liveBook · Manning

Rest

As explained in Statistics Slam Dunk

Rest is a period without scheduled games that allows a basketball team or player to recover before competing.

Rest is the number of days off before an NBA regular-season game. Comparing rest between opponents helps answer whether recovery is associated with winning. section 9.0

NBA schedules create unequal rest because teams follow an 82-game schedule shaped by opponents, venues, travel, television, and blackout dates. Analysis therefore compares home and away teams with equal, greater, or fewer prior days off. section 9.0

Grouping games by relative rest produces clear win-rate comparisons:

  • Equal rest: home teams won 57.7% of games, close to the overall 58.7% home winning percentage. section 9.5.3
  • Home-team advantage in rest: home teams won 74% of games when they had at least one additional day off. section 9.5.3
  • Away-team advantage in rest: away teams won 58% of games when they had at least one additional day off. section 9.5.3

Because wins and losses are categorical outcomes, the relationship is tested with a chi-square test for independence. Cramer's V complements it by measuring effect size; the null hypothesis assumes no meaningful relationship. section 9.0

finaldf %>%
 filter(teamDayOff == opptDayOff) %>%
 group_by(teamLoc) %>%
 summarize(wins = sum(n))
## # A tibble: 2 × 2
##   teamLoc  wins
##   <fct>   <int>
## 1 Away     1432
## 2 Home     1955

sqldf("select SUM(n) FROM finaldf WHERE teamLoc ='Home' and 
     teamDayOff = opptDayOff")
##   SUM(n)
## 1   1955

sqldf("select SUM(n) FROM finaldf WHERE teamLoc ='Away' 
     and teamDayOff = opptDayOff")
##   SUM(n)
## 1   1432

Filter equal-rest games, group by location, and sum wins with dplyr; equivalent sqldf queries provide the same counts. section 9.5.3

How Statistics Slam Dunk covers it

How deep it goes
Compares equal and unequal rest with grouped counts, then applies chi-square and Cramer's V rather than stopping at visualization.
Examples
Demonstrates equal, home-favored, and away-favored rest using dplyr and sqldf win counts.
What it assumes
Assumes grouped data manipulation and distinguishes categorical outcomes from numeric outcomes tested with t-tests.
What you can do after
Evaluate rest effects by grouping games, comparing win percentages, and testing categorical associations.
How it connects
Connects rest with categorical data, chi-square tests, Cramer's V, home advantage, scheduling, and travel.
Cover of Statistics Slam Dunk
liveBookmanning.com
Read Statistics Slam Dunk on liveBook Get the book at manning.com

This explainer was generated by Manning's knowledge graph from the text of Statistics Slam Dunk. The section links point to the parts of the book it draws on.