I don’t want this blog to deal very much with the work of other people, and I am not interested generally in being a statistical or graphical critic, but some things are just abhorrent to a true Arbitrarian, and I feel compelled to discuss them.
It is a fact of life in sports journalism in general that people just make things up all of the time. This is neither the time nor place to discuss this fully, but people make money when consumers watch their programs, listen to their shows, and read their blogs. Thus, there is an incentive to publish anything and everything you can think of, regardless of merit, just to get it in front of eyeballs. This is especially true of particularly strong or controversial opinions–more discussion and debate just generates greater attention and revenue. This is one of the reasons that so many shows (not just sports shows) offer an adversarial format in which each of several individual personalities adopt a certain position and attempt to justify it. It is not that they care about their position, or even necessarily believe it, it is just that there are always at least two positions that could be taken on any subjective matter, and arguments to be made. Television is not the only guilty party, however, in this shouting match.
We are in the midst now of a legitimate debate about who might be the most deserving of the NBA regular season MVP award. Depending on the criteria you use, certain players may rise to the top of your consideration (I will post on this in the very near future, with my own contention). The NBA loves this, just as the NCAA loves the BCS rankings and bowl determinations — the more ambiguous and arbitrary the process, the more discussion there will be, the more profit for the leagues and the media that report on the leagues. If, for example, the MVP award was automatically awarded to the player with the highest pts/gp on the team with the best record (which some may contend is a valid way of determining it), there might be some drama during the season if the race is close, but generally, there would be nothing to talk about, because the numbers alone would have it. Instead, the NBA holds a vote, so that any number of idiosyncratic factors may go into the determination of who’s “Most Valuable.” For the Arbitrarian, this is obnoxious, but to many, this is profitable.
I am always interested in reading a well-reasoned argument for some player as most valuable. I enjoy seeing assumptions and criteria stated, and a rational argument of some sort that those criteria are appropriate, and that one player or another is the best choice, given these criteria. What bothers me is when no real criteria are stipulated. What bothers me more is when “statistics” are used to offer the illusion of objectivity, while actually only covering up subjectivity.
A particularly egregious example of this, and the impetus for my writing this, can be found here: http://www.realgm.com/src_goaltending/136/20080416/finding_the_true_mvp/ (I hesitate to post the link, as traffic merely incentivizes this type of article). The author begins by ostensibly defining his criteria, his analysis will be “One that looks strictly at what advanced stats can tell us about which player actually has the most value to his team. Emphasis on ‘his team.'” He then proceeds to use one statistic, Pythagorean Wins Differential, which seems a reasonable choice, although he does not explicitly justify it, to construct a top five list.
However, transparently this list does not satisfy the author. He then performs a number of completely arbitrary transformations on the data, multiplying win differential by percentage of minutes played (?), multiplying by team winning percentage (which I thought would have already been considered in the win difference statistic), and adding (!) PER (!!!). None of these transformations make a lot of sense to me; at the very least, they are not fully explained by the author. Why don’t we also add the square root of each player’s blocks/home game, because, you know, we need to consider defense on the home court? The obvious reason for the author’s machinations are that he had in his mind a single player, or group of players who should be deserving of MVP, and his preferred statistic (the Pythagorean Win Differential) did not confirm his initial preferences. So, he added and subtracted and multiplied “advanced” statistics, until he found some transformation that supported his initial (unstated) opinion. Then, of course, he makes it official by calling his choice a 100 out of 100, and scaling everyone down from there. This scaling helps to make the reader forget all the nonsense that came before, and effectively decimates the convoluted units he had arrived at (Playing-time-and-Team-Success-weighted-Pythagorean Win Differential… Now with PER!). All we really need to know, of course, is that LeBron had a perfect season. “Remember, this is a measure of a player’s value to his particular team.” Or something like that. Remember to tune in next time, for “a look at the MVP scores for each individual team.”
I have nothing personal against this author, but his is an example of a larger phenomenon: subjectivity masquerading as objectivity. It is my belief that honest analysis is marked by commitment to methodology and consistency in rhetoric, and as an Arbitrarian, I hate to see the arbitrariness of so much that gets published.
Update: Apparently, Slate has a somewhat similar take.