Rebounding Margin: Simple Stats Tell Lies

[JoePgh]

In two recent games against Providence and Marquette, UConn was "out-rebounded" because the raw rebounding totals showed the opponent with a higher number of rebounds. This caused much gnashing of teeth on The Boneyard, and even Kara Wolters on SNY made disparaging comments about it. I'm no basketball expert nor did I stay at a Holiday Inn Express recently, but my "eye test" after having watched both of these games in person and on replay did not make me think that UConn was being dominated on the boards in either game.

But there are some basic arithmetical realities that distort these rebounding totals, at least when they are used to evaluate which team showed more rebounding skill or desire. In the discussion that follows, I'm going to try to try to explain these arithmetic facts, so I ask for a little patience ...

Case 0 (simple and very hypothetical): Imagine a basketball game in which there are no fouls, no turnovers, no held balls, and no offensive rebounds, and equal shooting efficiency for both teams. In such a game, each team would have the same number of possessions and take the same number of shots. (OK, exactly 4 possessions might end with time expiring in the quarter -- let's ignore that case.) Every possession would end either with a made field goal or a defensive rebound. To make it tangible, let's assume that each team takes 60 shots and makes 30 (for a 50% shooting efficiency). That means that the other team would rebound the 30 shots that each team misses, so the simple "rebounding margin" would be 0 (rebounds tied 30-30).

Case 1.1 (allow different shooting efficiency but otherwise just like Case 0): Now let's allow the two teams' shooting efficiency to vary by supposing that Team A shoots 50% (hence they make 30 out of 60 shots) while Team B shoots 33% (making 20 out of 60 shots). In that case, Team A will get 40 rebounds (all of Team B's 40 misses) while Team B will get 30 rebounds (all of Team A's 30 misses). In that case, Team A will have a 40-30 edge in rebounding, which is a +10 rebounding margin, but it would have nothing to do with greater rebounding skill or desire -- both teams are clearing their defensive boards with 100% efficiency. The favorable rebounding margin is a direct result of Team A's greater shooting efficiency.

Case 1.2 (just like Case 0 but allowing for turnovers at an unequal rate): Now let's go back to assuming equal shooting efficiency for both teams, but allow for turnovers to occur at an unequal rate. Let's say Team A turns the ball over 10 times, but Team B turns it over 20 times. So Team A starts with 60 possessions but takes only 50 shots because of its 10 turnovers, and Team B starts with 60 possessions but takes only 40 shots because of its 20 turnovers. So Team A ends up with 70 shots (60 - 10 + 20) while Team B ends up with 50 shots (60 - 20 + 10). If both teams shoot 50%, Team A will score 35 field goals and Team B will get 35 defensive rebounds, while Team B will score 25 field goals and Team A will get 25 defensive rebounds. The end result is that Team B will out-rebound Team A by 35-25. This margin of +10 for Team B (-10 for Team A) again has nothing to do with rebounding skill or desire, but simply reflects more defensive rebounds being available to Team B than to Team A. Remember that we're assuming that neither team ever gets an offensive rebound.

Case 2 (more like the real world, with UConn as Team A and Providence or Marquette as Team B): Case 1.1 proves that when a team like UConn shoots at a significantly higher percentage than its opponent, it should get more rebounds, but if it forces more turnovers than it gives up as in Case 1.2, it should get fewer rebounds. Which effect will predominate? That depends on the game, I.e., the specific differentials in shooting percentage and turnovers.

So, let's look at the actual cases of Providence and Marquette. In the Providence game, UConn shot 41% and the Friars shot 33% -- a significant but not massive differential. However, Providence turned the ball over to UConn 31 times, compared to just 11 UConn turnovers. That is a massive differential, and it resulted in UConn taking 63 field goal attempts to Providence's 40 FGA's. So, of course, more rebounds are going to be available at UConn's offensive end (37 misses) vs Providence's offensive end (27 misses). If you assume (as I do and as statistics confirm) that defensive rebounds are easier to get than offensive rebounds, then you would expect Providence to have a favorable raw rebounding margin -- and they did (by 37-27).

The Marquette game was fairly similar. UConn had a shooting efficiency advantage of 42% to 32%, significant but not massive. But UConn had a huge turnover margin in its favor (21 Marquette TO's vs. 5 for UConn). As a result, UConn had an advantage in Field Goals Attempted: 67-54. Thus, there were more rebounds to be had at UConn's offensive end than at Marquette's offensive end, so Marquette had a 41-33 rebounding margin.

The conclusion that I draw is that in both games, the unfavorable UConn rebounding margin is almost entirely explained by the massive turnover differential in UConn's favor, and should therefore not be of concern to The Boneyard or to Kara Wolters.

So how should one measure a team's meaningful rebounding performance? I think that the best reasonably simple metric is the ratio of the team's offensive rebounds to its missed field goals. That is given below:

In the Providence game, UConn misssed 37 FGA's and got back 9 offensive rebounds, for a ratio of 24%, while Providence missed 27 FGA's and got back 8 offensive rebounds, for a ratio of 30%. So Providence was slightly better on this metric in that game.

Against Marquette, UConn missed 39 FGA's and got back 8 offensive rebounds, for a ratio of 21%. Marquette missed 37 FGA's and got back 11 O-bounds, for a ratio of 30%. So, yes, Marquette was a bit better in recovering its misses.

Another way to measure the effect of rebounding is to look at second-chance points. Providence led in this metric by 9-4, but UConn beat Marquette by 14-4 in this category.

If you're curious about how these stats look in the most recent game against Villanova, here they are:

Raw rebounding margin: 36-30 UConn (as you would expect with a big advantage in shooting percentage and only a small advantage in turnovers)
Shooting percentage: 59%-33% in UConn's favor, a rather massive difference.
Turnover margin: 14-10 in UConn's favor, not so big as in the previous games.
Missed shot recovery rate: UConn 8/24 or 33%, Villanova 12/40 or 30%.
Second-chance points: UConn wins 8-6.

The basic moral of this story is that the simple ("raw") rebounding margin is misleading in any game where there is a big difference in either shooting percentage or turnover margin, and should not be used to judge rebounding performance in such games.

 

[Blond D.]

plus we play a lot of small ball

 

[Vowelguy]

Statistics tell no lies. They are only misinterpreted.

 

[stamfordhusky]

Your point is well taken. A key for people to remember is that a team will get close to 75% of available defensive rebounds while it will only get closer to 25% of available offensive reboundsThus rebounding margin is very dependent on whether available rebounds are skewed towards offensive or defensive rebounds. A few simple ways to have a good rebounding margin include.

1. Make a high percentage of your shots, thus not giving your opponent any chances to get the easy defensive rebounds.
2. Turn the ball over a lot, thus not giving your opponent chances to get the easy defensive rebounds.
3. Do a good job of challenging shots and not giving up easy ones, thus creating a lot of defensive rebounds for your team.
4. Don’t ever steal the ball from your opponent. Make them shoot it so there are more available defensive rebounds for your team. Also, by not going for steals, you will be in better position to box out and get a high percentage of those defensive rebounds.

Thus the only rebounding stats that are very meaningful are what percentage of available defensive rebounds a team gets, and what percentage of available offensive rebounds a team gets. In theory, they can do very well in both of those categories and yet see their opponent have a positive rebounding margin because of the factors mentioned above.

 

[fomalhaut]

Comparing each team's offensive rebounding percentage (or defensive, which amounts to the same thing) is indeed the correct way of assessing rebounding. Total rebounds is misleading for another reason: it counts deadball rebounds. Marquette outrebounded Connecticut 41-33, but 8 of the 41 were deadball rebounds, compared to only 1 for Connecticut, so the margin was much closer, 33-32.

On the other hand, Connecticut usually outshoots its opponents, and defensive rebounding is around 65-70%, so they should have an advantage of a couple of rebounds per game even assuming equal ability/effort.

 

[WiseWillie]

Nice summaries JoePgh and stamford husky. Lesson for all: Need to look at more than just the raw rebounding margin and dig a little into the factors outlined in these summaries (like FGA-FGM and TO's) that have contributed to it.

BY gold stars to each of you.

 

[WiseWillie]

And we must assume of course that the UConn WBB coaching staff with all of the stats that are crunched by several analysts up down and sideways knows all of this and surely a lot more that they have determined to be most important in over 40 years of stat crunching!

 

[Bone Dog]

When I see an opposing team get a lot of o-boards, especially if they use the full shot clock reset, I think about the wasted defensive energy. Those o-boards just mean more time spent defending. I’d rather Paige & Co expend all that energy in other ways.

 

[Woof 101]

In two recent games against Providence and Marquette, UConn was "out-rebounded" because the raw rebounding totals showed the opponent with a higher number of rebounds. This caused much gnashing of teeth on The Boneyard, and even Kara Wolters on SNY made disparaging comments about it. I'm no basketball expert nor did I stay at a Holiday Inn Express recently, but my "eye test" after having watched both of these games in person and on replay did not make me think that UConn was being dominated on the boards in either game.

But there are some basic arithmetical realities that distort these rebounding totals, at least when they are used to evaluate which team showed more rebounding skill or desire. In the discussion that follows, I'm going to try to try to explain these arithmetic facts, so I ask for a little patience ...

Case 0 (simple and very hypothetical): Imagine a basketball game in which there are no fouls, no turnovers, no held balls, and no offensive rebounds, and equal shooting efficiency for both teams. In such a game, each team would have the same number of possessions and take the same number of shots. (OK, exactly 4 possessions might end with time expiring in the quarter -- let's ignore that case.) Every possession would end either with a made field goal or a defensive rebound. To make it tangible, let's assume that each team takes 60 shots and makes 30 (for a 50% shooting efficiency). That means that the other team would rebound the 30 shots that each team misses, so the simple "rebounding margin" would be 0 (rebounds tied 30-30).

Case 1.1 (allow different shooting efficiency but otherwise just like Case 0): Now let's allow the two teams' shooting efficiency to vary by supposing that Team A shoots 50% (hence they make 30 out of 60 shots) while Team B shoots 33% (making 20 out of 60 shots). In that case, Team A will get 40 rebounds (all of Team B's 40 misses) while Team B will get 30 rebounds (all of Team A's 30 misses). In that case, Team A will have a 40-30 edge in rebounding, which is a +10 rebounding margin, but it would have nothing to do with greater rebounding skill or desire -- both teams are clearing their defensive boards with 100% efficiency. The favorable rebounding margin is a direct result of Team A's greater shooting efficiency.

Case 1.2 (just like Case 0 but allowing for turnovers at an unequal rate): Now let's go back to assuming equal shooting efficiency for both teams, but allow for turnovers to occur at an unequal rate. Let's say Team A turns the ball over 10 times, but Team B turns it over 20 times. So Team A starts with 60 possessions but takes only 50 shots because of its 10 turnovers, and Team B starts with 60 possessions but takes only 40 shots because of its 20 turnovers. So Team A ends up with 70 shots (60 - 10 + 20) while Team B ends up with 50 shots (60 - 20 + 10). If both teams shoot 50%, Team A will score 35 field goals and Team B will get 35 defensive rebounds, while Team B will score 25 field goals and Team A will get 25 defensive rebounds. The end result is that Team B will out-rebound Team A by 35-25. This margin of +10 for Team B (-10 for Team A) again has nothing to do with rebounding skill or desire, but simply reflects more defensive rebounds being available to Team B than to Team A. Remember that we're assuming that neither team ever gets an offensive rebound.

Case 2 (more like the real world, with UConn as Team A and Providence or Marquette as Team B): Case 1.1 proves that when a team like UConn shoots at a significantly higher percentage than its opponent, it should get more rebounds, but if it forces more turnovers than it gives up as in Case 1.2, it should get fewer rebounds. Which effect will predominate? That depends on the game, I.e., the specific differentials in shooting percentage and turnovers.

So, let's look at the actual cases of Providence and Marquette. In the Providence game, UConn shot 41% and the Friars shot 33% -- a significant but not massive differential. However, Providence turned the ball over to UConn 31 times, compared to just 11 UConn turnovers. That is a massive differential, and it resulted in UConn taking 63 field goal attempts to Providence's 40 FGA's. So, of course, more rebounds are going to be available at UConn's offensive end (37 misses) vs Providence's offensive end (27 misses). If you assume (as I do and as statistics confirm) that defensive rebounds are easier to get than offensive rebounds, then you would expect Providence to have a favorable raw rebounding margin -- and they did (by 37-27).

The Marquette game was fairly similar. UConn had a shooting efficiency advantage of 42% to 32%, significant but not massive. But UConn had a huge turnover margin in its favor (21 Marquette TO's vs. 5 for UConn). As a result, UConn had an advantage in Field Goals Attempted: 67-54. Thus, there were more rebounds to be had at UConn's offensive end than at Marquette's offensive end, so Marquette had a 41-33 rebounding margin.

The conclusion that I draw is that in both games, the unfavorable UConn rebounding margin is almost entirely explained by the massive turnover differential in UConn's favor, and should therefore not be of concern to The Boneyard or to Kara Wolters.

So how should one measure a team's meaningful rebounding performance? I think that the best reasonably simple metric is the ratio of the team's offensive rebounds to its missed field goals. That is given below:

In the Providence game, UConn misssed 37 FGA's and got back 9 offensive rebounds, for a ratio of 24%, while Providence missed 27 FGA's and got back 8 offensive rebounds, for a ratio of 30%. So Providence was slightly better on this metric in that game.

Against Marquette, UConn missed 39 FGA's and got back 8 offensive rebounds, for a ratio of 21%. Marquette missed 37 FGA's and got back 11 O-bounds, for a ratio of 30%. So, yes, Marquette was a bit better in recovering its misses.

Another way to measure the effect of rebounding is to look at second-chance points. Providence led in this metric by 9-4, but UConn beat Marquette by 14-4 in this category.

If you're curious about how these stats look in the most recent game against Villanova, here they are:

Raw rebounding margin: 36-30 UConn (as you would expect with a big advantage in shooting percentage and only a small advantage in turnovers)
Shooting percentage: 59%-33% in UConn's favor, a rather massive difference.
Turnover margin: 14-10 in UConn's favor, not so big as in the previous games.
Missed shot recovery rate: UConn 8/24 or 33%, Villanova 12/40 or 30%.
Second-chance points: UConn wins 8-6.

The basic moral of this story is that the simple ("raw") rebounding margin is misleading in any game where there is a big difference in either shooting percentage or turnover margin, and should not be used to judge rebounding performance in such games.

Dear God... Please afford me the time @JoePgh has in a day to address the topic.

 

[Rowdy831]

When you are playing mediocre teams no statistics really matter. UConn plays ten or so players in every Big East game because they have so much more talent that other teams and the games are usually not that close. If you are playing your second team for 1/3 of the game it stands to reason the play will become more sloppy and your fundamentals such as rebounding might be challenged.

Where my concern would be is when they play better teams. Face it UConn could win every game starting Paige and Sarah along with any other three players in the Big East this season. When they play ND or USC they were exposed or at least a couple players were exposed as not ready for that level of competition.

The challenge is how do you prepare your team to compete with the 5 or 6 best teams when tournament time comes. So far UConn is 0-2 against teams that could be considered top 6 with South Carolina still on the schedule. The rebounding doesn't matter in league but it will against the best teams. Problem is by time that is figued out the season is over.

 

[Cuango]

In two recent games against Providence and Marquette, UConn was "out-rebounded" because the raw rebounding totals showed the opponent with a higher number of rebounds. This caused much gnashing of teeth on The Boneyard, and even Kara Wolters on SNY made disparaging comments about it. I'm no basketball expert nor did I stay at a Holiday Inn Express recently, but my "eye test" after having watched both of these games in person and on replay did not make me think that UConn was being dominated on the boards in either game.

But there are some basic arithmetical realities that distort these rebounding totals, at least when they are used to evaluate which team showed more rebounding skill or desire. In the discussion that follows, I'm going to try to try to explain these arithmetic facts, so I ask for a little patience ...

Case 0 (simple and very hypothetical): Imagine a basketball game in which there are no fouls, no turnovers, no held balls, and no offensive rebounds, and equal shooting efficiency for both teams. In such a game, each team would have the same number of possessions and take the same number of shots. (OK, exactly 4 possessions might end with time expiring in the quarter -- let's ignore that case.) Every possession would end either with a made field goal or a defensive rebound. To make it tangible, let's assume that each team takes 60 shots and makes 30 (for a 50% shooting efficiency). That means that the other team would rebound the 30 shots that each team misses, so the simple "rebounding margin" would be 0 (rebounds tied 30-30).

Case 1.1 (allow different shooting efficiency but otherwise just like Case 0): Now let's allow the two teams' shooting efficiency to vary by supposing that Team A shoots 50% (hence they make 30 out of 60 shots) while Team B shoots 33% (making 20 out of 60 shots). In that case, Team A will get 40 rebounds (all of Team B's 40 misses) while Team B will get 30 rebounds (all of Team A's 30 misses). In that case, Team A will have a 40-30 edge in rebounding, which is a +10 rebounding margin, but it would have nothing to do with greater rebounding skill or desire -- both teams are clearing their defensive boards with 100% efficiency. The favorable rebounding margin is a direct result of Team A's greater shooting efficiency.

Case 1.2 (just like Case 0 but allowing for turnovers at an unequal rate): Now let's go back to assuming equal shooting efficiency for both teams, but allow for turnovers to occur at an unequal rate. Let's say Team A turns the ball over 10 times, but Team B turns it over 20 times. So Team A starts with 60 possessions but takes only 50 shots because of its 10 turnovers, and Team B starts with 60 possessions but takes only 40 shots because of its 20 turnovers. So Team A ends up with 70 shots (60 - 10 + 20) while Team B ends up with 50 shots (60 - 20 + 10). If both teams shoot 50%, Team A will score 35 field goals and Team B will get 35 defensive rebounds, while Team B will score 25 field goals and Team A will get 25 defensive rebounds. The end result is that Team B will out-rebound Team A by 35-25. This margin of +10 for Team B (-10 for Team A) again has nothing to do with rebounding skill or desire, but simply reflects more defensive rebounds being available to Team B than to Team A. Remember that we're assuming that neither team ever gets an offensive rebound.

Case 2 (more like the real world, with UConn as Team A and Providence or Marquette as Team B): Case 1.1 proves that when a team like UConn shoots at a significantly higher percentage than its opponent, it should get more rebounds, but if it forces more turnovers than it gives up as in Case 1.2, it should get fewer rebounds. Which effect will predominate? That depends on the game, I.e., the specific differentials in shooting percentage and turnovers.

So, let's look at the actual cases of Providence and Marquette. In the Providence game, UConn shot 41% and the Friars shot 33% -- a significant but not massive differential. However, Providence turned the ball over to UConn 31 times, compared to just 11 UConn turnovers. That is a massive differential, and it resulted in UConn taking 63 field goal attempts to Providence's 40 FGA's. So, of course, more rebounds are going to be available at UConn's offensive end (37 misses) vs Providence's offensive end (27 misses). If you assume (as I do and as statistics confirm) that defensive rebounds are easier to get than offensive rebounds, then you would expect Providence to have a favorable raw rebounding margin -- and they did (by 37-27).

The Marquette game was fairly similar. UConn had a shooting efficiency advantage of 42% to 32%, significant but not massive. But UConn had a huge turnover margin in its favor (21 Marquette TO's vs. 5 for UConn). As a result, UConn had an advantage in Field Goals Attempted: 67-54. Thus, there were more rebounds to be had at UConn's offensive end than at Marquette's offensive end, so Marquette had a 41-33 rebounding margin.

The conclusion that I draw is that in both games, the unfavorable UConn rebounding margin is almost entirely explained by the massive turnover differential in UConn's favor, and should therefore not be of concern to The Boneyard or to Kara Wolters.

So how should one measure a team's meaningful rebounding performance? I think that the best reasonably simple metric is the ratio of the team's offensive rebounds to its missed field goals. That is given below:

In the Providence game, UConn misssed 37 FGA's and got back 9 offensive rebounds, for a ratio of 24%, while Providence missed 27 FGA's and got back 8 offensive rebounds, for a ratio of 30%. So Providence was slightly better on this metric in that game.

Against Marquette, UConn missed 39 FGA's and got back 8 offensive rebounds, for a ratio of 21%. Marquette missed 37 FGA's and got back 11 O-bounds, for a ratio of 30%. So, yes, Marquette was a bit better in recovering its misses.

Another way to measure the effect of rebounding is to look at second-chance points. Providence led in this metric by 9-4, but UConn beat Marquette by 14-4 in this category.

If you're curious about how these stats look in the most recent game against Villanova, here they are:

Raw rebounding margin: 36-30 UConn (as you would expect with a big advantage in shooting percentage and only a small advantage in turnovers)
Shooting percentage: 59%-33% in UConn's favor, a rather massive difference.
Turnover margin: 14-10 in UConn's favor, not so big as in the previous games.
Missed shot recovery rate: UConn 8/24 or 33%, Villanova 12/40 or 30%.
Second-chance points: UConn wins 8-6.

The basic moral of this story is that the simple ("raw") rebounding margin is misleading in any game where there is a big difference in either shooting percentage or turnover margin, and should not be used to judge rebounding performance in such games.

It seems this eliminates the effort factor! Yes, percentages and turnovers etc are all factors in the number of rebounds available but effort and fundamentals are much more of a determining factor as to which team wins the rebounding battle. Imo

 

[oldhusky]

Comparing each team's offensive rebounding percentage (or defensive, which amounts to the same thing) is indeed the correct way of assessing rebounding. Total rebounds is misleading for another reason: it counts deadball rebounds. Marquette outrebounded Connecticut 41-33, but 8 of the 41 were deadball rebounds, compared to only 1 for Connecticut, so the margin was much closer, 33-32.

On the other hand, Connecticut usually outshoots its opponents, and defensive rebounding is around 65-70%, so they should have an advantage of a couple of rebounds per game even assuming equal ability/effort.

The rebounding % of available shots, both offensive and defensive is "the" stat that best measures rebounding effectiveness. Those stats are available from several sites and without them in front of me, I recall that we are about average among 362 schools in offensive rebounding, in the top 40 or so in defensive rebounding, and overall above average, but not nearly as good as we usually are.

Yes we will out-rebound our opponents most times, but compared to the other top contenders it is a relative weakness. Sarah and Jana are good rebounders, Ice for a big is not, and Paige and Morgan while good for a guard are below average when playing the 4, and the other guards are a huge drop-off if they are used at the 4.

Small ball has consequences in both directions. Our assists are better, turnovers lower, maybe fewer fouls committed, but way fewer fouls drawn. We are also not as good as Uconn usually is at blocked shots, above average? of course, but not anywhere close to elite which we frequently are.

As long as our rotation guards are considered better than Jana or Ice, small ball may be used quite a bit, but that dynamic could change with Aubrey becoming available soon. You could make a good case for her being better than any guard not named Paige or Azzi, and she clearly brings rebounding and blocked shots to the equation. We still need Jana or Ice to elevate their games a notch or two and provide a reliable 3rd big. Jana is the preferred choice from a rebounding standpoint, and if we had a three big rotation of Sarah, Jana and Aubrey, our rebounding would be more like it usually is.

 

[PacoSwede]

It seems this eliminates the effort factor! Yes, percentages and turnovers etc are all factors in the number of rebounds available but effort and fundamentals are much more of a determining factor as to which team wins the rebounding battle. Imo

the eye test trumps the math test? as a fan that's my take.

 

[Millenial23]

Comparing each team's offensive rebounding percentage (or defensive, which amounts to the same thing) is indeed the correct way of assessing rebounding. Total rebounds is misleading for another reason: it counts deadball rebounds. Marquette outrebounded Connecticut 41-33, but 8 of the 41 were deadball rebounds, compared to only 1 for Connecticut, so the margin was much closer, 33-32.

On the other hand, Connecticut usually outshoots its opponents, and defensive rebounding is around 65-70%, so they should have an advantage of a couple of rebounds per game even assuming equal ability/effort.

what are deadball rebounds?

 

[meyers7]

what are deadball rebounds?

You don't want to go there.

But, well first, I think formalhaut meant "Team Rebounds". Those are for a missed shot, change of possession without anyone actually grabbing the ball. e.g. a missed shot that bounced out of bounds, or is tipped out.

Now a "Deadball rebound" is basically an accounting trick. It's for missed shots that, I guess, don't necessarily change possession. e.g. 1st shot of a 2 shot foul. Or end of quarter heave that misses.

At least as I understand them. But this would probably not hold up in......ahem....court.

 

[meyers7]

In two recent games against Providence and Marquette, UConn was "out-rebounded" because the raw rebounding totals showed the opponent with a higher number of rebounds. This caused much gnashing of teeth on The Boneyard, and even Kara Wolters on SNY made disparaging comments about it. I'm no basketball expert nor did I stay at a Holiday Inn Express recently, but my "eye test" after having watched both of these games in person and on replay did not make me think that UConn was being dominated on the boards in either game.

But there are some basic arithmetical realities that distort these rebounding totals, at least when they are used to evaluate which team showed more rebounding skill or desire. In the discussion that follows, I'm going to try to try to explain these arithmetic facts, so I ask for a little patience ...

Case 0 (simple and very hypothetical): Imagine a basketball game in which there are no fouls, no turnovers, no held balls, and no offensive rebounds, and equal shooting efficiency for both teams. In such a game, each team would have the same number of possessions and take the same number of shots. (OK, exactly 4 possessions might end with time expiring in the quarter -- let's ignore that case.) Every possession would end either with a made field goal or a defensive rebound. To make it tangible, let's assume that each team takes 60 shots and makes 30 (for a 50% shooting efficiency). That means that the other team would rebound the 30 shots that each team misses, so the simple "rebounding margin" would be 0 (rebounds tied 30-30).

Case 1.1 (allow different shooting efficiency but otherwise just like Case 0): Now let's allow the two teams' shooting efficiency to vary by supposing that Team A shoots 50% (hence they make 30 out of 60 shots) while Team B shoots 33% (making 20 out of 60 shots). In that case, Team A will get 40 rebounds (all of Team B's 40 misses) while Team B will get 30 rebounds (all of Team A's 30 misses). In that case, Team A will have a 40-30 edge in rebounding, which is a +10 rebounding margin, but it would have nothing to do with greater rebounding skill or desire -- both teams are clearing their defensive boards with 100% efficiency. The favorable rebounding margin is a direct result of Team A's greater shooting efficiency.

Case 1.2 (just like Case 0 but allowing for turnovers at an unequal rate): Now let's go back to assuming equal shooting efficiency for both teams, but allow for turnovers to occur at an unequal rate. Let's say Team A turns the ball over 10 times, but Team B turns it over 20 times. So Team A starts with 60 possessions but takes only 50 shots because of its 10 turnovers, and Team B starts with 60 possessions but takes only 40 shots because of its 20 turnovers. So Team A ends up with 70 shots (60 - 10 + 20) while Team B ends up with 50 shots (60 - 20 + 10). If both teams shoot 50%, Team A will score 35 field goals and Team B will get 35 defensive rebounds, while Team B will score 25 field goals and Team A will get 25 defensive rebounds. The end result is that Team B will out-rebound Team A by 35-25. This margin of +10 for Team B (-10 for Team A) again has nothing to do with rebounding skill or desire, but simply reflects more defensive rebounds being available to Team B than to Team A. Remember that we're assuming that neither team ever gets an offensive rebound.

Case 2 (more like the real world, with UConn as Team A and Providence or Marquette as Team B): Case 1.1 proves that when a team like UConn shoots at a significantly higher percentage than its opponent, it should get more rebounds, but if it forces more turnovers than it gives up as in Case 1.2, it should get fewer rebounds. Which effect will predominate? That depends on the game, I.e., the specific differentials in shooting percentage and turnovers.

So, let's look at the actual cases of Providence and Marquette. In the Providence game, UConn shot 41% and the Friars shot 33% -- a significant but not massive differential. However, Providence turned the ball over to UConn 31 times, compared to just 11 UConn turnovers. That is a massive differential, and it resulted in UConn taking 63 field goal attempts to Providence's 40 FGA's. So, of course, more rebounds are going to be available at UConn's offensive end (37 misses) vs Providence's offensive end (27 misses). If you assume (as I do and as statistics confirm) that defensive rebounds are easier to get than offensive rebounds, then you would expect Providence to have a favorable raw rebounding margin -- and they did (by 37-27).

The Marquette game was fairly similar. UConn had a shooting efficiency advantage of 42% to 32%, significant but not massive. But UConn had a huge turnover margin in its favor (21 Marquette TO's vs. 5 for UConn). As a result, UConn had an advantage in Field Goals Attempted: 67-54. Thus, there were more rebounds to be had at UConn's offensive end than at Marquette's offensive end, so Marquette had a 41-33 rebounding margin.

The conclusion that I draw is that in both games, the unfavorable UConn rebounding margin is almost entirely explained by the massive turnover differential in UConn's favor, and should therefore not be of concern to The Boneyard or to Kara Wolters.

So how should one measure a team's meaningful rebounding performance? I think that the best reasonably simple metric is the ratio of the team's offensive rebounds to its missed field goals. That is given below:

In the Providence game, UConn misssed 37 FGA's and got back 9 offensive rebounds, for a ratio of 24%, while Providence missed 27 FGA's and got back 8 offensive rebounds, for a ratio of 30%. So Providence was slightly better on this metric in that game.

Against Marquette, UConn missed 39 FGA's and got back 8 offensive rebounds, for a ratio of 21%. Marquette missed 37 FGA's and got back 11 O-bounds, for a ratio of 30%. So, yes, Marquette was a bit better in recovering its misses.

Another way to measure the effect of rebounding is to look at second-chance points. Providence led in this metric by 9-4, but UConn beat Marquette by 14-4 in this category.

If you're curious about how these stats look in the most recent game against Villanova, here they are:

Raw rebounding margin: 36-30 UConn (as you would expect with a big advantage in shooting percentage and only a small advantage in turnovers)
Shooting percentage: 59%-33% in UConn's favor, a rather massive difference.
Turnover margin: 14-10 in UConn's favor, not so big as in the previous games.
Missed shot recovery rate: UConn 8/24 or 33%, Villanova 12/40 or 30%.
Second-chance points: UConn wins 8-6.

The basic moral of this story is that the simple ("raw") rebounding margin is misleading in any game where there is a big difference in either shooting percentage or turnover margin, and should not be used to judge rebounding performance in such games.

Eh, sort of. While that is somewhat true, yes with poorer shooting team, there's more rebounds available, but that means for UCONN there are more "defensive" rebounds to get. Which should make us a + on the rebound margin.

This shooting accuracy, efficiency, is really only applicable to offensive rebounding. A big reason UCONN doesn't get a lot of offensive rebounds. (they are not there to get)

TO's, eh, maybe a little bit? True with turnovers, there are less shots, less defensive rebounds available, but also less offensive rebounds available for the shooting team.

HOWEVER, against teams of the Big East caliber, UCONN should be outrebounding them regardless of TOs and efficiency. Maybe not "dominate", but still should be outrebounding them. And right now, our Rebounding Differential of +4.4 is probably the lowest I can remember in a long time. (actually I don't remember it every being that low) Usually we are up around +10 or so. Something for Geno and Co to work on.

 

[fomalhaut]

The rebounding % of available shots, both offensive and defensive is "the" stat that best measures rebounding effectiveness. Those stats are available from several sites and without them in front of me, I recall that we are about average among 362 schools in offensive rebounding, in the top 40 or so in defensive rebounding, and overall above average, but not nearly as good as we usually are.

Where do you find those stats, without hand calculating from the aggregate totals?

 

[Bone Dog]

To win, you must score more points than your opponent. To do this you must do any or all of these:

1) shoot more shots than they shoot, and/or make shots at a higher % than they do,
2) get more possessions than they get,
3) make more free throws than they make.

These are the only direct factors involved in winning. Rebounding is not a direct factor in any victory. However, it can be critical for #2 if we consider an o-board as a new possession. This is why people worry so much about winning the rebounding battle. The other contributing factor in #2 is creating turnovers. In recent years, we have been better at the turnover component of #2 than the rebounding component, though there have been exceptions. This season, we are winning the rebounding battle 36-32 and the turnover battle 18-10. This is a slight improvement (so far) over last season (38-33, 17-13), and clearly better than the year before that (41-31, 14-16). You have to go back to 2017-18 to find a significantly better rebounding/turnover advantage (38-27, 17-12), Which is on a par with the NC champion teams — a double figure rebounding advantage and a 4-5 turnover advantage.

As for #1, this is impacted by better shooting (and indirectly by better passing) and by better defense against opposing shooters. Rebounding also has an indirect role to play here.

 

[gandydancer]

I don't see this factoring in the 3 point shot. Plug in the numbers where one team shoots only 3 point shots and the other team shoots only 2 point shots.

 

[NycUcWbbFan]

Expounding more on BoneDog’s train of thought…

How does UConn win games?

• And why does it seem that Geno prioritizes, with a development program that is more transparent this year with a young team, certain items in a certain order?
• This order seems to be half-court offense, scoring defense (defense other than rebounding), press, … and then rebounding (and post play &1) closer to the true rebounding aptitude tests of TN (9th game hence) and SC (12th game hence).

A team wins if it outscores its opponent:

• Note that Points = Points Per Possession x Number of Possessions: PTS = PPP x p;
• UConn wins if PTS > PTS’ ==> PPP x p > PPP’ x p’; (‘ denotes opponent);
• Note that p = FGA possessions (p1) + non-FGA possessions (p2);
• Note also that p = p1 + p2 = p’ = p1’ + p2’ (may be off by 1 or 2 possessions) due to the definition of possessions and end-of-quarters end possessions;
• UConn wins if PPP > PPP’ and/or p1 > p1’;
• In short, UConn wins if it wins the possessions battle (scoring efficiency in own possessions, defensive efficiency in opponent’s possessions);

The possessions battle is won through the following:

• UConn has better efficient offense (better shot selection, better accuracy, better effective scoring (reflecting free throws), i.e. better PPP;
• UConn has more FGA possessions (better p1), i.e. less turnovers (efficient offense);
• The opponent has less efficient offense imposed by UConn’s scoring defense;
• The opponent has less FGA possessions (efficient scoring defense);
• With alternating possessions, the opponent’s non-FGA possessions result in transition opportunities (from opponent turnovers imposed by UConn’s scoring defense) on UConn’s end — (efficient) scoring defense feeding (efficient) offense — or trading an opponent’s empty possession (scoring defense without fouling) with UConn’s effective non-transition possession.

How does rebounding factor in?

• A prolific rebounding opponent can lose if it is inept on offense (multiple offensive rebounds in the same possession results in multiple FGAs but the same ultimate reward of 0-4+ points in one possession) exacerbated by UConn’s superb scoring defense;
• Even when UConn cannot get an offensive rebound of a missed FGA, UConn can win if it has an efficient offense (it has enough FGMs);
• one-and-done on a UConn FGA possession is the same as many-and-done (through offensive rebounding) on an opponent FGA possession — the same opportunity for 0-4+ points on one possession;
• Having more FGAs or more rebounds are not determinative of a win, in the same magnitude as having an efficient offense and an efficient scoring defense.

But a rebounding win margin can be win-determinative enough for an opponent when:

• UConn’s offense sputters; and
• UConn’s scoring defense is inadequate resulting in opponent better offense (PPP’) and less transition offense for UConn (p1 < p1’; p2’ is minimal).

So, way before the likes of TN and SC, UConn will hone on its offense and scoring defense as the surest way to win games;

• It seems that Geno may now devote more bandwidth to improve rebounding after its offense and scoring defense performance in the Villanova game (punctuating the team’s remarkable development backed by statistical evidence).

For the above reasons, imo, the frequently-litigated rebounding issue (reinforced by anemic early season rebounding statistics this year for this young team) is often a red herring.

&1 Geno does want Jana and Ice to develop from the very beginning because it makes UConn more formidable in the postseason. But he has a young team with many aspects to develop (and requires training bandwidth and thus far he only has had a 15-game training bandwidth in a 31-game regular season).

 

[Steelerone]

Dear God... Please afford me the time @JoePgh has in a day to address the topic.

I think I read War and Peace in half the time,