Stattleship xCase

Stattleship xCase
Rose Eckhardt
Process:
•
•
•
•
Importing data using the Stattleship API
Merging data tables: game logs, players, teams
Categorizing player positions: Pitcher, Catcher, Infielder, Outfielder, Other
Examining available variables
• Relevant to which position(s)?
• Quality of data
• Significant issue for infielders and outfielders
• Assessing normality/distribution
• Attempts to transform
• Creating custom metrics for each position category
• Normalized component metrics
• Distribution of custom metrics across the league by teams
• Attempts to validate custom metrics
Custom Metrics:
Catchers
-
Metric:
-
Fielding skills: Stealers Caught – Stealers
Allowed
Batting skills: Hits / At Bats
-
-
Unable to take mean of Batting Average
variable
50% to each skill, normalized before
combined
-
Mean: 0.0753
-
Median: 0.0222
-
Standard deviation: 0.6133
Custom Metrics:
Pitchers
-
Metric:
-
Fielding skills: Mean strike percentage
Hitting skills: Not applicable since many
pitchers do not bat
-
Mean: 0.0011
-
Median: 0.0517
-
Standard deviation: 1.0054
Custom Metrics:
Other positions
-
Metric:
-
-
Fielding skills: Data not available for
fielders or outfielders (only two available
variables, both had data quality issues)
Batting skills: Hits / At Bats
-
Unable to take mean of Batting Average
variable
-
Mean: 0.3172
-
Median: 0.3661
-
Standard deviation: 0.7087
Custom Metrics - Comparison
Custom Metric by
Teams
-
Metric:
-
Mean of all of the player metric values
for the team
-
Mean: 0.17940
-
Median: 0.14320
-
Standard deviation: 0.2180
Validating the Team Metric
R-squared: 0.102
P-value: 0.0857
Comparing Team Rankings:
My Metric vs. ESPN
Source: http://espn.go.com/mlb/story/_/page/2016mlbpowerrankings_week10/mlb-ranking-mlb-teams-ten-weeks