Post by eric on Jan 8, 2016 10:04:51 GMT -6
I devised a way to parse the play by play files to get offensive/defensive rebound data. I set up two identical teams (the Los Angeles Sailers and the Washington Illusion) with identical depth charts and roster settings, then simmed the same game over and over until I got ~20 seasons' worth of data. (Unfortunately the pbp parsing process takes about fifteen times as long as getting csv/mdb so I chose not to go to the traditional 80 seasons' worth.) I used four settings: base, giving one team's power forward 100 Offensive Rebounding, giving him 100 Defensive Rebounding, and giving him 100 in both. I will report the results in terms of team and opposing team ORB / (team ORB + opp DRB) = ORB%. Note that this is identical to 1 - opp DRB%...
1 - (opp DRB / (team ORB + opp DRB))
= (team ORB + opp DRB - opp DRB) / (team ORB + opp DRB)
= team ORB / (team ORB + opp DRB)
...and so ORB% gap (or DRB% gap) will reflect changes on both sides of the ball, and here it is:
I do not know if the gap in the base scenario is random noise or due to the fact that the Sailers were the home team in these simulations. Either way, it is clear that Offensive Rebounding is dramatically more important to team rebounding gap. To find out why, we should look more carefully at the rest of the team:
This table lists the player's rebounds (offensive and defensive), the team's, and thus the rest of the team's. Thoughts:
-Being a better rebounder in life sometimes means taking rebounds from an opposing player, but sometimes it means taking rebounds from a teammate, and this reality is reflected in the software. Additionally, we know in the software that a big man that gets more offensive rebounds is more likely to attempt put backs, exacerbating the effect.
-It doesn't look like the software has a "let's have three guys box out Tristan Thompson" effect where a better rebounder can give his team a better chance at the rebound even when he doesn't get it.
-As in life, a player getting offensive rebounds is much more important than defensive rebounds in tipping the rebounding battle, because he mostly takes ORBs from the other team and mostly takes DRBs from his own. Offensive rebounding is only slightly more important than defensive rebounding when it comes to wins, though, which probably has to do with preventing transition by abjuring the offensive glass.
-Finally, each rebounding attribute mostly affects its own side but there is a smaller (about 20%) effect on the other side.
1 - (opp DRB / (team ORB + opp DRB))
= (team ORB + opp DRB - opp DRB) / (team ORB + opp DRB)
= team ORB / (team ORB + opp DRB)
...and so ORB% gap (or DRB% gap) will reflect changes on both sides of the ball, and here it is:
scen sail illu gap
base 0.285 0.280 -0.005
PF O 0.284 0.304 0.020
PF D 0.275 0.276 0.001
PF b 0.274 0.297 0.022I do not know if the gap in the base scenario is random noise or due to the fact that the Sailers were the home team in these simulations. Either way, it is clear that Offensive Rebounding is dramatically more important to team rebounding gap. To find out why, we should look more carefully at the rest of the team:
scen plyr O plyr D team O team D rest O rest D
base 3199 7492 24012 60215 20813 52723
PF O 5267 8096 26989 62999 21722 54903
PF D 3886 13838 23463 61080 19577 47242
PF b 6551 14390 25518 60142 18967 45752This table lists the player's rebounds (offensive and defensive), the team's, and thus the rest of the team's. Thoughts:
-Being a better rebounder in life sometimes means taking rebounds from an opposing player, but sometimes it means taking rebounds from a teammate, and this reality is reflected in the software. Additionally, we know in the software that a big man that gets more offensive rebounds is more likely to attempt put backs, exacerbating the effect.
-It doesn't look like the software has a "let's have three guys box out Tristan Thompson" effect where a better rebounder can give his team a better chance at the rebound even when he doesn't get it.
-As in life, a player getting offensive rebounds is much more important than defensive rebounds in tipping the rebounding battle, because he mostly takes ORBs from the other team and mostly takes DRBs from his own. Offensive rebounding is only slightly more important than defensive rebounding when it comes to wins, though, which probably has to do with preventing transition by abjuring the offensive glass.
-Finally, each rebounding attribute mostly affects its own side but there is a smaller (about 20%) effect on the other side.
