      by wildsurf (535389) on Wednesday November 24, 2004 @10:20PM (#10915141) Homepage
      Just memorize the 13th root of every 100-digit number in existence. Sheesh.

      Let's just think about this for a minute.

      100-digit numbers will fall between 10^99 and 10^100. Thirteenth-roots of such numbers will lie between 10^(99 / 13) and 10^(100 / 13), or in the range [41246264 .. 49238826]. That's about 8 million possibilities, and the distribution is far from linear.

      But it's linear enough that the first nine digits of the 100-digit number yield a unique possibility for a root. And the last digit of the root will be the same as the last digit of the 100-digit number, because (N mod 10) always equals (N^13 mod 10). So the problem can be tackled from both ends, with the middle digits of the root being the hardest.

      Of course, if the audience members are clued in, they can still beat the mental calculator hands down. Type the first nine digits, take the thirteenth root, and start reading off the digits; round up slightly to make the eighth significant digit match the final digit of the 100-digit number. Done.

      A college professor of mine taught us how to square 3-digit numbers in our head in seconds using tricks like this; he was able to multiply arbitrary 5-digit numbers in his head, and often performed this onstage. And for the curious, yes, I do actually have a life outside slashdot. :-)
      • Re:That's easy. (Score:3, Informative)

        by Anonymous Coward
        if you're going to mention Professor Benjamin [] you mind as well use his name :)
        • Re:That's easy. (Score:3, Informative)

          by wildsurf (535389)
          if you're going to mention Professor Benjamin you mind as well use his name :)

          You beat me to it. (Retroactively.) Sorry Art. :-)

          P.S. Try solving one of these [] in eleven seconds. :-)
      • The article didn't say whether the number had an integral 13th root or not (While reading it, I was thinking he would give the result to 3d.p., or something, which seems like quite an amazing feat to do in 11.8 seconds).

        It would be a better challenge if the computer came up with a random base (3 - 30, say), and a fairly random large number (that had an integral root in that base), and then they timed the guy over say 10 of these challenges.
      • Re:That's easy. (Score:3, Interesting)

        by Rufus88 (748752)
        because (N mod 10) always equals (N^13 mod 10)

        I know I'm going to kick myself for asking this, but why is this necessarily true?
        • Re:That's easy. (Score:5, Informative)

          by stoborrobots (577882) on Thursday November 25, 2004 @02:36AM (#10916416)
          it isn't always - it's only true for numbers which are not even and not multiples of 5...

          As for why it's true otherwise, it's because of Fermat's Little Theorem [] and Euler's Totient Function []...

          Specifically, since the Totient of 10 is 4, any number which is coprime to 10 (i.e. not even and not a multiple of 5) when raised to a power of 4, yields a 1 in the units place, (i.e. N^4 = 1 mod 10 if gcd(N,10) = 1).

          Since if a number is coprime to 10, then all its powers are coprime to 10, N^12 = (N^3)^4 also has a 1 in its units place.

          Now N^13 = N*(N^12) will always have the same last digit as N, if N is coprime to 10.
        • Re:That's easy. (Score:5, Informative)

          by kylemonger (686302) on Thursday November 25, 2004 @02:58AM (#10916485)
          (n mod 10) = (n^k mod 10) iff (k mod 4) = 1. (n > 0, k > 0)

          Since we use base 10 arithmetic (n mod 10) means we just look at the last digit. Digits repeat every fourth iteration when computing the powers of a natural number.

          Numbers ending with:
          1 -> 1,1,1,1,1,1,1,1,1,...
          2 -> 2,4,8,6,2,4,8,6,2,...
          3 -> 3,9,7,1,3,9,7,1,3,...
          5 -> 5,5,5,5,5,5,5,5,5,...

          You can see the period 4 cycles for 4, 6, 7, 8, and 9 as well. Since the digits repeat, the value of (n^k mod 10) must also repeat as k increases.
    • The number of elementary particles in the universe is estimated to be around an 80-digit number. It would be impossible to even write every 100-digit number in existance--you'd run out of matter in the universe first. Even if that were possible, just imagine the time it would take to even look at each one...

      It's really interesting to think of all the hard limits in the universe caused by things like this.
    • Just memorize the 13th root of every 100-digit number in existence . Sheesh.

      Thanks for mentioning that, imagine the time I would have wasted had I tried to memorize all the 100-digit numbers that DONT'T exist. Thank you!

  • Family guy (Score:4, Funny)

    by comwiz56 (447651) <comwiz&gmail,com> on Wednesday November 24, 2004 @09:39PM (#10914897) Homepage
    Obligitory Family Guy quote:

    Lois: Peter, why would they make you presidesnt?
    Peter: Maybe it's because I can recite all 50 states in a quarter of a second - RARF!
    Lois: Peter, that was just a loud yelping noise
  • by Anonymous Coward on Wednesday November 24, 2004 @09:39PM (#10914898)
    ...Mittring will now go for the record of longest lifespan without losing one's virginity.
    • Mittring will now go for the record of longest lifespan without losing one's virginity.


      It says he's already got the 13th root, that's 12 more than required!

  • by fred911 (83970) on Wednesday November 24, 2004 @09:40PM (#10914907)
    Five. Everyone knows that!
  • 38, ohhh (Score:2, Interesting)

    by photon_chac (306576)
    According to Neumann's thoery, a math guy reaches his peak at about 26, could this _Gert Mittring_ be a bit more 'number-crunching' at that age?
  • by dupper (470576)
    Definitely 11.8 seconds.
  • by taradfong (311185) * on Wednesday November 24, 2004 @09:49PM (#10914973) Homepage Journal
    Just as I read this article, what would start playing in my playlist but Mr. Roboto. I wonder if he has parts made in Japan?
  • by NotQuiteReal (608241) on Wednesday November 24, 2004 @09:50PM (#10914977) Journal
    ... lurking near the ATM, looking over my shoulder, memorizing my PIN.
  • did he figure that one out?

  • For all of the real records of achievement that exist in this world, there are numerous examples in their book of utter crap that doesn't amount to anything. The most people in a pie eating contest in one place. The most times running through a field serviced by multiple beehives. The largest cheese sandwich. The Houston 500. Etc, etc, etc.
  • by GreenPenInc (792018) on Wednesday November 24, 2004 @09:56PM (#10915012)
    When I was a kid, my dad lent me a book of Shakuntala Devi's book, "Figuring". She was famous some years ago (in the 50s, I believe) for her own computational ability, multiplying two 13-digit numbers in her head in 28 seconds.

    The book itself was an interesting read, and at the time I just ate it up. It has a lot of tricks regarding number theory, mathematical riddles, calendar tricks, and calculation of pi, for example. It teaches how to figure the day of the week for any Gregorian date of any time in a few seconds, a trick which I still remember and use today!

    As for the Pi, it contained a few poems and sayings whose letter counts signified the individual digits. I started trying to memorize pi, with my sights set firmly on the world record (as I am not without my own mathematical and mnemonic prowess). However, around grade 9, I decided to abandon my quest in order to get a life. I had memorized 1350 digits at that point.

    One such quote held little significance for me at the time, but has since become hilarious. "How I want a drink, alcoholic of course, after the heavy chapters involving quantum mechanics!" Needless to say, my quantum prof found it quite funny. :)

    • "It teaches how to figure the day of the week for any Gregorian date of any time in a few seconds, a trick which I still remember and use today!"

      Me too. Although I always forget what doomsday is this year (Sunday iirc).
    • I'm not sure which method you use, but I included one method in an article I wrote on memory improvement [] which some slashdot readers might find interesting.

      With some practice, you really can get to the point where you can calculate days of the week for any date in just a few seconds. People don't realize it's not all that difficult so it's a nice parlor trick.

      Also included in that article are methods for remembering long-digit numbers, the order of a deck of cards, etc.

    • It teaches how to figure the day of the week for any Gregorian date of any time in a few seconds, a trick which I still remember and use today!

      Care to share? Sounds interesting!
      • by GreenPenInc (792018) on Wednesday November 24, 2004 @10:49PM (#10915301)
        Absolutely. :) Let's see if I can type this by the end of the lecture!

        First, figure out the "year number". This part -- and the month number -- take some practice. Here's the first few to get you started:
        1900 - 0
        1904 - 5
        1908 - 3
        1912 - 1
        1916 - 6
        1920 - 4
        1924 - 2
        1928 - 0
        And it repeats thusly. Note that the "year number" starts at 0 for the beginning of the century, and is decreased by two (modulo seven) every leap year.

        In case you're interested in the other 75% of the time, simply add one to the year number for every year you add. Thus, 1901 becomes 1, 1902 becomes 2, etc.

        The "month" number requires memorization of another table, which cannot be recalculated as quickly as the year number:
        Jan - 0
        Feb - 3
        Mar - 3
        Apr - 6
        May - 1
        Jun - 4
        Jul - 6
        Aug - 2
        Sep - 5
        Oct - 0
        Nov - 3
        Dec - 5
        Add the month number to the year number. If your year is a leap year and your month is January or February, subtract 1.

        Next, add the day number. The day number is the day. :P

        Now, add or subtract sevens as necessary until you end up with a number between 0 and 6:
        0 - Sunday
        1 - Monday
        2 - Tuesday
        3 - Wednesday
        4 - Thursday
        5 - Friday
        6 - Saturday
        The result will be the day of the week.

        If your desired date does not begin with a "19", you have to add a century number as well. I believe 2000 is a leap year, since every 100 years is not but every 400 years is. Thus, the century number of 2000 is 6 (or, equivalently, -1). 1800 is 5, 1700 is 3, etc. (I am not certain of these.)

        As an example, today's year number is 5, the month number is 3, and the day number is 24. After compensating for the century by subtracting 1, we obtain 31. This reduces to 3 (by subtracting 28), which corresponds to Wednesday. Since it is Wednesday, and since I am in a large empty room, I further deduce that the lecture has ended.

        • by Zach Garner (74342) on Wednesday November 24, 2004 @10:54PM (#10915339)
          You know, that reminds me of the time I caught the ferry over to Shelbyville. I needed a new heel for my shoe, so, I decided to go to Morganville, which is what they called Shelbyville in those days. So I tied an onion to my belt, which was the style at the time. Now, to take the ferry cost a nickel, and in those days, nickels had pictures of bumblebees on 'em. "Give me five bees for a quarter", you'd say.

          Now where were we? Oh yeah - the important thing was I had an onion on my belt, which was the style at the time. They didn't have white onions because of the war. The only thing you could get was those big yellow ones...
    • High pi (Score:5, Funny)

      by dexter riley (556126) on Wednesday November 24, 2004 @10:35PM (#10915229)
      I read somewhere that you only need about 50 digits of pi to describe a circle the size of the observable universe to within the diameter of a proton, let alone a chocolate donut.

      This isn't to say that 1350 digits wouldn't be useful! If you ever wake up in an alternate universe (you were warned about operating quantum machinery while drunk!) just look up pi in a math book. The degree of trouble you're in could correlate to the digit at which your memorized value, and the local value of pi, diverge.

      If pi only diverges after 1000 or more digits, you're probably alright, except for having to re-memorize pi.
      If pi diverges after 100 digits, there may be some minor historical divergences, like, say, President Nixon being impeached, or Bush winning a second term. The mind boggles!
      If pi diverges after 30 or 40 digits, look out the window. Do dinosaurs roam the earth? Since you're surrounded by ruthless, math-book-publishing carnivores, consider donating yourself to the primate house of the zoo.
      If local pi begins with a number other than 3, you should start to get worried, or maybe implode.
      • For another such bit of trivia (and this one's much easier to work out), give this riddle to your friends:

        Imagine a (perfect) sphere the size of Earth, with a rope tight around some equatorial line, circumnavigating it. If you wanted to have the rope one meter above the ground all the way around the sphere, how much rope would you have to tie on to the end to do so?

        The answer, as it's very easy to see after knowing the answer, is 2Pi meters, or about 6.28 meters. Most people, without checking the
      • Bush DID win a second term. *tear*
      • Re:High pi (Score:5, Interesting)

        by Nyh (55741) on Thursday November 25, 2004 @04:45AM (#10916828)
        I read somewhere that you only need about 50 digits of pi to describe a circle the size of the observable universe to within the diameter of a proton, let alone a chocolate donut.

        Well, let us see:
        radius universe: about 15e9 lightyears
        radius proton: 1.2e-15 m

        circle with the size of the universe divided by diameter proton:
        2*pi*15e9*365*24*3600*300000000/(2*1.2e-1 5)=3.7e41 .
        So 42 digits of pi will do.

        42? Where did I see this number before?

  • My Turn (Score:2, Funny)

    by Stupidhead (834364)
    1 One Thousand
    2 One Thousand
    3 One Thousand
    4 One Thousand
    5 One Thousand
    6 One Thousand
    7 One Thousand
    8 One Thousand
    9 One Thousand
    10 One Thousand
    11 One Thousand
    12 One Thousand
  • by Rai (524476) on Wednesday November 24, 2004 @10:02PM (#10915044) Homepage
    Get this guy some sappho juice.
    • It is by will alone I set my code in motion.

      It is by the juice of caffiene that thoughts aquires speed, the hands develop shakes, the shakes become a warning.
      It is by will alone I set my code in motion.

      --Coder's litany.
  • I so call bullshit (Score:5, Interesting)

    by tomstdenis (446163) < minus cat> on Wednesday November 24, 2004 @10:04PM (#10915057) Homepage
    Unless there is some really trivial algorithm for finding 13th roots I totally call bullshit. If it takes him four seconds to memorize a 22 digit number how can he manipulate and find a 13th root for a 100 digit number in just over twice that amount of time?

    There has to be a trick to it aside from "thinking really fast"

    • by kfg (145172) on Wednesday November 24, 2004 @10:24PM (#10915167)
      There has to be a trick to it aside from "thinking really fast"

      Well of course, there is. Probably two or three tricks combined. . .plus thinking really fast, as well as having a good memory for numbers.

      Walking a tightrope is more than just having "good balance," and it's really just a trick, and not necessarily a very useful one, but. . .

      It is still pretty impressive and you can't do it.

  • ahh (Score:5, Funny)

    by nomadic (141991) <nomadicworld@gma ... inus threevowels> on Wednesday November 24, 2004 @10:05PM (#10915061) Homepage
    It's worth noting though that his feat will not be recognized by The Guinness Book Of World Records because of the difficulty of standardizing such mathematical challenges.

    That's the problem when dealing with a highly subjective field like mathematics.
  • by jmcmunn (307798) on Wednesday November 24, 2004 @10:12PM (#10915097)

    Probably breaking codes for some government or another. Someone with talent with numbers and such will catch the eye of someone out there. Could it be that this was just to show off his talent as a sort of "job interview"? Probably not, but I expect he will get some calls about it anyway.
  • The future is here (Score:4, Interesting)

    by forgetmenot (467513) < minus city> on Wednesday November 24, 2004 @10:14PM (#10915108) Homepage
    When I hear about people like this I can't help but think of "Dune" and it's Mentats.

    I would like to know how much of this ability is genetically determined and how much is due to training and from what age did his "gifts" become apparent.

    Either he needs to be stuck into some kinda breeding program (perhaps solving his virginity problem *hyuk hyuk*) or his training regimen needs to be studied and duplicated en masse. Imagine an advanced state-of-the-art military computer system that runs on 3-square meals a day and isn't susceptible to EMP bursts.
  • dumb tricks... (Score:2, Interesting)

    by Vellmont (569020)
    It sounds impressive, but how usefull is doing something a machine can already do more quickly and efficiently? John Henry [] learned this the hard way. As others have pointed out there's tricks and shortcuts that people use to doing these calculations, so most of it just amounts to mathematical parlor tricks.

    The implicaton is this guy is a genius. Maybe he is, but calculating roots quickly doesn't make you a genius, it just means you know some math tricks. Isn't this just the mathematical equivalent of
    • Which John Henry? Your link leads to a disambiguation page.
  • by G4from128k (686170) on Wednesday November 24, 2004 @10:19PM (#10915137)
    Gert Mittring was disqualified when judges noted a small sticker on his chest in a post-event checkup. It was discovered that he had Intel Inside.

    The news set off a legal feeding frenzy. SCO sued Mr. Mittring for using the company's super secret 13th root finder source code. Microsoft then added to the man's woes by suing for patent infringement over Microsoft's patents on 100 digit numbers. RIAA then sued him for including "8675309" in the answer -- obviously a stolen clip from "Jenny" by Tommy Tutone.
  • by Magickcat (768797) on Wednesday November 24, 2004 @10:24PM (#10915164)
    A photo of Gert Mittring can be found here. []

    Please note his rather tasteful attire.

    The page also has information on the actual rules on calculating the 13th root of a 100 digit number.
  • by bzBetty (787223)
    pppfffft, my 286 can do it faster...
  • by _Hellfire_ (170113) on Wednesday November 24, 2004 @10:30PM (#10915201) Homepage
    This guy appears to have "superhuman" math ability, and I would imagine that it's just the way this guy's brain is wired that allows him to do that.

    I always wonder if there is a condition that works in the opposite way, a bit like dyslexia for reading/writing for maths, a sort of "mathlexia" if you will. Just as dyslexia doesn't mean you're stupid, it's just that your particular model of brain doesn't comprehend words straight away, a person with "mathlexia" can't add up 137 and 48 in their head to save their life, let alone do anything complicated like division or factorisation.

    If there is such a thing as mathlexia, I'd say I've definately got it. The funny thing is, I love computers, I love programming (in C among other languages, though a mastering of assembly has persistently eluded my efforts), and I can understand even engineering diagrams and other geeky stuff. I kicked ass in English Literature at high school (even though I didn't particularly enjoy it and it's not where my passions lie); but I cannot do maths in my head if my life depended on it. Even with a calculator I get lost in the process of doing a complicated sum, but I would say I'm at least a half decent programmer. It's not that I have a problem with a logical process, it's the math part that throws me.

    Is it just the way my brain is wired? Is there a big secret no-one's telling me that will make this all easy? Am I destined for a life of going "uh huh? righto..." when someone explains a (pure) math concept to me? Or is there some hope for a math dummy like me?

    If anyone knows the answer(s) to any of this I would be eternally thankful.
    • "Is there a big secret no-one's telling me that will make this all easy? "

      There is a simple trick to math, that sounds bleeding obvious, but it took me years to truly figure out- once I did, it all became really easy.

      Math is really easy. Just learn the process, apply it, and you will always get the right answer. You say you have no problem with the logical processes- well thats all math is.

      The big secret is that there isn't a big secret. Unless you've forgotten your addition and multiplication tables,
      • That's not entirely true. What's the process for proving a theorem? There isn't really one. It requires a certain amount of creativity and skill to come up with the right steps to go through.

        In any case, going through the process is really easy to screw up unless you understand the underlying concept. That understanding (like what exactly does an integral mean?) is what can be hard for some people.

  • That involves some memory...

    Fastest time to find the char-2 differential profile of a random bijective 4x4 look up table. No gimmicky tricks there just pure nlogn work in your head ;-)

    From what others have posted and I've read on the net the 13th root is a trick to a large part. The leading digits are a strong indicator of the value of the root, etc, etc, etc.

    Or something with more practical implications... fastest time to perform an inverse cosine transform [type used in MPEG video] of an 8x8 matrix i
  • This is a 100 digit number: 19283740592837485932081293847560293618273458192031 17346932745397452409864082460814617651293753975329 Now. Get the 13th root of it..... In 11 seconds...
  • by laughingcoyote (762272) <barghesthowl@exc ... m minus caffeine> on Wednesday November 24, 2004 @10:35PM (#10915232) Journal

    I can do the 23rd root of a 163 digit number in 5.8 seconds, and I wasn't even trying. I've climbed Mt. Everest in an hour and a half. I can rewrite the Linux kernel in under an hour. I can count up to ten thousand coins in no more than a minute.

    And yet, curiously, it takes me almost...-checks watch- five minutes to make a stupid, useless post on /. Strange eh?

  • If I recall... (Score:4, Interesting)

    by gravteck (787609) on Wednesday November 24, 2004 @10:36PM (#10915237)
    I don't remember if this was the same guy I saw on TV. But the guy I saw was performing large multiplications and finding large roots in front of an elementary school class. They later showed doctors or scientists doing brain imaging on him while he solved math problems. What they found was that he was using parts of his brain that most people utilize during visualization (not sure how they were able to separate it from him actually seeing something). He said he visualizes the number in his head and then he can perform various manipulations on them and he can "see" the math work itself out. Obviously some is probably genetic, but he also commented on practicing his methods for 5-7 years. He also appears to not be the only root master [].
  • by jwise (106316) on Wednesday November 24, 2004 @10:38PM (#10915249)
    And how much about the problem did he know in advance? Did he know it would be a 13th root of a 100-digit number? Did he know that the number would be a perfect 13th power of an integer? I find it impossible to believe he calculated a 13th root of a 100-digit number in 11.8 seconds without knowing any of these things. Knowing all of them makes the problem a lot easier.

    The 13th root of a 100-digit number will always have 7 digits. If you memorize the first few digits of the 13th powers of numbers between 49 and 58 and you are given a 100-digit number, then you immediately know the first 2 digits of the 13th root. Memorize the initial digits of 13th power of numbers between 491 and 588 and you immediately know the first 3 digits. By memorizing the terminal digits of 13th powers of numbers less than 100, you could similarly immediately get the last 3 digits. That leaves 1 digit to compute, which is a slightly less impressive-sounding feat for 11.8 seconds. It's not a trivial calculation, though, and not at all shabby for 11.8 seconds.

  • at a math museum in Giessen

    a math museum ??? can someone explain what a math museum contains? surely not the pickled brain of Leibnitz next to Pascal's toothbrush?
  • by Spoing (152917) on Wednesday November 24, 2004 @10:48PM (#10915300) Homepage
    This is an amazing book. It will make you a math wiz even if you are an ace or suck at math. Just put in the time. It will even give you some appreciation for how numbers 'feel'.

    Cheap new. Even cheaper used (check Amazon).

    The book is thin and has a white cover with blue and red lettering.

  • by e_lehman (143896) on Wednesday November 24, 2004 @10:51PM (#10915315)
    The 13th root of a 100-digit number is an 8-digit number. Here's how YOU can find TWO of those 8 digits in an instant.

    1. The leading digit is ALWAYS 4.

    2. The last digit of the 13-th root of N is always the same as the last digit of N.

    (The first fact follows because Floor[N[(10^100 - 1)^(1/13)]] = 49238826 and Floor[N[(10^99 - 1)^(1/13)]] = 41246263. The second holds because N^13 is congruent to N modulo 10.)

    With minimal practice, you can get the second-highest digit from the magnitude. Beyond that I can only speculate what he's doing. But by taking an alternating sum of the digits, you get its value mod 11, which gives you the value of the root mod 11, which buys you another digit. Now you're halfway there...
  • by pvg (152136) on Wednesday November 24, 2004 @11:02PM (#10915402)
    Are described here. Rest of the site is also informative and insane. []

  • Met him last week! (Score:3, Interesting)

    by rxmd (205533) on Thursday November 25, 2004 @12:51PM (#10918523) Homepage
    Actually, I was quite astonished to see this on Slashdot, as I had lunch with the guy last Thursday where I work []. He's nice in persion, but one of the secretaries at work said he stinks and should wash more often. I'm afraid I didn't notice it quite as badly...

    He has an interesting way of getting along financially. Basically, he's living off an exclusive contract with the German TV station RTL [] where he's featured every now and then in shows. He also gives lectures on mathematical topics; RTL makes him charge a very steep EUR 2500 per lecture (about $3000). I think originally he studied psychology; he's still running the psychiatrist's practice in Cologne that he startet off with.

    We were joking about him tackling the Millenium Problems now; I wonder if he's serious about that... but then, there's more to it than calculating in your head really fast.

"Text processing has made it possible to right-justify any idea, even one which cannot be justified on any other grounds." -- J. Finnegan, USC.