Monday, February 13, 2006

LHC olympics: big success

We were informed that the LHC olympics were fabulous. Skeptics have been converted, and almost everyone will try to participate in the next rounds. Those who thought that no one would solve anything and no one would learn anything new have been proved wrong.

In the case you don't know, LHC olympics is a competition in which various teams face the challenge to determine the right model from a set of fictitious raw data from the LHC. This setup emulates the conditions that people may face as early as in 2007.

There were three boxes - in this case, all of them were supersymmetric models - that were supposed to be determined:
  • Harvard box (by Nima Arkani-Hamed)
  • Michigan box (by Gordon Kane)
  • Seattle box (by Matthew Strassler)
The teams that constructed the boxes obviously tried to solve the remaining two only. There were three more teams participating:
  • Princeton University
  • John Hopkins University
  • Cornell University
The contest was slightly unfair because Harvard's team was crowded with stellar phenomenology graduate students. This includes Philip Schuster and Natalia Toro who have created a powerful Mathematica package that everyone else was using - as well as Can Kilic, Jesse Thaler and others. There were no medals awarded but if they had to be distributed, the golden medal was not really in question because the Harvard team has used intellectual brute force to nail down both models, including parameters with a percent-like accuracy. If the medals were a part of the olympics, the two valuable medals were:
  • Gold: Harvard University grad students
  • Silver: Princeton University string theorists
Local string theorists in Europe went to see the event because their colleagues Herman Verlinde and Leonardo Rastelli of Princeton University were participating, and they did a really impressive job given the difficult initial conditions. The bronze medal, if it had to be given away, would probably be tied. Matt Strassler did a heroic job - without much help of others - and if there were a bronze medal, he would probably share the medal with the Cornell group - Patrick Meade and Matt Reece - who almost solved the Harvard box and impressed the olympic spectators, too.

Both the Washington and the Cornell group chose not to see the revelation of the Harvard model so that they can continue to work on it and hopefully nail it down.

Matt Strassler took a very rational approach but made an incorrect assumption from a subset (1/8) of the data which led to him to a wrong direction. The Princeton team presented a lot of evidence for their idea what the Washington model was - graphs that essentially matched - except that the model was not quite right. The Harvard team got it right.

The general predictions of The Reference Frame have been confirmed: when powerful companies such as the Harvard phenomenology thinking machine start to work hard, it may take about one weekend to figure out the right model from the data which was actually the case. Indeed, the Harvard team nailed both models down, including the harder Michigan box. The Harvard solution of the Michigan box was similar to Matt Strassler's solution from the last year, obtained from 40% of the data. By the way, the typical amount of data included in the boxes was about 5 inverse femtobarns which corresponds to approximately 1 year of the LHC data.

The skepticism has evaporated and many people, including the experimentalists, said that they learned a great deal of stuff from this "game". It is more or less guaranteed that many more people will participate in the next olympics. Such a contest puts the physicists in front of the "real" challenges and hidden prejudices that they may have if they think about a model they know. If you're thinking about a model that you don't yet know, the reasoning works very differently.

I think that all physicists who know how to do phenomenology should think how to create powerful local teams that will be able to compete with the Harvard phenomenologists in the near future. The teams must be sufficiently powerful - for example, Western Europe should probably create its own unified team, much like the West Coast or Asia.