Mathematician Daniel Tammet counts pi to 22,500 digits

[ElliotSmith]

Summary: wow!

Mathematician and autism-spectrum savant Daniel Tammet counts pi to 22,500 digits.

 

[MathematicalPhysicist]

Good for him, the next step is 22,501.

 

[mfb]

Good for him, the next step is 22,501.

He did 22514 ;)

 

[Spinnor]

Truly amazing. I have read that people can improve their memory and use certain tricks to make it better. But this just seems to be "God given" perfection, no tricks involved?

 

[Greg Bernhardt]

The mind can be so powerful!

 

[Filip Larsen]

I would think a next step of 31415 digits will be kind of cool ...

 

[MathematicalPhysicist]

He did 22514 ;)

Then off you go to 22515, can he do all the infinite digits?!
It will take some time though...

 

[Klystron]

The mind can be so powerful!

Agree! Without offering proof I do not think there has to be a connection between his mathematical ability and a diagnosis with autistic spectrum disorder (ASD).

From a medical point of view this correlational reminds one of the myth that the visually impaired develop extraordinary hearing ability rather than, lacking visual cues, the blind concentrate more on their remaining senses for orientation and navigation. Sighted people can also learn to clap their hands and listen for returns from surrounding structures or feel ground surfaces through soft footwear, for example, as well as the visually impaired but lack the necessity. A similar argument pertains to touch-reading Braille and learning sign-languages.

There are many people with ASD with normal ability at mathematics and superb mathematicians with normal or well developed communication and social skills and no diagnosis of autism.

 

[cosmik debris]

Can he tie his shoelaces?

Cheers

 

[epenguin]

But how long has it taken him? At this rate to finish the job might take for ever!

 

[phinds]

Agree! Without offering proof I do not think there has to be a connection between his mathematical ability and a diagnosis with autistic spectrum disorder (ASD).

There are many people with ASD with normal ability at mathematics and superb mathematicians with normal or well developed communication and social skills and no diagnosis of autism.

But are there any normal people who can do the kind of amazing math stuff that ASD people can do? I've never heard of any.

 

[StatGuy2000]

But are there any normal people who can do the kind of amazing math stuff that ASD people can do? I've never heard of any.

First of all, it's worth keeping in mind that the overwhelming majority of people on the autism spectrum do not have any special ability with respect to mathematics.

According to the Guardian article below that states that an estimated 10% of the autistic population and an estimated 1% of the non-autistic population have savant abilities of the type mentioned in the original post (note: citations needed for the estimates mentioned in the article). That means that an estimated 90% of the autistic population do not have such abilities.

https://www.theguardian.com/theguar...It's mental imagery.,no one knows exactly why.

Second, the article mentions that an estimated 1% of the non-autistic population have savant abilities. While that does not sound like much, keep in mind that the population size of non-autistic people is orders of magnitude larger than that of the autistic population. Which means that there are many more people (in terms of absolute numbers) who have savant abilities who are non-autistic.

 

[Vanadium 50]

1. I don't think I am comfortable defining people not "on the spectrum" as normal, and by implication those who are as abnormal.
2. Memorizing a pile of digits is not mathematics, and I can't see any obvious correlation between these two skills.
3. Memorizing 22,500 words-that-happen-not-to-be-digits is hardly unprecedented. The fact that we have Homeric poems today is a fortunate example of this.
4. I have known many mathematicians and mathematics professors who are not "on the spectrum". It is neither a necessary nor sufficient condition.

 

[TeethWhitener]

Memorizing 22,500 words-that-happen-not-to-be-digits is hardly unprecedented. The fact that we have Homeric poems today is a fortunate example of this.

Also of relevance:
https://en.wikipedia.org/wiki/Hafiz_(Quran)

 

[collinsmark]

For what it's worth, btw, 22,500 digits, while quite impressive, is not a record.

The current record is just over 70,000 digits:
http://pi-world-ranking-list.com/index.php?page=lists&category=pi&sort=digits

(for awhile, I used to hang out with/study with #8 on the list.)

[Edit: If you're curious, I, myself, am able to memorize about 6 digits.]

 

[phinds]

Memorizing a pile of digits is not mathematics, and I can't see any obvious correlation between these two skills.

I think you misunderstand what they are doing. They do not memorize anything. They state the digits as they compute them. That IS math, even if only arithmetic.

EDIT: see below. I had this wrong. It IS memorization.

 

[collinsmark]

I think you misunderstand what they are doing. They do not memorize anything. They state the digits as they compute them. That IS math, even if only arithmetic.

No, it's definitely memorization.

Rajan Mahadevan explained to me how he memorized PI to over 30,000 digits. He associated each digit [from 0 to 9] with a set of objects or actions and then used those objects or and actions to form a story. After that, it was just a matter of memorizing the story.

I doubt there's any of these record holders who mathematically (arithmetically) generated digits of PI in their head, on the fly.

[Edited for clarity]

 

[phinds]

No, it's definitely memorization.

Rajan Mahadevan explained to me how he memorized PI to over 30,000 digits. He associated each digit with a set of objects or actions and then used those objects or actions to form a story. After that, it was just a matter of memorizing the story.

I doubt there's any of these record holders who mathematically (arithmetically) generated digits of PI in their head, on the fly.

Interesting. That had not been my understanding. Are you sure there are no cases where they just say the digits as they compute them? I was sure that I saw that mentioned in a science program (but it WAS a pop-sci TV program and we all know how reliable those are).

Also, are you saying that they don't even compute the digits themselves?

 

[collinsmark]

Interesting. That had not been my understanding. Are you sure there are no cases where they just say the digits as they compute them? I was sure that I saw that mentioned in a science program (but it WAS a pop-sci TV program and we all know how reliable those are)

I can't say for certain because I've only known one record holder (very limited sample size).

But I'm confident enough that I'd bet the farm that it's all memorization.

 

[phinds]

I can't say for certain because I've only known one record holder (very limited sample size).

But I'm confident enough that I'd bet the farm that it's all memorization.

Yeah, I've been poking on the internet and it certainly appears that you are right and I had it wrong.

Sorry, @Vanadium 50, you had it right and I had it wrong.

 

[mfb]

There is no way to compute these digits on the fly. The first few maybe, let's say 10 to be optimistic or 20 to be extremely optimistic, but not thousands.

 

[Vanadium 50]

There is no way to compute these digits on the fly

Actually, there are algorithms that can determine the n-th digit without knowing the previous n-1. This is what they use to test computers that calculate pi to a zillion digits - otherwise, how could you tell that the output is pi? However, as far as I know, none of these are in base 10.

However, what we're seeing here is memorization.

 

[Vanadium 50]

There is a base 10 algorithm. See http://stanleyrabinowitz.com/bibliography/spigot.pdf

 

[mfb]

I know there are algorithms, but there is no way to run them in the head. They need about as many intermediate digits as you get digits of pi - or more.

 

[collinsmark]

Another memory demonstration that Rajan Mahadevan would like do that I found very impressive was as follows: He would have someone write down a 10x10 matrix of single digit numbers, in whatever random fashion the writer chose, on a piece of paper (or napkin, or whatever was lying around). 'Something that looked similar to this:

9 6 7 7 4 0 2 2 3 9
0 8 9 2 2 9 3 1 6 3
3 2 8 0 4 8 5 9 9 3
2 0 7 5 6 9 6 4 4 1
1 2 5 1 9 6 4 0 2 9
3 9 5 6 4 7 5 3 2 9
4 3 1 0 9 0 1 1 2 7
1 0 2 8 4 9 5 5 8 0
4 9 8 3 2 9 4 0 3 2
2 9 0 3 8 5 0 3 8 9

He would then take the paper and concentrate on it for maybe a couple of minutes -- and I'm not exaggerating on the time, it was typically no more than a couple of minutes.

He would then give the paper back. When everyone was ready, even it was after a small break, he would rattle off the numbers, in order, row by row.

Then he would speak the numbers backwards, starting from the last number of the last row back up to the first number of the first row.

Then he would rattle off the number by column, instead of row.

Finally he would give you the diagonals.

That was pretty impressive.