With two games down in the World Series, and the Yankees and Phillies tied at one game apiece, this year’s champion comes down to a five game series. After 162 regular season games, and two playoff series each, the winner of a five game series (essentially) is going to be crowned. The playoffs are an interesting beast in any sport, but in baseball especially. Teams battle through long regular season campaigns to determine playoff matchups, accumulating records that generally mean very little in the playoffs. To say that playoffs are the most effective way of identifying the best team would be foolish.
Based on a correlation study over the last 19 seasons, measuring various season statistics against whether or not teams won the World Series (WS), it turns out that there is a weak correlation between winning regular season games and winning the Championship. In fact, the winningest team in the regular season wins the World Series about a quarter of the time. The team with the best run differential wins the WS about…yep, about a quarter of the time*. Why? Because playoffs are a crapshoot.
Let’s take two random teams, say, the Yankees and Phillies. Oh, they’re in the WS this year? What a coincidence. The Yankees won 63.6% of their games this season, while the Phillies won 57.4% of theirs. If we assume that this is a good estimate of each team’s true ability, and then play another 162 games, the chances that the Phillies finish with more wins than the Yankees is 12.7%**. Basically this means that the better team in this case will finish with a better record nearly 90% of the time.
If we put them up against each other in a seven game series, and assume the Yankees win about 53% of those matchups (63.6/(63.6+57.4), then probabilistically, New York has a 57% to win that 7-game series. So all of a sudden, New York has gone from a 90% chance of proving it is indeed the better team in the regular season (assuming it is), to just a 57% chance in the series.
Looking at the first two games of this series in particular, luck seems to be playing a role as it usually does in small sample sizes. The winner of each game has been the team that had both more hits on balls in play (BABIP), and a lower percentage of runners stranded on base (meaning more runners scored). Offenses have some control over these stats, but the Yankees’ BABIP jumping from .273 in game one to .353 in game two was likely a product of good fortunes, and also the difference in a 3-1 game.
Obviously there are a lot of variables, such as differences in the two leagues (both rules and level of competition) and the nature of a series vs. a regular season that allows teams like New York to go to a three-man rotation. But my point is still intact: teams have a much better chance to prove their true ability in 162 games than in 7, or 5. Don’t get me wrong, I love watching the playoffs. They are fun, exciting and emotional (and make a lot of rich people richer), but they do not often weed out the best team.
However, if I have to pick a winner, my brain says the Yankees, but my heart wants the Phillies.
*I am not exactly saying that in the last 20 years, 25% percent of all the regular season’s best teams won the WS. What I mean is that the regression curve predicted that each regular season champ had about a 25% theoretical shot at winning the WS.
**This assumes a normal distribution, using the binomial variance with n=162, p=team winning percentage.