Usain Bolt Is No Takeru Kobayashi

DESCRIPTIONPhoto: laverrue

I blogged last week about how progress in lowering the world record in the 100-meter dash has been extremely slow, even with the improvements in track surfaces, training techniques, steroids, etc. The world record has been lowered at an average of 0.1 percent per year over the last 40 years.

Compare that kind of progress with the revolution that Takeru Kobayashi started in competitive eating. The Nathan’s 4th of July hot dog eating contest is said to have started in 1916 with the winner eating 13 hot dogs that year. By 1978 the record was only up to 17 hot dogs, according to Wikipedia. That rate of progress is not so different than what has been observed in sprinting. By 2000, the record had been stretched to an incredible 25 hot dogs.

In 2001, Kobayashi shows up and eats 50 hot dogs! He doubles the world record. He reigns supreme for six years before Joey Chestnut shows up, and he and Kobayashi both shatter the record with 66 and 63 hot dogs respectively. In this year’s event, Chestnut somehow manages to down 68 hot dogs and buns in 10 minutes.

What is so interesting about this competitive eating example is that, like running, humans have been eating forever. There is no particular reason why people should suddenly be getting so much better at it. A reasonable person might have argued 20 years ago that eating 20 hot dogs in 10 minutes was bumping up against human limitations of stomach size. There was little or no room for improvement. And yet these guys are able to quadruple the world record that stood in 1978. Truly amazing.

Competitive eating was the activity I had in mind when I posed the quiz. As usual, it didn’t take long for a blog reader to get the answer; just a few minutes after the post went up, a reader named Josh wrote “eating hot dogs really fast” to be the first-prize winner.

I also said I would give a prize to the best answer other than competitive eating. There were many great examples given, ranging from mountain climbing to female marathon running to racial integration of the work force to domino tumbling.

But the activity that most captured my awe is something I had never even heard of called “piphilology.” Believe it or not, the point of this activity is to memorize and recite the digits of pi. Before reading further, stop for a moment and take a guess at how many digits of pi people have been able to memorize.

Remarkably, the world record in 1973 stood at 930 digits, according to a page devoted to this activity.

But that was mere child’s play. By 1977, the record was up to 5,050. By 1980, it was 20,013. By 1987, it was 40,000. The current world record is 67,890. It took the gentleman from China over 24 hours to recite those 67,000+ digits.

Now a Ukranian doctor claims to know the first 30 million digits, although he has not yet had the chance to recite them in order. If he went at the pace of recitation that the current record holder used, it would take the doctor over a year to get to the end. It is said, however, that when asked for specific digits he could deliver them.

I’m grateful to blog reader John C. for mentioning this fascinating example. He also wins Freakonomics swag, although perhaps he deserves punishment instead of accolades. I had been feeling pretty good about myself for finally having learned my 16-digit credit card number after roughly a decade of trying (my wife, tellingly, had learned the number many years earlier).


For all the people who want to cry foul about Usain Bolt, there are just as many accusations about modern competitive eaters. There is the possibility that surgery and/or drugs have allowed them to go beyond the bounds of normal human limits.


Time in the 100m is a different beast compared to both hot dogs and pi. Theoretically, there are no limits on the number of digits of pi one could memorize and the same is true for hot dog eating (though probably much more theoretical). Conversely, the fastest one could run the 100m is bounded by zero seconds. I would argue that making gains in a bounded activity is the more impressive feat.


There was a step-change in hot-dog eating. It was the soaking of the buns. Kobayashi dunks the buns in water so they aren't dry anymore and much easier to eat. Once that barrier was broken it opened the door for records to keep advancing. He also made it so popular that other people realized that you had to train to eat that much. I think before that people just showed up hungry.


I would guess that things like hotdog eating have exploded because the potential was always there, but no one was capitalizing on it. Kobayashi revolutionized the activity by eating the hotdogs separately from the buns. The change in technique demonstrated that the choke point of progress (pun fully intended) was not in the stomach, but much earlier in the alimentary channel.

But both competitive eating and feats of memorization were stuck at the same limitation for many years: Who wanted to bother training for these things? Why would anyone do it? It needed to be a competition of some sort before large numbers of people would focus effort on figuring out how to do it better.

By comparison, running is something humans have been doing for hundreds of thousands of years, and since survival depended on it, virtually all humans have done it. Our forebears got the basic techniques down long before anyone made a sport out of it.


Mike B

I think that the competitive eating feat may be somewhat bias by the fact that around 2000 it turned from an activity into an honest to god sport. At some point in the recent past eating contests morphed from carnival side show attraction featuring local overweight individuals to a sport where competitors undergo years of scientific training regimens and then compete full time on a circuit for large amounts of money. There were also many breakthroughs in technique that delivered the same performance boosts as the Fosbury Flop or fiberglass pole vaulting poles.



What makes you think that there are no theoretical limits to hot-dog eating, or pi-memorizing? It's not as if our bodies have unlimited volume, or our brains unlimited memory. Both are capped at some value - and depending on how you personally represent digits in your head, it could be a very hard bound too. Though I do admit that there seems to be much more leeway in the case of pi-memorizing.

As for the Ukranian doctor, his ability to give digits upon request reminds me of PCP's from the field of Computational Complexity - but that's another story altogether.


The hot dog eating can be easily converted to a bounded feat. How long does it take to eat 50 hot dogs? As of 2001 it took 10 minutes. Prior to that, a lot longer. Currently, less. Conversely, we could convert the sprint to "How much distance can be covered in 10 seconds?"

I don't see a lot of difference.


Now I'm hungry....


Asking for specific digits of PI is the clue that another technique may be in use since there are formulas that provide the digits at arbitrary positions. If you practiced this computation and were exceptional at mental arithmetic, you could pretend you have the digits memorized and recite them slowly for the rest of your life.


Also, both those exercises are benefiting primarily from new interest. Who cared about competitive eating until ESPN started broadcasting it and other channels starting pushing similar style shows? Who ever cared about Pi?

People have been trying to break the record in the 100M for ever. Usain Bolt is the most successful at it, where hundreds of thousands of others failed. How many people were actively trying to break the hot dog eating record? Not many, until recently, and they seemingly all have succeeded. The same can be said for Pi digits. If a record can be broken every time someone simply applies himself to breaking it, how impressive is that?

Running approached it's asymptote far sooner than either of these modern, niche activities because it's been going on for far longer. We won't continue to see the exponential rise of hot dog eating and pi digits in the future. You are comparing two activities in their infancy to one in its maturity. I won't deny that the facts are accurate. But the assertion that one is more impressive than the other is a farce.



Rich - actually the record for running the 100 meters is bounded by the speed of light. If an althlete ran faster than that he'd almost certainly fail a drug test.

The doctor has the right approach I'd have to say. If you think that he has memorised 30 million digits of pi then the probabilistic (statistical) approach to verifying it is the most sensible one. If he claimed to be able to recite all (that is, any) of the digits of pi, then it's you only choice.


Nutjob is right, saying that the 100m is bounded by the speed of light. The fastest possible time is 3.33564095 × 10^-7 seconds.

However, if a person beat that speed, they could trivially go back in time to alter the results of their drug test.


Well, to say "people have been eating forever" is a little misleading. What Chestnut and Kobayashi do is a far cry from actually eating.


"Asking for specific digits of PI is the clue that another technique may be in use since there are formulas that provide the digits at arbitrary positions."

No. There are formulas that approximate pi to arbitrarily high degrees of precision but they don't give you a short cut to say, the millionth digit.


Although impressive, I still think the advances in piphilology seem to skirt the rules of the contest. Modern computing allows anybody to see/study the digits of pi to their hearts' delight. In the 1970s how many people even had access to the information necessary to compete?


Jason, I agree that you're right about me being right, but what if in going back in time to alter the drug test, he inadvertently was late to the race, it started without him, and he never ran it?



His point is that the race isn't framed as "How much distance can be covered in 10 seconds?". It's "How fast can you run 100 meters." Wording it like that changes the nature of the event to something it isn't.

The only bound you can implement on hot dog eating is stomach capacity, but we clearly haven't reached the human limits (since records are still being broken).


" ... humans have been eating forever. There is no particular reason why people should suddenly be getting so much better at it."

I hardly call shoving dozens of hot dogs down your throat over a short period of time "getting better" at eating. This whole topic is totally disgusting.


The pi list (which includes two of my former math students) seems to go down quite far, the bottom memorizers boasting only 20 or so digits. That seems like it would be the piphilological equivalent of about three hot dogs or a 17-second hundred meter dash.


How has the evolution of the 100M record time compared to other track and field events?