A happy coincidence: some nice Duke insight from my favorite Michigan fansite

[scottdude8]

I've mentioned mgoblog.com a few times as my go to for Michigan news (in particular their fantastic coverage of Roy's famous "my coaching sucks" game). They also do a lot of good stuff when it comes to crunching the numbers from various analytics sites. A recent post (https://mgoblog.com/content/let%27s-have-some-fun-bart-torvik-dot-com) did some digging using Bart Torvik's numbers and, in looking into how good Michigan's defense has been, also ended up showing how good Duke's is:


Here is a chart on which the best defensive teams in the country are to the top and the best at preventing threes are to the right. High major teams are the big bubbles; low majors are the little dots.

https://mgoblog.com/sites/default/files/image_564.png

Michigan is sui generis amongst the best Ds in the country, more heavily dependent on their ability to prevent launches from deep than the other Ds that have poked their head above the fray. (For the record: those are Duke, Virginia, Texas Tech, Oklahoma, and the purple one way to the left is Kansas State.)

So essentially what this analysis is showing is that Duke is playing Top-5 level defense, but in a slightly different way than Michigan. Michigan is relying more on defending the 3 to yield their elite defense (which makes sense given the Wolverines have some incredible perimeter defenders in Zavier Simpson, Charles Matthews and Jordan Poole who aren't as stellar at the rim, but they can play aggressively on the perimeter because 7-foot-1 Jon Teske is in the paint), while Duke is relies more on defending the rim, which we know is the case based on our stellar block rate. That's not to say our 3 point defense isn't good, it certainly is (I'm pretty sure we're the next "top-right" most dot), but we're certainly not the outlier that Michigan is.

An interesting data point not only showing how good our D has been so far this year, but also how it's been that good, especially in comparison to some other teams.

[JasonEvans]

Great stuff and interesting to note that should Duke and Michigan ever meet this makes me feel more confident about our chances. Michigan succeeds on D by keeping teams from shooting well from the outside... Duke succeeds on O by not really relying on the outside shot, but by taking the ball to the hole. It would seem that their best D matches up quite poorly with our strategy on O.

[HereBeforeCoachK]

Great stuff and interesting to note that should Duke and Michigan ever meet this makes me feel more confident about our chances. Michigan succeeds on D by keeping teams from shooting well from the outside... Duke succeeds on O by not really relying on the outside shot, but by taking the ball to the hole. It would seem that their best D matches up quite poorly with our strategy on O.

BEAT
ME
TO
IT.......the point about Duke matching up well with Michigan (from Duke standpoint). That was what jumped out at me....

[scottdude8]

Great stuff and interesting to note that should Duke and Michigan ever meet this makes me feel more confident about our chances. Michigan succeeds on D by keeping teams from shooting well from the outside... Duke succeeds on O by not really relying on the outside shot, but by taking the ball to the hole. It would seem that their best D matches up quite poorly with our strategy on O.

FWIW my die-hard Michigan family agrees. My Dad and Uncle gave me a call on their way back home from the victory over Indiana this weekend over the moon about our 15-0 start and how good we looked. The conversation turned to a deal my Dad and I have that if Michigan and Duke both make it to the Final Four, we're going (which has been so damn close to happening on the two occasions Michigan made it to the National Championship game it makes my heart hurt, haha). My Uncle then chimed in that, "As good as we (Michigan) look there's no way we beat Duke".

I definitely agree that Duke poses a particularly bad matchup for Michigan (honestly, we do for most teams, but Michigan especially)... Tre is one of the few PGs in the country that can play Zavier Simpson to a draw defensively (and Tre is superior offensively, although if you watched the Indiana game you know that Zavier's jump hook is pretty amazing), and the matchup problems that Michigan can cause it's opponents with long and hyper-athletic shooters at the 3 and 4 (Charles Matthews and Iggy Brazdeikis) obviously won't fly against Cam, RJ and Zion. That said, Michigan is a superior 3 point shooting team compared to Duke, meaning that if we play Michigan on a night where the 3s are falling they can always beat us. It's one of those situations where if Duke and Michigan played 10 games, I'd very confidently assert Duke would win at least 6 or 7 of them... but on a given night Michigan can beat anyone if the 3s are falling from Poole, Matthews, Brazdeikis and the bench.

Let's just say if this matchup comes to fruition I'll be super happy (because it'll likely be in the Final Four the way things are looking) but I also may not have any hair left when the game ends. Don't worry, I'll be rooting for Duke though... I just wouldn't be as completely devastated after a potential loss as I normally am, haha.

P.S. My apologies for continually using "we" and "our" for both Michigan and Duke. It's a habit that probably will never die, haha, but hopefully it's always clear in the context who I mean.

[devildeac]

FWIW my die-hard Michigan family agrees. My Dad and Uncle gave me a call on their way back home from the victory over Indiana this weekend over the moon about our 15-0 start and how good we looked. The conversation turned to a deal my Dad and I have that if Michigan and Duke both make it to the Final Four, we're going (which has been so damn close to happening on the two occasions Michigan made it to the National Championship game it makes my heart hurt, haha). My Uncle then chimed in that, "As good as we (Michigan) look there's no way we beat Duke".

I definitely agree that Duke poses a particularly bad matchup for Michigan (honestly, we do for most teams, but Michigan especially)... Tre is one of the few PGs in the country that can play Zavier Simpson to a draw defensively (and Tre is superior offensively, although if you watched the Indiana game you know that Zavier's jump hook is pretty amazing), and the matchup problems that Michigan can cause it's opponents with long and hyper-athletic shooters at the 3 and 4 (Charles Matthews and Iggy Brazdeikis) obviously won't fly against Cam, RJ and Zion. That said, Michigan is a superior 3 point shooting team compared to Duke, meaning that if we play Michigan on a night where the 3s are falling they can always beat us. It's one of those situations where if Duke and Michigan played 10 games, I'd very confidently assert Duke would win at least 6 or 7 of them... but on a given night Michigan can beat anyone if the 3s are falling from Poole, Matthews, Brazdeikis and the bench.

Let's just say if this matchup comes to fruition I'll be super happy (because it'll likely be in the Final Four the way things are looking) but I also may not have any hair left when the game ends. Don't worry, I'll be rooting for Duke though... I just wouldn't be as completely devastated after a potential loss as I normally am, haha.

P.S. My apologies for continually using "we" and "our" for both Michigan and Duke. It's a habit that probably will never die, haha, but hopefully it's always clear in the context who I mean.

For clarity, you could always use one of the preferred spellings by tOSU fans: Xichigan or M*chigan.

;):rolleyes:

Good analysis on the above BTW.

[Kedsy]

I've mentioned mgoblog.com a few times as my go to for Michigan news (in particular their fantastic coverage of Roy's famous "my coaching sucks" game). They also do a lot of good stuff when it comes to crunching the numbers from various analytics sites. A recent post (https://mgoblog.com/content/let%27s-have-some-fun-bart-torvik-dot-com) did some digging using Bart Torvik's numbers and, in looking into how good Michigan's defense has been, also ended up showing how good Duke's is:

So essentially what this analysis is showing is that Duke is playing Top-5 level defense, but in a slightly different way than Michigan. Michigan is relying more on defending the 3 to yield their elite defense (which makes sense given the Wolverines have some incredible perimeter defenders in Zavier Simpson, Charles Matthews and Jordan Poole who aren't as stellar at the rim, but they can play aggressively on the perimeter because 7-foot-1 Jon Teske is in the paint), while Duke is relies more on defending the rim, which we know is the case based on our stellar block rate. That's not to say our 3 point defense isn't good, it certainly is (I'm pretty sure we're the next "top-right" most dot), but we're certainly not the outlier that Michigan is.

An interesting data point not only showing how good our D has been so far this year, but also how it's been that good, especially in comparison to some other teams.

I think this analysis is a little bit rudimentary. All it's showing is that Michigan's opponents don't shoot many threes (Michigan is 3rd in the nation in lowest opponent %threes; FWIW, Duke is 50th). It doesn't, for example, say how well the opponents shoot the three when they take it (in that category, Michigan is 42nd; Duke is 19th). One could argue that having your opponents take a lot of threes at a bad percentage is better than fewer threes at a good percentage -- though I remember a Pomeroy article where he argued that 3-point shooting percentage is not really something the defense has a lot of control over so keeping them off the line is more important.

Another thing this analysis doesn't show is why the opponents don't shoot a lot of threes. For example, two of Duke's recent teams that gave up a low percentage of threes were 2014 (24.1% of that team's opponents' shots were threes) and 2012 (24.2%). Both of those teams gave up a much smaller percentage of threes than this year's Michigan team (28.9%). But one reason those teams' opponents' shot so few threes was those teams interior defense wasn't very good (the 2014 team's opponents shot over 50% on two-point shots, for example). But this isn't the case for Michigan, whose opponents shoot only 42.5% from two-point range (Duke's opponents shoot just 41.1% on two-point shots).

How good a team's defense is, is basically governed by the four factors: (1) opponents' eFG% (which indirectly takes into account how many threes an opponent takes); in this category, Michigan is 11th in the country and Duke is 3rd; (2) opponents' turnover%; in this category, Michigan is 127th and Duke is 50th; (3) defensive rebounding%; in this category, Michigan is 19th and Duke is 263rd; and (4) opponents' free throw rate; in this category, Michigan is 3rd and Duke is 14th. (all rankings based on unadjusted numbers)

Looking at the above, it would seem Michigan's defensive prowess is based on making their opponents inefficient in their shooting (a part of which is how good they are at keeping their opponents off the three-point line), combined with how good they are at keeping their opponents off the free throw line and how good they are at defensive rebounding. Duke's defensive success is also due to making their opponents inefficient in their shooting, along with keeping their opponents off the FT line and turning them over.

(It's also worth noting that Michigan's opponents are tied for the 15th lowest FT% in the country -- Duke's opponents are tied for 41st lowest. This would appear to be something that Michigan's defense has little or no control over, though it does help their overall efficiency.)

I guess my point is that saying Michigan's defensive success is because they run their opponents off the three-point line is overly simplistic and probably misleading.

[timmy c]

I think this analysis is a little bit rudimentary. All it's showing is that Michigan's opponents don't shoot many threes (Michigan is 3rd in the nation in lowest opponent %threes; FWIW, Duke is 50th). It doesn't, for example, say how well the opponents shoot the three when they take it (in that category, Michigan is 42nd; Duke is 19th). One could argue that having your opponents take a lot of threes at a bad percentage is better than fewer threes at a good percentage -- though I remember a Pomeroy article where he argued that 3-point shooting percentage is not really something the defense has a lot of control over so keeping them off the line is more important.

Another thing this analysis doesn't show is why the opponents don't shoot a lot of threes. For example, two of Duke's recent teams that gave up a low percentage of threes were 2014 (24.1% of that team's opponents' shots were threes) and 2012 (24.2%). Both of those teams gave up a much smaller percentage of threes than this year's Michigan team (28.9%). But one reason those teams' opponents' shot so few threes was those teams interior defense wasn't very good (the 2014 team's opponents shot over 50% on two-point shots, for example). But this isn't the case for Michigan, whose opponents shoot only 42.5% from two-point range (Duke's opponents shoot just 41.1% on two-point shots).

How good a team's defense is, is basically governed by the four factors: (1) opponents' eFG% (which indirectly takes into account how many threes an opponent takes); in this category, Michigan is 11th in the country and Duke is 3rd; (2) opponents' turnover%; in this category, Michigan is 127th and Duke is 50th; (3) defensive rebounding%; in this category, Michigan is 19th and Duke is 263rd; and (4) opponents' free throw rate; in this category, Michigan is 3rd and Duke is 14th. (all rankings based on unadjusted numbers)

Looking at the above, it would seem Michigan's defensive prowess is based on making their opponents inefficient in their shooting (a part of which is how good they are at keeping their opponents off the three-point line), combined with how good they are at keeping their opponents off the free throw line and how good they are at defensive rebounding. Duke's defensive success is also due to making their opponents inefficient in their shooting, along with keeping their opponents off the FT line and turning them over.

(It's also worth noting that Michigan's opponents are tied for the 15th lowest FT% in the country -- Duke's opponents are tied for 41st lowest. This would appear to be something that Michigan's defense has little or no control over, though it does help their overall efficiency.)

I guess my point is that saying Michigan's defensive success is because they run their opponents off the three-point line is overly simplistic and probably misleading.

Awesome analysis of the numbers, Kedsy! Imagine what Duke's defense would look like if there defensive rebounding % improved!

[scottdude8]

I guess my point is that saying Michigan's defensive success is because they run their opponents off the three-point line is overly simplistic and probably misleading.

That's an interesting perspective Kedsy, and one that is definitely born out by watching them play... as I've mentioned a ton when this topic comes up, Jon Teske is one of the most improved players in the entire NCAA and has become a shot-altering, shot-blocking, and rebounding machine down low. So a lot of the non-three point shots that opponents are getting end up being difficult because of the presence of Teske in the lane... and that's something that the numbers alone don't show.