MBB: Duke vs NCSU Pregame and In-Game Thread (Sat 1400 EST, CBS)

[CDu]

Well, big line-up or small line-up, it also screams poor defense to me. I agree with you that playing against a small line-up or big line-up doesn't really matter, but I don't think UVa making a comeback was a fluke. It's bad, lazy defense.

The fluke-ish part of it is how quickly those points came. I agree that part of it is our poor defense, but beyond that I think there was some fluke-ishness there. I completely agree that we have just been a bad defense - regardless of type/size of opponent.

[GGLC]

I agree with CDu entirely in this debate. I also agree that people aren't crediting our periodic, sustained offensive struggles as much as they could in trying to explain our losses.

[roywhite]

The fluke-ish part of it is how quickly those points came. I agree that part of it is our poor defense, but beyond that I think there was some fluke-ishness there. I completely agree that we have just been a bad defense - regardless of type/size of opponent.

Points came in bunches...a 3-pt shot by a mediocre shooter (Justin Anderson) and then 3-pt plays via basket and foul shot, and a steal mixed in. Points came quickly from Notre Dame and Clemson also.

Not sure what the exact diagnosis is, but this recurrent late second half malady is nasty and needs to be countered for the overall health of this team.

[FerryFor50]

I agree with CDu entirely in this debate. I also agree that people aren't crediting our periodic, sustained offensive struggles as much as they could in trying to explain our losses.

Agreed. The defense is generally consistently bad all game. It's the offense that is inconsistent at times.

[Kedsy]

Where 1 is made three point attempt, and 0 is a missed three point attempt, one of these is a snap shot of Andre this season, and the other four are from a random number generator (using Andre's current 45.2% 3pt goal percentage). Can you pick out Andre?

A. 11101011100100111100000110110010010111101101100
B. 00101001000011111100001010001000011111111100100
C. 10000111011000101010100011010100001111111011111
D. 00011100110011101011010000001000001001100110101
E. 11010001101110000001101100010010111011011110010

It's a trick question. None of them show 3 for 14 on either end, and that's what he's shot in his last 14 attempts. So I say none of the above.

Also, I don't quite get the point of your exercise. If you assume Andre is a 45.2% three-point shooter, then there will always be the same number of 1s and 0s in any such line of digits. And the 1s will be randomly bunched. But in Andre's career before this season, he'd only been a 40.1% three-point shooter. And before the last four games, he was shooting 50.8% from three this season. So is he a 40% shooter who's shooting better than normal so far? Is he a 50% shooter who went into a slump? Or is he a 45% shooter and the fluctuation is random? Or something else? You don't know. Andre doesn't even know.

[FerryFor50]

It's a trick question. None of them show 3 for 14 on either end, and that's what he's shot in his last 14 attempts. So I say none of the above.

Also, I don't quite get the point of your exercise. If you assume Andre is a 45.2% three-point shooter, then there will always be the same number of 1s and 0s in any such line of digits. And the 1s will be randomly bunched. But in Andre's career before this season, he'd only been a 40.1% three-point shooter. And before the last four games, he was shooting 50.8% from three this season. So is he a 40% shooter who's shooting better than normal so far? Is he a 50% shooter who went into a slump? Or is he a 45% shooter and the fluctuation is random? You don't know. Andre doesn't even know.

His point is that Andre is a machine and can only understand binary code.

[gus]

His point is that Andre is a machine and can only understand binary code.

You'd have another spork, except "[I] must spread some Comments around before commenting on FerryFor50 again."

[FerryFor50]

You'd have another spork, except "[I] must spread some Comments around before commenting on FerryFor50 again."

That's ok. Continue your fine work, Dr Magnus!

[gus]

It's a trick question. None of them show 3 for 14 on either end, and that's what he's shot in his last 14 attempts. So I say none of the above.

Also, I don't quite get the point of your exercise. If you assume Andre is a 45.2% three-point shooter, then there will always be the same number of 1s and 0s in any such line of digits. And the 1s will be randomly bunched. But in Andre's career before this season, he'd only been a 40.1% three-point shooter. And before the last four games, he was shooting 50.8% from three this season. So is he a 40% shooter who's shooting better than normal so far? Is he a 50% shooter who went into a slump? Or is he a 45% shooter and the fluctuation is random? Or something else? You don't know. Andre doesn't even know.

The point is that Andre is a consistent shooter, and that any "slumps" or "streaks" are primarily a result of the clustering illusion.

And one of them absolutely is a snap shot of Dawkin's shots this season. If I chose his last 14, it wouldn't be much of a challenge. I could take a larger sample, and it will still be difficult to see which is Dawkins, and which is a machine.

[FerryFor50]

The point is that Andre is a consistent shooter, and that any "slumps" or "streaks" are primarily a result of the clustering illusion.

And one of them absolutely is a snap shot of Dawkin's shots this season. If I chose his last 14, it wouldn't be much of a challenge. I could take a larger sample, and it will still be difficult to see which is Dawkins, and which is a machine.

Statistics has been, and always will be, about sample size.

Sure, you could take the last 14 and say "Andre is a bad shooter" (not saying that's what Kedsy is saying). Or you could take a stretch where he shot 8-14 (57% on 11/15 and 11/18) or 11-17 (65% on 12/28 and 12/31) and say Andre is the best shooter in the history of basketball.

Lies, damned lies and statistics.

[Kedsy]

The point is that Andre is a consistent shooter, and that any "slumps" or "streaks" are primarily a result of the clustering illusion.

And one of them absolutely is a snap shot of Dawkin's shots this season. If I chose his last 14, it wouldn't be much of a challenge. I could take a larger sample, and it will still be difficult to see which is Dawkins, and which is a machine.

I agree that Andre is a consistent shooter. But in order for your exercise to make sense you'd have to know (1) what is Andre's "true" shooting percentage; (2) that all shots are of equal difficulty; and (3) that there's no psychological component to shooting.

There's no way of knowing #1; #2 is clearly untrue; and you seem to believe in #3, but you've never explained why.

[Wander]

The point is that Andre is a consistent shooter, and that any "slumps" or "streaks" are primarily a result of the clustering illusion.

And one of them absolutely is a snap shot of Dawkin's shots this season. If I chose his last 14, it wouldn't be much of a challenge. I could take a larger sample, and it will still be difficult to see which is Dawkins, and which is a machine.

Nice demonstration. A quantitative next step would be to look at how often 1's follow other 1's, and if that percentage is significantly different than 45%.

[gus]

Nice demonstration. A quantitative next step would be to look at how often 1's follow other 1's, and if that percentage is significantly different than 45%.

I did that with Andre's first 3 seasons some time ago. Whether he made a shot, a series of shots, a miss or a series of misses had absolutely no predictive value on the next shot.

[gus]

I agree that Andre is a consistent shooter. But in order for your exercise to make sense you'd have to know (1) what is Andre's "true" shooting percentage; (2) that all shots are of equal difficulty; and (3) that there's no psychological component to shooting.

There's no way of knowing #1; #2 is clearly untrue; and you seem to believe in #3, but you've never explained why.

(1) I don't see the point of trying to distinguish between actual and "true" shooting percentage.
(2) One of Andre's strengths is that he rarely takes bad shots, and rarely passes on good ones. That's part of why he's consistent. Other players will think the're "hot" and take a crazy off balance, guarded, step back three pointer. *cough* parker *cough*. Elsewhere I argued that Dawkin's Clemson shooting percentage (0-4) is a bit of an outlier, because the game situation required him to force 3 shots at the end of the game. So no- I disagree with your contention that my analysis requires that all shots are of equal difficulty.
(3) I've never argued that. Rather, I've argued that the psychological component to shooting is vastly overrated. And we're talking about Dawkins, who had well publicized emotional struggles, and yet his shooting remained robotically consistent.

[Mtn.Devil.91.92.01.10.15]

I would just like to take a moment to appreciate this level of discourse. Perhaps the difference between DBR and Inside Carolina goes deeper than which team has players that can't read, but this sort of analysis is much more engaging than the whining and bickering I hear tell of at that website just 9 miles down the road.

Go Duke!
Go Andre!
Beat State!

[roywhite]

I would just like to take a moment to appreciate this level of discourse. Perhaps the difference between DBR and Inside Carolina goes deeper than which team has players that can't read, but this sort of analysis is much more engaging than the whining and bickering I hear tell of at that website just 9 miles down the road.

Go Duke!
Go Andre!
Beat State!

Yeah, I think we're close to requiring certification and continuing education requirements.

[devildeac]

Where 1 is made three point attempt, and 0 is a missed three point attempt, one of these is a snap shot of Andre this season, and the other four are from a random number generator (using Andre's current 45.2% 3pt goal percentage). Can you pick out Andre?

A. 11101011100100111100000110110010010111101101100
B. 00101001000011111100001010001000011111111100100
C. 10000111011000101010100011010100001111111011111
D. 00011100110011101011010000001000001001100110101
E. 11010001101110000001101100010010111011011110010

I couldn't figure it out. Perhaps I need a different/better binary translator.

[Lar77]

I agree with CDu entirely in this debate. I also agree that people aren't crediting our periodic, sustained offensive struggles as much as they could in trying to explain our losses.

The comebacks against us have seemed to result from poor judgment on defense and offense often leading to offensive struggles. I understand, and support, slowing the game down when we are ahead, but we seem to wait until countdown time before thinking about execution. Result > missed shot, rebound to the other team, not getting back on defense or trying for a steal leading to an open shot and basket for other team. We know that this is coached, but as someone (Greybeard? Sagegrouse?) has said before, it takes time to unlearn previously successful but bad habits and learn new habits. And that process forces a lot of thinking until it becomes learned and a lot of mistakes along the way.

[Kedsy]

(1) I don't see the point of trying to distinguish between actual and "true" shooting percentage.

Your machine can't spit out a series of 1s and 0s without you telling it what percentage should be 1s. The true clusters only resemble the random clusters if the data you fed the computer is accurate.

(2) One of Andre's strengths is that he rarely takes bad shots, and rarely passes on good ones. That's part of why he's consistent. Other players will think the're "hot" and take a crazy off balance, guarded, step back three pointer. *cough* parker *cough*. Elsewhere I argued that Dawkin's Clemson shooting percentage (0-4) is a bit of an outlier, because the game situation required him to force 3 shots at the end of the game. So no- I disagree with your contention that my analysis requires that all shots are of equal difficulty.

This seems circular to me. He's consistent therefore he's consistent? Andre isn't as bad about this as some players, but he clearly takes "heat check" shots from time to time.

(3) I've never argued that. Rather, I've argued that the psychological component to shooting is vastly overrated. And we're talking about Dawkins, who had well publicized emotional struggles, and yet his shooting remained robotically consistent.

Or, maybe, he was really a 50% three-point shooter who during his first three seasons suffered myriad short term slumps due to his volatile psychological state, and thus the cold streaks really were cold streaks, even though they looked like random clusters if you assumed he was a 40% shooter. I admit I'm no mathematician, but I don't think there's any way to tell.

[Wander]

(1) I don't see the point of trying to distinguish between actual and "true" shooting percentage.
(2) One of Andre's strengths is that he rarely takes bad shots, and rarely passes on good ones. That's part of why he's consistent. Other players will think the're "hot" and take a crazy off balance, guarded, step back three pointer. *cough* parker *cough*. Elsewhere I argued that Dawkin's Clemson shooting percentage (0-4) is a bit of an outlier, because the game situation required him to force 3 shots at the end of the game. So no- I disagree with your contention that my analysis requires that all shots are of equal difficulty.
(3) I've never argued that. Rather, I've argued that the psychological component to shooting is vastly overrated. And we're talking about Dawkins, who had well publicized emotional struggles, and yet his shooting remained robotically consistent.

I think you can go a step further even. None of these three points is necessary to show streakiness.

1. You don't need to know a "true" shooting percentage to see if a distribution is unusually streaky or not. All you need is the series of 0's and 1's in order.
2. You don't need to make the assumption that all shots are of equal difficulty. There can be variability in difficulty of shot, it just needs to not be strongly correlated with time.
3. Even if the psychological component to shooting was huge and not overrated, it wouldn't matter for your exercise. Once you've shown that Andre is either streaky or consistent, other people could argue that THE REASON for the behavior is due to psychological factors or not, but you don't need to make any assumptions about psychology to simply show if he's streaky.

I did that with Andre's first 3 seasons some time ago. Whether he made a shot, a series of shots, a miss or a series of misses had absolutely no predictive value on the next shot.

Nice. It'd be interesting to know how general this is, and if it applies to other notable three point shooters.