I scraped game results for the full 2015-2016 season from KenPom, and then filtered my results on games that were rematches from earlier in the season. I found 2,961 games that matched this criteria across all of D-1, from the ACC all the way down to the MEAC. I then adjusted these results for home/away splits, accounting for the 4-point advantage gained by home court.
You can propose a game theory argument for this idea: after playing the first game, you can adjust your schemes to the strategy you saw your opponent employ in the first meeting, and then deploy these adjustments in the subsequent meetup before your opponent can react. Basically, Team B (your opponent) made their moves in Game 1, and then you (Team A) make your counter-moves in Game 2. The obvious counterargument to this, however, is that the same thing applies to your opponent - they saw your strategy in Game 1, and can adjust as well in Game 2 before you have a chance to react. In the end this should result in a wash.
The data backs this up: teams "improved" in their second or third matchup only 50.35% of the time. I defined "improvement" as closing the margin of defeat (if they lost the first game) or increasing their margin of victory (if they won the first game), accounting for home/away. On average, this "improvement" actually decreased this margin of victory/defeat slightly by 0.08 points per game. This is over all conferences in Division 1, and contradicts the notion that it's harder to beat a team twice.
No comments:
Post a Comment