Improving this recommendation system is the focus of the prize competition. The prize announcement explains that, " Netflix is all about connecting people to the movies they love. To help customers find those movies, we've developed our [proprietary] world-class movie recommendation system...[but] if there is a much better approach it could make a big difference to our customers and our business..."

Prize: $1 million.

Records in the research dataset: 100 million+

If I had tenure already, I would be psyched to work on this problem, not just for the possibility of the fame and fortune but also for a chance to see what one might find in a dataset that big. Indeed, it's a bit surprising that there has until now been no mention of it in OR/MS Today and no one at the most recent INFORMS annual meeting in Pittsburgh seemed to be talking about it — even though the prize was publicly announced on Oct. 2, 2006.

Anyway, I was still thinking about this prize when I found an unread book entitled "Longitude" [1] on my shelf. When I finally started reading it over the holidays, I couldn't put it down.

Early sea voyagers experienced many harrowing incidents and tragic accidents that were all attributable to the difficulty of accurately determining a ship's longitude (and thus one's actual location). Under pressure from merchants and seamen in the early 18th century, a royal commission including Sir Isaac Newton implored the British Parliament to encourage the development of solutions to this vexing problem and "to reward success handsomely" [2]. The result of this commission's efforts was the Longitude Act of 1714, which offered enormous sums of money for the development of a reliable and repeatable method for determining the longitude at sea (cash prizes were on a sliding scale based on the level of precision achieved by the longitude estimates).

From the outset, there was a subtle but significant bias towards using astronomical techniques such as the plotting of the moons of Jupiter or the estimation of the distance between the moon and the sun by day to determine longitude. Subsequent holders of the "astronomer royal" post worked incredibly hard on measurements and calculations intended to provide a mathematically elegant solution to one of the greatest problems of the day.

To calculate longitude at sea, one needed to know the time on board the ship, the time at some fixed point on land, and the difference between the two. This problem, which seems trivial to citizens of the modern world, was "utterly unattainable up to and including the era of pendulum clocks. On the deck of a rolling ship, such clocks would slow down, or speed up, or stop running altogether" [3].

To us, the solution seems obvious: build a better clock! Yet, remarkably, the renowned watchmakers of the day never did manage to do so. Instead, the longitude problem was ultimately solved by John Harrison, a previously unknown man who became obsessed with the quest. He had no formal training, no apprenticeship with an established watchmaker, and no reliable means of support beyond small grants from the Longitude Commission and small loans from friends. Yet, over the course of his amazing lifetime, this self-taught craftsman and perfectionist produced an amazing series of five sea clocks, each of them more precise than the last, and all of them far more effective than the celestial approaches investigated by men such as Galileo, Newton, Halley and the great astronomers who followed them.

For his trouble, Harrison was abused repeatedly by those later astronomers, who as it happened, were key members of the Britain's Longitude Board. His clocks were subjected to a variety of sometimes arbitrary tests that were conducted carelessly, at times even maliciously. As soon as one of his clocks would pass one test, his enemies would dream up another.

A talented and tenacious artisan and innovator who "developed a series of virtually friction-free clocks" [4], Harrison was nevertheless no match politically for his enemies, who eventually forced him to explain his innovations in detail, after which they took possession of his clocks, discredited what he had done, and left his lifetime's works to decay.

Only an impassioned plea to the king enabled an aged and beaten down Harrison to finally receive the recognition and part of the compensation that was due to him for what he had achieved. And only thanks to a fortuitous circumstance were the five Harrison sea clocks rediscovered and restored in the early 20th century. Once restored, they continue to run to this day.

In endowing their prize, Netflix appears to be cognizant of history. The company has gone to great lengths to make the competition's rules clear and unambiguous and have publicly spelled out the measurement methodology in great detail (www. Netflixprize.com/rules). Anyone who reads "Longitude" will fervently hope that, unlike the 18th century British astronomers who tormented the heroic Harrison, the Netflix folks adhere to their own rules.

And, what the heck, if John Harrison's experience is any guide, then perhaps my tenure clock isn't really going to be an impediment to participation in the Netflix Prize competition after all. If no one wins any time soon, I'll get started in a couple of years. Anybody care to join me?



Vijay Mehrotra (vjm@sfsu.edu) is a faculty member in the Decisions Sciences Group in the College of Business at San Francisco State University and an operations management consultant.

References

1. Sobel, D. 1995, "Longitude: The True Story of a Long Genius Who Solved the Greatest Scientific Problem of His Time," New York: Penguin Books.
2. Sobel, p. 53.
3. Sobel, p. 5-6.
4. Sobel, p. 10.

Table of Contents

OR/MS Today Home Page


OR/MS Today copyright © 2007 by the Institute for Operations Research and the Management Sciences. All rights reserved.


Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com


Web Site © Copyright 2007 by Lionheart Publishing, Inc. All rights reserved.