Reply With QuoteThey could have at least printed the number in the article.
...from memory!
http://www.msnbc.msn.com/id/8456677/
They could have at least printed the number in the article.
I can't even begin to imagine how it's possible to remember something like that, much less recall it with that much insane accuracy. I'm sure I'd have a problem just counting up to 83,000 without losing my place or drifting off to sleep somewhere along the way. This guy's probably a genius that will now fail to do something monumentally important to the world of science since he has devoted so much of his brain to this fantastic feat.
.... and he gets his underwear at Kmart. After watching Wapner! :lol:
They forgot.Originally Posted by nutsak
Imagine being the poor bastard who has to check for accuracy.
I'm not saying the feat isn't an impressive display of memory, but it hardly qualifies him as a genius. He divides quickly. Wow. (He also lost his place 3 hours after he started)
I was wondering about this. I've never looked into methods for calculating pi, but you're saying you can calculate as you go? I mean it's not progressively larger calculations for each digit.
That seems more plausible, yet news reports of this sort of thing always say they do it from memory.
There are no known algorithms that spit out pi's decimal expansion digit by digit (bizarrely, one is known that spits out its *hexadecimal* expansion). If you wanted to produce the decimal digits by computation, you would have to choose some series expansion, but this would necessarily entail keeping track of a working area of many digits, and only producing a digit when your series terms had got beyond that point.
Everyone who does this kind of thing does it by memorizing. Excise the 'He divides quickly' from Bill's comment, and I agree entirely :)
"Say," I asked the chauffeur, "you know the value of pi?"
"You mean that 3.14 whatzit?"
"That's the one. How many decimal places do you know?"
"I know it to thirty-two places," the driver tossed out. "Beyond that, well . . ."
"Thirty-two places?"
"There's a trick to it, but yes. Why do you ask?"
"Oh, nothing really," I said, crestfallen. "Never mind."
.... er, why wouldn't this work, then?
step 1: get a digit (or two) from the hex expansion
step 2: convert hex -> decimal
That falls under 'keeping track of a working area of many digits, and only producing a digit when your series terms had got beyond that point'.
Anyway, I don't think the hex-digit-producing algorithm is particularly well-suited to mental calculation.
I'm pretty sure most of these digit memorizing people memorize a story in which each word has as many letters in it as the current digit of the number.
e.g.
See, I knew I could elucidate!
=
3.14159
I like apple pi.
Shut your pi hole.
This is an absolutely amazing story, and almost hard to swallow. Cog psy is the bread and butter of my studies and false memory is my focus therein. i can tell you with out doubt, what this man did is next to impossible. All humans take information and transfer it through no less than 3 stages of memory to get that information to a long term storage place. The first is sensory memory (SM), and lasts a very fleeting 1 or 2 seconds at the outside. After moving through SM the information reaches short term memory. This is where memory becomes a challenge for most people. The average shortterm memory (STM) can only process 7 + or - 2 pieces of information and lasts for 3 - 20 seconds. The STM is also subject to interruptions.
The most accepted way to move information from STM to Long term memory (LTM) is by assigning meaning to them, and integrating them into Schemata, or interrelated bits of information used to form mental models of our perception of the world. This is why songs and other mneumonic devices work for us we make them important to us. Imagine how hard it would be to move 83,000 bits of information to LTM if all you have to do this is about 7 bits of working space. Humans have a knack for getting bigger stuff into LTM or we would all be as smart as a lizard. Usually we do this through clumping. Take phone numbers for instance. We clump the area code, the prefix, and the number. This is how most people remember a 10 digit phone number.
Even for person that falls on the lower end of the spectrum and as a STM 5 bits big, they are remembering 3 clumps not the whole set as individual digits. This however still does not account for this guys amazing ability. Taking the above clumping theory the man would have memorized about 9222 clumps of information. That is absolutely amazing even if you are not getting into the information retrieval side of the equation. Remember the Schemata? We clump related information together to make it easy to get too. This is why leading questions will make some people remember and why smell can make you remember seemingly unrelated events. I would personally like to know how the hell you assign sufficient meaning to 83,000 numbers or 9222 clumps so that you can remember them.
Wow. That is frightening similar to an essay I wrote the Memory chapter of Psych.
What he did is amazing even taking into consideration chunking, as you noted. But it is certainly possible. I read an example of a man who memorized some 1,000 digits and he said them back in groups of 3 or 4. When asked to explain how he did it, he mentioned "Well, the first group is my 3-mile time, the second is 10-mile time, then someone's age, someone's age, etc."
Still, over 9,000 of these tidbits? That's impressive. I certainly consider him a genius.
Liz/AttAdude: Yeah, it's impressive, but it's not like he's regurgitating a string of 83,000 random digits after a single exposure. He's clearly put a LOT of time into studying/practicing.
This is far from what the typical person can do, though. I'm not sure if a typical person could do the same with an equal amount of time spent practicing.
See: "After 200 hours of practice, two Carnegie-Mellon University students were able to increase their memory span from the typical 7 chunks to more than 80."
But he's not increasing his ability to "chunk" up to 83,000 -- if he was, he'd be able to repeat back any random list of 83,000 digits you gave him. Rather, he's memorizing a fixed, known string of digits. Almost anyone could memorize, say, 100 digits with a bit of practice, (actually, almost everyone has -- quick, what's 8 x 7?), so it's not at all inconceivable that someone can do 83,000.
It's perhaps more remarkable that there's someone who's not only able, but eager to do so -- I was surprised that he appears to have a normal job, and isn't some sort of Rain-Man-esque savant. Talk about a boring hobby!
Oh, I didn't know that. In which case, *that's* the real story here - "savant does savant-y stuff" isn't much of a news item, after all. Mind you, he is Japanese, and they're just plain weird :lol:
You mean, like our lungs? Oh, you meant "mnemonic." And you left out sleep. Sleep assists in LTM formation. Maybe you haven't gotten to that chapter yet in your CogPsych book.
Animals. Food recipes. A city map. A world map. Again, it's not easy. However, it's also not I Can't Believe It's Not Impossible! either.
I think i have read about that one. Was there a follow up done to see if the change was lasting?
Bill:
What did you want a full blown dissertation on memory acquisition? My intent was to hit the salient and most widely accepted parts in an effort to show why this guy is so amazing. If you would like a treatise on genius level cognitive process i would recommend a semi recent study on how chessmasters process the game.
Aside from that if you guys actualy think this is so easy, why dont ya try it. You do realize the previous record holder was only able to do about half of what he did. I would be willing to wager my PC that there is not a single person here that could do 5000 even after 3 months of study. I have doubts one of us could even do 1000. My own memory tends to lean twords the episodic so i know for a fact i could not do it.
Well, how do you know that he can't after equal amounts of practice? Pi's digits are rather random to most people. Nonetheless, I certainly see your point.
Not that I read. Something to look up, though.
Oh, I don't think it was easy -- memorizing 83,000 digits would be a pain in the ass. However, I don't boggle at it quite the way you do, either. I am confident that I could learn a given string of 1000 digits if I had sufficient motivation. Would take a while, sure, but definitely doable.
Ever played Simon?
Because he'd be showing off THAT ability, rather than just regurgitating digits of Pi? The ability to repeat 80k digits error-free, after seeing/reading them only once, now THAT would be staggering.Well, how do you know that he can't after equal amounts of practice? Pi's digits are rather random to most people. Nonetheless, I certainly see your point.
That said, I wouldn't be shocked if a savant of some sort actually was found who could do that (or similar) eventually. Human brains can do some wild, wild shit.
Will there ever be born a boy who could swim faster than a shark?
Actualy i have played simon. It is a little different however, because it uses a different type of memory than memorizing 83000digits.
Are you saying my custom built gaming rig is not good enough reason to put this to the test? 5000 digits from memory and you get a new pc what better motivation could a geek need?
:P
No, it's not good enough reason -- it would take many many many hours to do, enough so that I'd be better off getting a part time job and buying a new PC. Give me six months to a year (or thereabouts), and I'd be able to do 'em.
But, yeah, I'm not willing to follow through on my conviction, so, may as well quit arguing -- I'm willing to forego the mudslinging stage if you are.
"Hitler!" There, we're done :)
FUCK YES KNOCK ME DOWN WITH YOUR ACUMEN COLLIJ BOY. Actually, no, but you could have not completely ignored the bits that might have contradicted your claims that this feat is "nearly impossible."
I already know about genius level cognitive process, Buddy. I'm living it. Seriously though, whatever your intentions may have been, you coincidentally left out anything that might have contradicted your legitimate right to declare awe. This old Jap fella, he's been memorizing pi for fucking years. When you were getting trounced in tetherball during recess, he was memorizing pi. When you were begging for a passing grade in your high school Composition class, he was memorizing pi. When you were taking your sister to the prom, he was memorizing pi. When you were floundering through your Psych Phys class, he was memorizing pi. Last December? Pi. Easter? Pi. Today? I bet that crazy motherfucker is still memorizing pi. "Pi 'til I die," he's got a T-Shirt or screensaver somewhere that says that, you can bet on it (your PC, even!).
Why don't you put her in charge? If you listed all the things you cannot do, you'd at least be close to the target number. Incidentally, nobody said it was "so easy." Just not "next to impossible." Next to impossible is throwing up the components of a wristwatch and having it land assembled and wound, or something.
5000 in 3 months, gee thanks. I'd get to 83000 in less than 5 years. I'm wagering Senor Pi-pants took a bit longer. 1000? 3 months? For your PC, estimated value somewhere around 1-2 thousand bucks?
3 months, to memorize this:
3.1415926535 8979323846 2643383279 5028841971 6939937510
5820974944 5923078164 0628620899 8628034825 3421170679
8214808651 3282306647 0938446095 5058223172 5359408128
4811174502 8410270193 8521105559 6446229489 5493038196
4428810975 6659334461 2847564823 3786783165 2712019091
4564856692 3460348610 4543266482 1339360726 0249141273
7245870066 0631558817 4881520920 9628292540 9171536436
7892590360 0113305305 4882046652 1384146951 9415116094
3305727036 5759591953 0921861173 8193261179 3105118548
0744623799 6274956735 1885752724 8912279381 8301194912
9833673362 4406566430 8602139494 6395224737 1907021798
6094370277 0539217176 2931767523 8467481846 7669405132
0005681271 4526356082 7785771342 7577896091 7363717872
1468440901 2249534301 4654958537 1050792279 6892589235
4201995611 2129021960 8640344181 5981362977 4771309960
5187072113 4999999837 2978049951 0597317328 1609631859
5024459455 3469083026 4252230825 3344685035 2619311881
7101000313 7838752886 5875332083 8142061717 7669147303
5982534904 2875546873 1159562863 8823537875 9375195778
1857780532 1712268066 1300192787 6611195909 2164201989
1 month:
3.1415926535 8979323846 2643383279 5028841971 6939937510
5820974944 5923078164 0628620899 8628034825 3421170679
8214808651 3282306647 0938446095 5058223172 5359408128
4811174502 8410270193 8521105559 6446229489 5493038196
4428810975 6659334461 2847564823 3786783165 2712019091
4564856692 3460348610 4543266482 1339360726 0249141273
7245870066 0631558817 4881520920 9628
1 week:
3.1415926535 8979323846 2643383279 5028841971 6939937510
5820974944 5923078164 0628620899 86280
1 day:
3.1415926535 89
Your episodic memory need only get through an episode about a dozen or so digits long every day to make it. You've no idea of the power you have, McCloud! Sure, it builds up, but part of my point is that every day goes by is another sleep cycle, another big dump into long term memory. The key to this whole feat is LTM, as you said. Except, that's the winning key (in terms of sleep contirbuting to LTM), not the defeating caveat as you imply. Still not as fun as getting punched in the face, but not nigh impossible. There are some very bright people around here.
Also, 1000 per 3 months gives me a little under 23 years to get to 83K. Ol' Akira is 59. He's a tenured psych prof. Now, you of all people know how much free time those guys have. The dude he (allegedly - neither this latest nor his prior effort have been verified as correct) beat was some college kid. Now that kid was cool.
This is just like Contact, the book, just different.
--- Alan
Posting Permissions