Slashdot Mirror
Pi Recited to 100,000 Digits
DiAmOnDirc writes "Akira Haraguchi, 60, needed more than 16 hours to recite the number to 100,000 decimal places, breaking his personal best of 83,431 digits set in 1995, his office said Wednesday. He made the attempt at a public hall in Kisarazu, just east of Tokyo. Haraguchi, a psychiatric counselor and business consultant in nearby Mobara city, took a break of about 5 minutes every one to two hours, going to the rest room and eating rice balls during the attempt, said Naoki Fujii, spokesman of Haraguchi's office. Fujii said all of Haraguchi's activities during the attempt, including his bathroom breaks, were videotaped for evidence that will later be sent for verification by the Guinness Book of Records."
More specifically, he memorizes the digits by making a story, probably from the sound of the numbers.
In Japanese, you can make a play on words by the sound of the numbers called goro-awase. For example, if there is a sequence of numbers such as "3341", it can be read as "sa-mi-shi-i" which means "sad". By having a series of these play on words, he can make up a story, which is much easier to remember than a sequence of numbers.
If you're curious, here is the article (in Japanese) that mentions that the guy makes a story to memorize the digits.
Um, he's in his 50s, he's married, and his wife was there when he was reciting it.
Just remember 355/133 (3.1415929...). It's the most accurate fraction possible with only a three-digit numerator and denominator. (WP)
http://www.TheGamerNation.com/Forums
He didn't do it this way; too ambiguous.
He developed his own method, whereby each digit was assigned to a row of hiragana (e.g. 1 = anything from the ka line, 2 = anything from the sa-line, etc.) and built up a memorizable string of words using those sounds.
Still means he had to memorize a 100,000 syllable story, though.
I saw this linked above by an AC. http://www.asahi.com/national/update/1004/TKY20061 0040185.html Its an article in the Asahi Shinbun about the feat. My brief non-literal translation follows (if its inaccurate, sorry in advance, for accurate translations you can pay me my hourly):
"Using equivilence rules like 3 = sa [n.b. all numbers in Japanese have a variety of syllables which they can be read as -- thus, you can remember a phone number as roughly a two to three word phrase, like my bank being 555-GOT-MONEY], you can memorize the first N of the infinite digits of pi by constructing a story of sufficient length and memorizing that. His previous record was seven years ago.
After reciting the 100k digits they were checked against a computer printout. Mr. Haraguchi then retired with his family. They brought him his favorite beer, which he proceeded to chug. He commented 'Its good that I was able to relax'*"
* This is ambiguous in Japanese: my guess is he is referring to his ability to have been relaxed while reciting the digits, but eh, doesn't really matter either way.
By the way: my back of the envelope math suggests 100k digits of pi would leave you with a Japanese text about a tenth as long as the Bible, give or take. So its neither impossible nor a mean feat to have memorized a text of that length.
Help poke pirates in the eyepatch, arr.
Just remember 355/133 (3.1415929...). It's the most accurate fraction possible with only a three-digit numerator and denominator.
It's not that accurate: 355/133 = 2.66917... . I think maybe you meant 355/113.
actually, that is the most accurate fraction with a denomenator 10,000. When you go to 5 digit denomenators the best is 312689 / 99532 with 6 digits : 3126535 / 995207 with 7 digits : 5419351 / 1725033 with 8 digits you hit the magic 245850922 / 78256779. the error is -7.8179366199075e-17 so the difference is zero in double precision arithmatic. -Rob
Umm, no it doesn't. 0.5 is 1/2, just like 0.50 or 0.500. It only contains 1 decimal place. Think of the number 7. You could call that number 07.0, but that doesn't make it any less correct to just call it 7
I can easily remember pi up to that precision. My first "good" calculator had as 3.1415927, and I thought that was pretty odd, because Commodore 64 had one more decimal place: 3.14159265.
And you could type the pi symbol in C64, and it served as the constant's name. Haven't figured out how to do that in PC, aside of setting kb layout to greek, which is kind of complicated... =) (Argh, Slashcode seems to filter π, too. Welcome to year 2006, character sets are still a big problem in computing!) And people are complaining about proposed Unicode operators in Perl 6 - heck, in Commodore 64, not only we had non-ASCII operators (up-arrow for exponents), we had non-ASCII constant names, for pi! Er, PETSCII constant names. =)
I just tried to remember more decimal places than that. I guessed "3542", but apparently I got the last two wrong (should be "89"), I can't possibly guess why, as I'm not that huge fan of Douglas Adams and I'm running out of ideas besides that. Ah...
> Memorizing large amounts of random information has more to do with effective
> mnemonic techniques and capacity for intense concentration than base retentive
Innate ability does help, but mostly memorization is a learned skill.
I should know. I'm heavily involved with a quizzing program -- not the random-trivia type of quizzing, but the sort where they give you a particular text to study for a few weeks and all the questions come from that text. Naturally the best quizzers memorize the whole material. At the national level, practically all of the quizzers memorize the entire material, and the best ones know things like how many times certain words are used. (I know several of these best quizzers personally. I happen to live in the North Central Ohio district, which takes first place in the nation more often than any other district in the program.)
So as I said, innate ability helps, and especially it helps inherently somewhat brighter people to get "into" the program more quickly and get started more easily, but the quizzers who take home the trophies are not the ones who started out smartest; they're the ones who studied most and/or the ones who have been quizzing seriously for the most years. With *rare* exceptions, new quizzers don't start out memorizing. At first they read over the material every day and try to learn enough of it to answer some of the questions. Because certain sections of the material are designated for especial memorization, most quizzers eventually reach the point where they start memorizing at least those sections. After they've done a few, it gets easier. Nobody memorizes *all* of the memory sections the first time he memorizes everything. They start out with a few, and work their way up, until they're learning all of them. Like I said, it gets easier after you've done more of it. Once they reach the point of having memorized all of the memory sections, it's about a quarter of the whole material, so they start to realize then that they could just flat memorize all of it. In NCO we try not to put anybody on the district team (i.e., the team we send to nationals) unless they've reached the point where they're willing to seriously attempt that.
The quizzing program doesn't just teach memorization (although the learning of the text is an important goal for pretty much everyone involved). Because the quizzers have to answer questions, not just recite, and because they have to *finish* the question before they can answer it (assuming the quizmaster didn't already finish it, which at the higher levels of quizzing he almost never does, because invariably the level of competition is such that the quizzers are off their lights before the question is half done), there are other kinds of mental activity as well.
Incidentally, the quizzers in this program are kids and teens. I'm far too old to quiz these days, but I coach and I serve as quizmaster (i.e., ask the questions) and sometimes judge. I did quiz when I was younger, and I was on the district team in 1992.
Cut that out, or I will ship you to Norilsk in a box.