The Residue of Design

Yesterday's post critiqued the theory postulating that because luck impacts postseason success, it's unnecessary to build the best team possible.

Let's try to evaluate using the facts of baseball history. There is no perfect method to do so; but one way to examine this is to look at MLB postseason history and see whether the teams with better records have beat the teams with losing records more often than not.
This method has some limitations:

• Teams face schedules of different strengths.
• The team playing in October can be much different than the team that played most of the year. Trades, injuries, and late-summer callups all come into play.
• Team win-loss records themselves can reflect a fair amount of luck: an 88-74 team that lost a few hard-luck games might be better than a 95-67 team that had some lucky wins.
• Looking at wins only does not account for slight differences in games played due to rainouts.
• When teams clinch early they may take the foot off the pedal.

Despite these limitations, given a large enough sample size, teams' season records overall should reflect the quality of teams; and this method has the advantage of simplicity. Future posts will examine this issue in different ways, but today let's examine the results from all 263 post-season series since MLB postseasons started in 1903. In 11 series both teams had the exact number of wins, so in the remaining 152:

• The team with more wins won 134 times
• The team with less wins won 116 times.

This means that the team with more wins won 53.6% of the time.

Now, let's eliminate series between 2 teams with similar numbers of wins; this will tend to capture series between teams of truly different quality. The following table shows results for when we only look at series where the 2 teams had 5-, 10-, and 15-or-more differences in season wins, respectively. If the quality of a team matters - if there is more than just dumb luck involved - I'd expect the percentage of series won by the team with the better record to increase in each row. As you can see, this is definitely the case:

Win Difference Range	More wins	Less wins	More wins - %
5+ only	89	62	58.9%
10+ only	44	22	66.7%
15+ only	18	3	85.7%

This suggests that historically, building the better team does indeed improve your odds of winning a postseason series.