In Fact, I argue that the Olympic Games are more a fanfare of nationalist outpouring and pride. Since I am an extreme nationalist (sometimes boarding on Jingoism), I’ve found American dominance to be particularly fulfilling. Fortunately, the Olympic Games allow for multiple winners in this respect. Total medal count doesn’t necessarily capture competitive release nations enjoy watching their athletes vie for the gold. Despite the Soviets’ leadership in golds (and near parity in overall medal count) in the 1980 Olympic games, the United States triumph in Ice Hockey was celebrated as an amazing victory. I wasn’t alive at the time, but I can imagine that people were more pleased than had the US won a myriad of other events and achieved a high medal count. Other success stories include the Kenyan dominance of the Marathon and Phelps’ epic streak of eight golds. None of these stories encompass an overall victory for their respective nation, but certainly provide something to the citizens of their respective nations. And different nations obviously emphasize different aspects of the Olympics. For some, participating at all on the world stage is exhilarating. Overall, I question the relevance of the medal count to a nation’s success at the Olympics.

Of course the fact that the United States has recently killed on both gold medals and total medal count (summer games: 1996*, 2000, and 2004) is not lost on me. Note that the US hosted the 1996 Olympics — the home field factor is apparent in their tromping over everyone else (similar to the explosion of Chinese gold medals in 2008). Germany enjoyed a similar dominance in the past in the Winter games, decimating all other nations in 1998 and 2006. This factor might be built in to some sort of analysis, but to do so misunderstands why I bring it up. I do simply to show where the US unquestioningly won the medals race (although as stated before, this might not be correlated to their Olympic success). So the question how to weigh medals is only significant in the handful of years where the total medal winner and gold medal winner were different anyway.

This brings us to this point where people are presenting various models accounting for the variation among nations in attempt to isolate some sort of elusive “Olympic variable.” Given that the controls (population, GDP, etc) are equal, which nation would succeed the most? Are the results are supposed to give us some insight as to which nations are the best? Perhaps this super variable could be correlated with other factors such as economic growth, well-being, or life expectancy. But if the goal is to find a great nation, why bother going through all of this work when other statistics such as the Human Development Index already incorporate a basket of other metrics? It seems a rather roundabout way to compare countries on a normalized basis.

Perhaps more puzzling is the notion of comparing normalized values to begin with. As we delve into what makes nations great, I support an analysis in the absolute sense. Won’t factors such as GDP and population (or more likely population distributions) affect the quality of life in a nation? If so, then why attempt to explain them away with statistical tricks? This implies the last question about the validity of statistical manipulation. As the spreadsheet and the variety of other models permeating the Internet aptly show, the choice of assumptions can factor into the eventual outcome of the analysis. Jeff did a great job pointing out that the Fourth Place Medal article created an ill-defined marquee distinction which propelled the United States to first place without which Canada would have remained on top. This is just the kind of silly nonsense that makes me question their entire method and more broadly any explanatory model not based on a natural law.

The ideas of statistical manipulation, measuring subjective well-being, and understanding what factors contribute to an outcome are important ideas. But their frequent misuse is cause for concern. The Olympic Medal Count in itself is a contrived statistical measurement, and unlike the t-statistic, it is mostly useless. It represents only a fraction of the information about the games, discounting the fourth, fifth, sixth, seventh (and so on) finishers. It glosses over margin of victory and rewards risk-taking by awarding all-or-nothing prizes. To compute a real analysis of the Olympics (perhaps to measure the effectiveness of training programs, national coaching, or something else altogether), it’s important to analyze the inputs (talents of individuals) and outcomes (athletes’ performance) using as much as data as possible. Any medal count analysis will fall fantastically short of this requirement. Oh, and the natural law in use here? A simple balance: Coaching ability = Outputs – inputs.