This week brought a close to the 2009 season with the announcement of the annual MVP and Cy Young award winners. While I happen to agree with all four of the baseball writers' picks for those awards this year, it's a fair question to ask whether the process used to determine the MVP and Cy Young awards is accurate and fair.
For a little background, the modern MVP Awards were first issued in 1931, when the Baseball Writers Association of America picked up the void left by the defunct league-sponsored awards. As it currently stands, two writers from each city in each league are selected from the local chapters of the BBWAA to vote on each award. Therefore just 32 voters in the NL and 28 voters in the AL have the power to select the MVP, Cy Young and Rookie of the Year awards in their respective leagues.
A poll of objective experts in the field is a good way to determine awards in any area, and back in 1931 sportswriters were the most qualified individuals to vote on baseball's Most Valuable Player award. With no television and limited radio broadcasts, the writers were among the few objective professionals who were able to view games and players on a daily basis. Now, of course, pretty much everyone has the ability to watch limitless games on television and the Internet, and many more objective professionals work in baseball. Despite this massive shift in the industry, MLB has allowed the BBWAA to keep control of the voting process and permit only print media (and recently a select few Internet writers) who regularly attend games to become BBWAA members, and thus become voters for the postseason awards.
Who's to say that these 32 writers, basically a sample of many equally qualified baseball experts, will choose the correct player for the award? In a close contest between two strong MVP candidates, there will, of course, be disagreement among experts on which player is most deserving. But how can we be sure that a sample of just 32 out of hundreds of well-qualified experts will accurately represent the broader consensus?
Of course, this year's MVPs were easy choices, and we can rest assured that Joe Mauer and Albert Pujols would have been the selections no matter who was asked. But the NL Cy Young was a different matter. In this case there was clearly disagreement between the experts on who deserved the award. Eleven voters choseTim Lincecum, nine chose Chris Carpenter and 12 chose Adam Wainwright. In the end, Lincecum won the award because he had more points overall. But was the writers' pick definitive? Do we have confidence that Tim Lincecum was actually the consensus expert pick for the NL Cy Young? While Lincecum's name will go down in immortality, Cardinals fans will wonder if their guys got jobbed. And who can blame them? A re-vote among similarly qualified experts easily could have turned out differently.
As a hypothetical example, let's say that a particular player enjoys the support of 55% of all qualified baseball experts. He should win the MVP, right? Indeed he should, but since MLB takes a poll of only 32 of these experts, there is actually a pretty fair chance (23%) that the player gets unlucky, receives fewer than half of the 32 votes and loses the award. Because the number of voters is so small, the current system allows the wrong man to win fairly often in a close race.
The current setup is analogous to determining the Presidency by taking a poll of 32 voters. If the election is lopsided, the poll will likely accurately reflect the will of the people as a whole, but in any kind of close contest there is great potential for the less deserving man to win.
Consider this: There are 539 BBWAA members who vote for the Hall of Fame. Most of them don't get an award vote. Why would the BBWAA consider someone qualified to vote on a player's immortality in Cooperstown but not on the MVP? Then there is the plethora of TV and radio broadcasters who cover baseball on a daily basis. They don't get an award vote, either. When you add in the many other writers, analysts and experts who follow baseball for a living and have inside knowledge of the game, the number of highly qualified individuals probably numbers close to a thousand. A sample of 32 of these people simply isn't enough to determine the true consensus MVP or Cy Young in many cases.
How confident can we be in the current writers' selections? I studied this question in an article at Baseball Analysts this week and came to the conclusion that the probability was only 53% that Tim Lincecum was really the true consensus Cy Young pick among a broad group of qualified experts. Meanwhile, there was a 28% chance that Carpenter was the true consensus pick and a 19% chance that Wainwright was the choice. Because only 32 out of hundreds of potential experts got a vote, we'll never know whether Lincecum was truly the "right" choice for the Cy Young, or whether he was just lucky and happened to have a disproportionate share of supporters in that particular group of 32 voters.
Of the other major awards this year, there was no uncertainty surrounding Pujols, Mauer or Zack Greinke's honors. However, the small sample size does cast some doubt upon the Rookie of the Year award, with Chris Coghlan having only a 70% chance of having been the true consensus NL Rookie of the Year (runner up J.A. Happ had a probability of 30%), and Andrew Bailey having an 80% chance of being true consensus RoY in the AL (11% forElvis Andrus and 9% for Rick Porcello).
When compared to other sports, the number of baseball MVP voters is extremely low. The AP's NFL MVP is determined by 50 experts, the NBA MVP is determined by 125 experts and the NHL MVP by 131 experts. The Heisman Trophy has 870 experts voting on the award. I would make the case that several hundred voters could be added to baseball's pool while keeping the expertise of the voters constant. Having the benefit of these voters' opinions would be a major improvement, because it would drastically reduce the chance that the wrong player could win by simple chance.