MBB: Duke vs. Michigan Pre-Game and In-Game Thread

[Fish80]

I did "12 take 5," which (I believe) is the number of possible combinations of lineups for a 12-man hoops team. That calculates as 12!/(7!x5!) and works out to be 11x9x8 = 792. Your calculation may be the number of permutations, which (again, I believe) is a different problem.

Why don't you PM if you disagree, so we don't have to bore the entire planet with our discussion?

sagegrouse

I think a more plausible estimate of possible lineups is 3,675.

Let's lay out players by potential position.
Quin 1 or 2
Tyler 1 thru 4
Rasheed 1 thru 4
Andre 2 thru 4
Matt 2 thru 4
Rodney 2 thru 4
Alex 3 or 4
Semi 3 thru 5
Jabari 3 thru 5
Amile 4 or 5
Josh 4 or 5
Marshall 5

We have three potentially at the 1, 6 potentially at the 2 (less one already at the 1), nine potentially at the 3 (less 2 at the 1 or 2), 10 potentiall at the 4 (less 3 at the 1, 2 or 3) and 5 potentially at the five. Multiply out 3 x (6-1) x (9-2) x (10-3) x 5 and you get 3,675.

[Kedsy]

I would argue that Matt Jones makes sense to get major playing time, given his potential is unknown, whereas Sheed and Dre are more of a known commodity.

Because the unknown is always better than the known?

I think a more plausible estimate of possible lineups is 3,675.

Don't forget to multiply by 3 to 5 games.

[howardlander]

I did "12 take 5," which (I believe) is the number of possible combinations of lineups for a 12-man hoops team. That calculates as 12!/(7!x5!) and works out to be 11x9x8 = 792. Your calculation may be the number of permutations, which (again, I believe) is a different problem.

Why don't you PM if you disagree, so we don't have to bore the entire planet with our discussion?

sagegrouse

No need, you are right and I'm wrong. I forgot the 5!. That will teach me to do combinatorics without looking at Wikipedia. In my defense, I did finish grad school in 1988...

Howard

[Goduke2010]

Because the unknown is always better than the known?

Yes, always.

[sagegrouse]

I think a more plausible estimate of possible lineups is 3,675.

Let's lay out players by potential position.
Quin 1 or 2
Tyler 1 thru 4
Rasheed 1 thru 4
Andre 2 thru 4
Matt 2 thru 4
Rodney 2 thru 4
Alex 3 or 4
Semi 3 thru 5
Jabari 3 thru 5
Amile 4 or 5
Josh 4 or 5
Marshall 5

We have three potentially at the 1, 6 potentially at the 2 (less one already at the 1), nine potentially at the 3 (less 2 at the 1 or 2), 10 potentiall at the 4 (less 3 at the 1, 2 or 3) and 5 potentially at the five. Multiply out 3 x (6-1) x (9-2) x (10-3) x 5 and you get 3,675.

Nope. I disagree. If there are only 792 possible lineups from a roster of 12 players, there is no other logic that can produce a larger number. Here's a website that may be helpful or may just confuse the issue.

If you have a roster of six players, there are six possible lineups.
If you have a roster of seven players, there are 21 possible lineups.
With eight players, there are 56 lineups.
With nine players, there are 126.
With ten players, there are 252.
With 11 players, there are 462.
A roster of 12 players produces 792 possible lineups.

Adding restrictions like you propose just reduces the number of possible lineups. No way one can one get to the number you suggested.

Now this approach doesn't assign players by position; it just puts them on the floor. I believe that is the Coach K approach. Now if you said, "I want to assign each player to a position 1 thru 5, in addition to deciding which players are on the floor." Then for each set of five players there are 5!, or 120, possible permutations and the overall number of combinations and permutations becomes much larger.

Is there any right-minded poster, who cut his eye teeth on calculating the odds of poker hands (52 take 5) or bridge hands (52 take 13), who can weigh in on this?

sagegrouse

[Eakane]

You know, after reading the Arizona game thread and seeing this, maybe there really is something here. I propose the sole way Duke can get out of its horrible funk is if the only players who see the court had in their careers played fewer than 20 minutes for Duke coming into the season. Thus, we need to start a lineup of Matt at PG, Rodney at SG, Semi at SF, Jabari at PF, and Marshall at C. Only then can we achieve Nirvana.

I reject your sarcastic proposal, but I bet that team would win 20+ games. A frontline of Semi, Jabari and MPIII would be formidible, and Rodney and Matt are talented enough to handle the guard duties.

What we really need tonight is for Cook to play confident, Hood to play aggressive, Rasheed to come out of his funk, and Jabari to be Jabari. If we get that, with four significant scoring threats, we can rotate Amile/Hairston to emulate Lance Thomas, and give Semi and MPIII opportunities to show that they've progressing. If Rasheed continues to struggle, we could give Matt and Andre a shot. Alas, I think we'll see 7...

[wilson]

Huh, I get a much larger number. There are 12 choices for the first player, 11 for the second player, 10 for the third, 9 for the 4th and 8 for the fifth. So I get 95040 possible 5 player combinations (12 * 11 * 10 * 9 * 8). What did I do wrong?

Howard

you are multiple-counting each lineup. you want combinations, not permutations.

Let me be the first math wannabe to weigh in. Let's see. Twelve players and five on the court at any time. I get 792 possible combinations, which, at 20 minutes per lineup, would take 396 games to test.

sagegrouse

But the OP said it wouldn't count unless you tested it for 3 to 5 games. Have you worked that into your analysis? Or do we need between 1,188 and 1,980 games?

Because that suggests that a line-up of Cook, Sulaimon, Hood, Parker, and Jefferson is different than a line-up of Sulaimon, Cook, Hood, Parker, and Jefferson.

From math class to English class: the word "than" should be replaced by "from."

/grammar Nazi mode

Only on a Duke board.

[Gthoma2a]

I reject your sarcastic proposal, but I bet that team would win 20+ games. A frontline of Semi, Jabari and MPIII would be formidible, and Rodney and Matt are talented enough to handle the guard duties.

What we really need tonight is for Cook to play confident, Hood to play aggressive, Rasheed to come out of his funk, and Jabari to be Jabari. If we get that, with four significant scoring threats, we can rotate Amile/Hairston to emulate Lance Thomas, and give Semi and MPIII opportunities to show that they've progressing. If Rasheed continues to struggle, we could give Matt and Andre a shot. Alas, I think we'll see 7...

I agree with this. The biggest problem for Rasheed is over-penetrating, though. He gets in the lane great, but he keeps going too far and gets caught in the air. A floater would serve him well. It was the problem Nolan had until his junior season. I would still see no problem in letting the other two guys get some time in at the spot, anyway. Depth isn't a bad thing for a team that needs fresh legs to outscore its opponents. That's who we are until we communicate better on defense. The backdoor cuts that killed us in the last game could have been called out.

[flyingdutchdevil]

Can't be different. Coach K teams don't have positions.

But in theory, to fully test the permutations, Amile would have to bring the ball upcourt, and Quinn guard the post.

Is that different than Thornton guarding the post? Like against Wiggins?

Kidding aside, I think roles have been established if you play the 1-4 offensively for Duke. If you play the 5 offensively, roles are in the air like RoyWill's job security for the next 5 years...

[Fish80]

Nope. I disagree. If there are only 792 possible lineups from a roster of 12 players, there is no other logic that can produce a larger number. Here's a website that may be helpful or may just confuse the issue.

If you have a roster of six players, there are six possible lineups.
If you have a roster of seven players, there are 21 possible lineups.
With eight players, there are 56 lineups.
With nine players, there are 126.
With ten players, there are 252.
With 11 players, there are 462.
A roster of 12 players produces 792 possible lineups.

Adding restrictions like you propose just reduces the number of possible lineups. No way one can one get to the number you suggested.

Now this approach doesn't assign players by position; it just puts them on the floor. I believe that is the Coach K approach. Now if you said, "I want to assign each player to a position 1 thru 5, in addition to deciding which players are on the floor." Then for each set of five players there are 5!, or 120, possible permutations and the overall number of combinations and permutations becomes much larger.

Is there any right-minded poster, who cut his eye teeth on calculating the odds of poker hands (52 take 5) or bridge hands (52 take 13), who can weigh in on this?

sagegrouse

3,675 lineup combinations assumes that

. . . a line-up of Cook, Sulaimon, Hood, Parker, and Jefferson is different than a line-up of Sulaimon, Cook, Hood, Parker, and Jefferson.

3,675 lineup combinations assumes that it matters what position they play. To some extent, it matters. For example, a lineup of Cook, Sulaimon, Hood, Parker, and Plumlee works better with Cook at the point than Plumlee at the point.

I'll concede the point that there are 792 possible lineup combinations. But I'll continue to assert that it matters what position they play.

[Indoor66]

I have never seen so many Angels dance on the head of so many pins. What a group. Will someone please cut the weeds?

[rsvman]

... Depth isn't a bad thing for a team that needs fresh legs to outscore its opponents. ...

True, but fresh legs depend upon that "depth" being exploited. In other words, other players would have to actually play. Just having them sitting on the bench will preserve the freshness of their legs, but not the freshness of the legs belonging to the guys who are actually playing the game. They're still going to get exhausted.

[Gthoma2a]

True, but fresh legs depend upon that "depth" being exploited. In other words, other players would have to actually play. Just having them sitting on the bench will preserve the freshness of their legs, but not the freshness of the legs belonging to the guys who are actually playing the game. They're still going to get exhausted.

Oh, I know. I am advocating lengthening the rotation.

[mgtr]

Nope. I disagree. If there are only 792 possible lineups from a roster of 12 players, there is no other logic that can produce a larger number. Here's a website that may be helpful or may just confuse the issue.

If you have a roster of six players, there are six possible lineups.
If you have a roster of seven players, there are 21 possible lineups.
With eight players, there are 56 lineups.
With nine players, there are 126.
With ten players, there are 252.
With 11 players, there are 462.
A roster of 12 players produces 792 possible lineups.

Adding restrictions like you propose just reduces the number of possible lineups. No way one can one get to the number you suggested.

Now this approach doesn't assign players by position; it just puts them on the floor. I believe that is the Coach K approach. Now if you said, "I want to assign each player to a position 1 thru 5, in addition to deciding which players are on the floor." Then for each set of five players there are 5!, or 120, possible permutations and the overall number of combinations and permutations becomes much larger.

Is there any right-minded poster, who cut his eye teeth on calculating the odds of poker hands (52 take 5) or bridge hands (52 take 13), who can weigh in on this?

sagegrouse

Well, I cut my eyeteeth on teaching statistics in college, both grad and undergrad. You, sir, are 100% correct in your assertions, particularly your line that you cannot come up with more possible lineups than 12 combination 5 = 792.

[El_Diablo]

The next person to post anything about the upcoming game, or anything about the Michigan Wolverines basketball team, gets a free spork from me.

[Gthoma2a]

The next person to post anything about the upcoming game, or anything about the Michigan Wolverines basketball team, gets a free spork from me.

Jabari Parker > McGary! Count on it tonght. hehe. If he outrebounds him, I'll lose my mind, with joy!

[Dr. Rosenrosen]

The next person to post anything about the upcoming game, or anything about the Michigan Wolverines basketball team, gets a free spork from me.

Spork slut here...

Prediction: Beilein throws crazy zone at us tonight. Dre' has a career night blowing it up!

[El_Diablo]

Spork slut here...

Prediction: Beilein throws crazy zone at us tonight. Dre' has a career night blowing it up!

Sorry, but Gthoma2a beat you to it, and I must spread some comments around before sporking you again anyway.

News reports still have Stauskas as uncertain for tonight's game. I am thinking he will play, even if he still has a little pain...Duke is not Coppin State.

[Goduke2010]

Nice article on Jabari's NBA potential / draft position.
http://preview.tinyurl.com/n5wlwel

[peterjswift]

4. Can the ACC reclaim Big 10 Challenge supremacy?
The ACC hasn't won the Challenge since 2008 (tied last year). This year, Indiana @ Syracuse, Penn State @ Pittsburgh, Northwestern @ NC State, and Wisconsin @ Virginia all look winnable. Duke is rightly favored tomorrow night, but will need to play well to win. It could well be a rubber match when all is said and done, because the remaining 7 games look like tossups (Notre Dame @ Iowa, Florida State @ Minnesota, Illinois @ GA Tech, Boston College @ Purdue) or ACC losses (unc @ Michigan State, Maryland @ Ohio State, Miami @ Nebraska).

I was just thinking that this would be the last ACC/Big10 Challenge for Maryland. Then I realized my error...