Boxscore to many to massive splits
Posted: Sun Sep 21, 2025 3:31 pm
A whole new boxscore could be net of player performances vs their averages rather than league averages as in ESPN net points. That might be useful for tracking interactions between stats.
For defense, there are measures of opponent efg% vs. average / expected performance. Could do that for every stat and for defense and offense.
1 boxscore could become at least 4:
regular, vs league averages, vs. player averages and vs. opponent averages.
How much of the changes could be ascribed to changes in minutes and usage could be additional boxscores and be measures of player change and coaching impact especially on minutes change.
6 boxscores over 2 or 1.
Comparisons to league average could be replaced by or supplemented by league average by league average by "position". 5, 3, or possibly more than 5 archetypes.
Or players could be in boxscore as separate lines based on apparent matchup position in specific stibts or plays. 6 boxscores could become 11 or more.
Boxscores and boxscore sets could be done for major lineups or all lineups or all lineup matchups.
And play calls vs random and by specific playcalls and plays vs types of defense.
Dozens to many dozens of groupings to break out, track. If they are to be analyzed, used to manage. Wouldn't stop where the use stops but would want to note, track, discuss where the apparent use stops. How many levels of awareness do players, Coach and management really have / use?
Want the game level data but maybe do roll-ups by week, month, season, various moving averages, by various opponent splits (conference, overall quality, factor strengths, pace).
Many dozens of boxscore outputs could push 100 or more...
Boxscores supplemented by game level overall metrics, including RAPM and RAPM hybrids...
What is potentially available vs. actually available, read, used ...
Tracking data overall and side by side with boxscore and with every split. Tracking data is interesting, could be managed overall but by splits is potentially more useful.
How far have teams gone? How far are they willing to go?
What was this thought exercise worth? If any team saw / did anything more than than they are currently doing it would be worth something, especially if analytic work leads to active consideration > impact.
If you increase or greatly increase analytic output, you should increase or greatly increase the amount of hours studying the output, making recommendations, discussing, making decisions, tracking, refining data, data analysis and decision analysis. More staff doing analysis of inputs and probably bigger presence up the decision analysis process...
Comments?
For defense, there are measures of opponent efg% vs. average / expected performance. Could do that for every stat and for defense and offense.
1 boxscore could become at least 4:
regular, vs league averages, vs. player averages and vs. opponent averages.
How much of the changes could be ascribed to changes in minutes and usage could be additional boxscores and be measures of player change and coaching impact especially on minutes change.
6 boxscores over 2 or 1.
Comparisons to league average could be replaced by or supplemented by league average by league average by "position". 5, 3, or possibly more than 5 archetypes.
Or players could be in boxscore as separate lines based on apparent matchup position in specific stibts or plays. 6 boxscores could become 11 or more.
Boxscores and boxscore sets could be done for major lineups or all lineups or all lineup matchups.
And play calls vs random and by specific playcalls and plays vs types of defense.
Dozens to many dozens of groupings to break out, track. If they are to be analyzed, used to manage. Wouldn't stop where the use stops but would want to note, track, discuss where the apparent use stops. How many levels of awareness do players, Coach and management really have / use?
Want the game level data but maybe do roll-ups by week, month, season, various moving averages, by various opponent splits (conference, overall quality, factor strengths, pace).
Many dozens of boxscore outputs could push 100 or more...
Boxscores supplemented by game level overall metrics, including RAPM and RAPM hybrids...
What is potentially available vs. actually available, read, used ...
Tracking data overall and side by side with boxscore and with every split. Tracking data is interesting, could be managed overall but by splits is potentially more useful.
How far have teams gone? How far are they willing to go?
What was this thought exercise worth? If any team saw / did anything more than than they are currently doing it would be worth something, especially if analytic work leads to active consideration > impact.
If you increase or greatly increase analytic output, you should increase or greatly increase the amount of hours studying the output, making recommendations, discussing, making decisions, tracking, refining data, data analysis and decision analysis. More staff doing analysis of inputs and probably bigger presence up the decision analysis process...
Comments?