RTC: A dissenter’s opinion [Archive]

View Full Version : RTC: A dissenter’s opinion

timmy c

11-17-2010, 11:34 AM

MChambers

11-17-2010, 11:48 AM

pfrduke

11-17-2010, 11:53 AM

The analysis seems shallow. Our defense is going to force far more turnovers, leading to easy offensive opportunities, hence an increase in offensive efficiency. Saying this team could be like Duke 2007-08 is just silly.

Leaving the analysis aside for the moment, it's worth noting that the 2007-08 team a) was very good (28-5, 13-3); b) won 11 of its 13 ACC games by double digits; and c) finished in the top 11 nationally in both offense and defense. We overweigh, I think, the fact that it lost 3 of its last 5 games and only just snuck by Belmont; that it had a rough spot at the wrong time in the season does not mean that it was not a legitimate contender for the title when the tournament started.

superdave

11-17-2010, 11:54 AM

The analysis seems shallow. Our defense is going to force far more turnovers, leading to easy offensive opportunities, hence an increase in offensive efficiency. Saying this team could be like Duke 2007-08 is just silly.

We had entire games last year without breakaway, fast break scores. We should have between 10-20 fast break points every game this year which will hold our offensive efficiency at a high level. Now our half-court offensive efficiency should drop because we played so deliberately last year and Scheyer took care of the ball so well, but fast breaks will balance it out even with a higher turnover rate.

loran16

11-17-2010, 11:55 AM

I replied on RTC. Not going to retype it here. But basically, he doesn't write anything knew. Duke fans (and most other fans) knew all along that replacing Zoubek would be tough, but he basically ignores all of the following:

Any change of Duke's style from the team.
Any improvement in current players
Any additions from new players.

Those are not exactly small things to miss....and thus he compares us to 07-08 and 08-09....it's like he's like Oh it's Duke but without the good stuff from last year, they're the Bad Duke again! Very poor. (Also he called Zoubek 6'10")

BlueintheFace

11-17-2010, 11:59 AM

Easily the BIGGEST factor he misses is the increase in possessions that should serve to make up for the dive in efficiency

loran16

11-17-2010, 12:07 PM

Easily the BIGGEST factor he misses is the increase in possessions that should serve to make up for the dive in efficiency

Errr, no that doesn't truly work quite that way......Efficiency is Per Possession. If you're less efficient, it doesn't matter how much faster or slower you run, you'll still be worse.

Efficiency stats are adjusted for tempo. Now, increasing or decreasing tempo could result in your team playing better or making other teams play worse for various reasons (comfort level at certain speeds, capabilities of players) and this will be reflected in you efficiency stats increasing.

But an increase in pace by itself DOES NOT make up for a loss of efficiency. Lets use the stat right please.

Kedsy

11-17-2010, 12:08 PM

Check out Rush the Court’s behind the numbers write-up (http://rushthecourt.net/2010/11/17/behind-the-numbers-new-duke-same-as-the-old-duke/
).

Kellen Carpenter gets ahead of the crowd by offering a dissenters opinion as to why Duke won’t be as good as last year. His argument boils down to this:
“Turnovers will probably increase and offensive rebounds will probably decrease, which means offensive efficiency is facing a drop, even if Duke somehow manages to keep up its world-class defense and replace all other lost offensive production. Now, am I saying Duke is going to be bad? No, I’m not. Duke will be very good, but unless a couple things happen, they won’t be in the same class as the national champion 2009-10 team.”

I think he’s got if half right. This team won’t be anything like last year. Playing faster will lead to more turnovers; however it will also lend itself to more scoring opportunities. If turnovers and possessions rise proportionally, this will be a wash. And, offensive rebounds didn’t become a key part of the Duke NC swagger until mid-season.
Coach K is brilliant at helping his team hone their identity and use it to obliterate the opponents. I expect the method to be different, but the ’10-’11 results will be very similar.

This is a shoddy article. He rightly points out why we were efficient last year, and then concludes that because we'll have more turnovers and fewer offensive rebounds we will have "a few more particularly surprising losses and a much shorter postseason." It seems in his move from premise to conclusion, he forgot to bring the logic.

Offensive efficiency is a measure of how many points you get per 100 possessions. According to Pomeroy, our adjusted offensive efficiency was 123.5, best in the nation. In 2009 and 2008 (the seasons he predicts will be similar to this year) we had an offensive efficiency of 117.5 (10th) and 118.2 (11th). This year after two games our OE is 121.2 (1st). The problems with his logic are myriad. First, he ignores defensive efficiency. The top offensive team doesn't win the championship every year. Last year according to Pomeroy we were 4th in the nation in defensive efficiency, but we were 20th in 2009 and 9th in 2008 (FWIW, so far after two games we are 1st in DE as well as OE). So why does he think the improvement from 2009 to 2010 was solely about our improvement in offensive efficiency and not our even bigger improvement in defensive efficiency. It makes no sense.

Second, and possibly more importantly, he ignores the other paths to a stellar offensive efficiency. In 2009, UNC was tops in OE (124.2), but they didn't do it with offensive rebounding and a lack of turnovers. They did it because so many of their baskets were high percentage shots coming from their fast break.

The author of this article ignores Duke's change in style. Our three-point shooting percentage should go up and we (like UNC in 2009) should get a lot of high-percentage baskets on run-outs after turnovers.

Bottom line: his logic doesn't follow. If we don't achieve our team goals this season, it won't be because our offense couldn't score enough.

loran16

11-17-2010, 12:19 PM

This is a shoddy article. He rightly points out why we were efficient last year, and then concludes that because we'll have more turnovers and fewer offensive rebounds we will have "a few more particularly surprising losses and a much shorter postseason." It seems in his move from premise to conclusion, he forgot to bring the logic.

Offensive efficiency is a measure of how many points you get per 100 possessions. According to Pomeroy, our adjusted offensive efficiency was 123.5, best in the nation. In 2009 and 2008 (the seasons he predicts will be similar to this year) we had an offensive efficiency of 117.5 (10th) and 118.2 (11th). This year after two games our OE is 121.2 (1st). The problems with his logic are myriad. First, he ignores defensive efficiency. The top offensive team doesn't win the championship every year. Last year according to Pomeroy we were 4th in the nation in defensive efficiency, but we were 20th in 2009 and 9th in 2008 (FWIW, so far after two games we are 1st in DE as well as OE). So why does he think the improvement from 2009 to 2010 was solely about our improvement in offensive efficiency and not our even bigger improvement in defensive efficiency. It makes no sense.

Second, and possibly more importantly, he ignores the other paths to a stellar offensive efficiency. In 2009, UNC was tops in OE (124.2), but they didn't do it with offensive rebounding and a lack of turnovers. They did it because so many of their baskets were high percentage shots coming from their fast break.

The author of this article ignores Duke's change in style. Our three-point shooting percentage should go up and we (like UNC in 2009) should get a lot of high-percentage baskets on run-outs after turnovers.

Bottom line: his logic doesn't follow. If we don't achieve our team goals this season, it won't be because our offense couldn't score enough.

I agree with you, as I say above, but one minor nitpick: The pomeroy stat you're citing for this year is calculated by weighting our 2 games this year against his projection, which is given the weight of 5 games essentially right now. So it's more projection than actual result, and I wouldn't take it into consideration. We haven't done anything to discourage the projection, but we haven't played tough teams yet.

(By contrast, our D has made kenpom's projection look foolish and we've upgraded from the #3 defensive team according to his projection to the #1 according to the projection combined with the actual results)

Kedsy

11-17-2010, 12:45 PM

I agree with you, as I say above, but one minor nitpick: The pomeroy stat you're citing for this year is calculated by weighting our 2 games this year against his projection, which is given the weight of 5 games essentially right now. So it's more projection than actual result, and I wouldn't take it into consideration. We haven't done anything to discourage the projection, but we haven't played tough teams yet.

(By contrast, our D has made kenpom's projection look foolish and we've upgraded from the #3 defensive team according to his projection to the #1 according to the projection combined with the actual results)

I understand and I agree with you. I wasn't taking it into consideration. I only included this year's numbers for the purposes of being thorough.

airowe

11-17-2010, 12:45 PM

First, he ignores defensive efficiency. The top offensive team doesn't win the championship every year. Last year according to Pomeroy we were 4th in the nation in defensive efficiency, but we were 20th in 2009 and 9th in 2008 (FWIW, so far after two games we are 1st in DE as well as OE). So why does he think the improvement from 2009 to 2010 was solely about our improvement in offensive efficiency and not our even bigger improvement in defensive efficiency. It makes no sense.

Correct. Actually, the best way to use kenpom's numbers to predict tournament results is by looking at the deviation between offensive efficiency and defensive efficiency.

There was an article about it last season just before the tourney and it explained why Duke had the best chance to win it all. Not just because our offensive efficiency was so great or our defensive efficiency was so great, but because they were both at an extremely high level.

COYS

11-17-2010, 12:45 PM

(By contrast, our D has made kenpom's projection look foolish and we've upgraded from the #3 defensive team according to his projection to the #1 according to the projection combined with the actual results)

Our D has been off the charts incredible so far, even if the competition isn't top notch. We'll obviously face stiffer tests in the near future, but any worries that the defensive effort wouldn't be there this year seem to have been misplaced.

Kedsy

11-17-2010, 12:57 PM

Our D has been off the charts incredible so far, even if the competition isn't top notch. We'll obviously face stiffer tests in the near future, but any worries that the defensive effort wouldn't be there this year seem to have been misplaced.

I hope you're right. If so, we are going to be very, very difficult to beat.

CDu

11-17-2010, 01:22 PM

Easily the BIGGEST factor he misses is the increase in possessions that should serve to make up for the dive in efficiency

No, you can't make up for a lack of quality with quantity. An increase in the number of possessions for us would also increase the number of possessions for the opponent. So if we're less efficient on offense, we'll need to be more efficient on defense in order to end up with the same results. That's true regardless of the number of possessions.

Where the argument attributed to the writer misses the boat is that turnovers and offensive rebounds are not the only variables that determine offensive efficiency. At the most basic level, four things come into play: turnover rate, offensive rebound percentage, shooting percentage, and the rate of fouls drawn (and FT shooting percentage). The argument presented in the first post (attributed to the Rush the Court author, but the link didn't work for me) ignored the last two.

We are likely to have a higher turnover rate and we are likely to have a lower offensive rebound percentage. But we may very well get higher percentage shots (via more fast breaks). It remains to be seen whether we'll draw more fouls or shoot a higher percentage. I'd guess not, since Scheyer was pretty good at drawing fouls and REALLY good at making free throws. But that's an unknown. But we could end up being more efficient simply on getting easy transition baskets much more frequently.

gus

11-17-2010, 01:51 PM

Errr, no that doesn't truly work quite that way......Efficiency is Per Possession. If you're less efficient, it doesn't matter how much faster or slower you run, you'll still be worse.

Efficiency stats are adjusted for tempo. Now, increasing or decreasing tempo could result in your team playing better or making other teams play worse for various reasons (comfort level at certain speeds, capabilities of players) and this will be reflected in you efficiency stats increasing.

But an increase in pace by itself DOES NOT make up for a loss of efficiency. Lets use the stat right please.

I think you missed the point.

efficiency * possesions = points

blueintheface is arguing that a drop in the first factor will be compensated by an increase in the second. Duke will score more this season than prior season, despite the lower efficiency.

Listen to Quants

11-17-2010, 01:51 PM

No, you can't make up for a lack of quality with quantity. An increase in the number of possessions for us would also increase the number of possessions for the opponent. So if we're less efficient on offense, we'll need to be more efficient on defense in order to end up with the same results. That's true regardless of the number of possessions.

Where the argument attributed to the writer misses the boat is that turnovers and offensive rebounds are not the only variables that determine offensive efficiency. At the most basic level, four things come into play: turnover rate, offensive rebound percentage, shooting percentage, and the rate of fouls drawn (and FT shooting percentage). The argument presented in the first post (attributed to the Rush the Court author, but the link didn't work for me) ignored the last two.

We are likely to have a higher turnover rate and we are likely to have a lower offensive rebound percentage. But we may very well get higher percentage shots (via more fast breaks). It remains to be seen whether we'll draw more fouls or shoot a higher percentage. I'd guess not, since Scheyer was pretty good at drawing fouls and REALLY good at making free throws. But that's an unknown. But we could end up being more efficient simply on getting easy transition baskets much more frequently.

Indeed. A *balanced* look (as begun by the above) is needed. I had never considered it, but the Miami game brings up another (plausible) effect of the running game. *Defensive* rebounding percentage may go up. I always wondered, last year, how such an excellent offensive rebounding team (7th best in D1 according to KenPom) could be such an ordinary defensive rebounding team (149th). Perhaps part of the problem was other teams were not afraid to send offensive rebounders to the (Duke defensive) boards. Miami certainly chose otherwise.

CEF1959

11-17-2010, 02:00 PM

Bloggers are a dime a dozen, and their opinions are worth no more than that of your next door neighbor's. All this chatter just makes them feel relevant. I suggest either ignoring or posting more useful insights from your next door neighbor.

Duvall

11-17-2010, 02:05 PM

SCMatt33

11-17-2010, 02:19 PM

Let me start with this. I am a regular follower of RTC, and I love the work they do. They put in legit effort and for the most part, do not have an agenda against power programs like many other general college basketball sites. I also think that the scenario presented in the article is entirely plausible. It is not that the numbers arguments presented are wrong or shoddy, they are simply incomplete. I had actually been thinking a lot about Duke’s four factors in recent days since Jumbo asked in phase I “Can Duke force enough turnovers to offset last year’s offensive rebounds?” My gut was that while the two seem related from a game standpoint, with each creating an extra opportunity for Duke, they wouldn’t be so intertwined in a strictly four factors argument.

For those who don’t regularly follow tempo-free stats, or aren’t entirely familiar with the four factors, here’s a quick background on how they work (Here’s a link to a more detailed description (http://www.rawbw.com/~deano/articles/20040601_roboscout.htm)). The four factors (eFG%, TO%, OR%, FT Rate) identify, in that order, the most important, independent factors to winning a game. To note exactly how important they are relative to each other, shooting is usually almost as important as turnovers and rebounds combined, though it can vary from game to game.

While the RTC scenario is valid and plausible, it is probably the worst case plausible scenario. The article mainly sticks to the four factors on offense, but they apply to defense as well (which is in essence, the offensive factors for Duke’s opponents). I offer my own breakdown below of Duke’s factors from 2010, and how they can change in the best case, and worst case scenarios, with the actual numbers likely falling in the middle. Of course, the numbers could very well end up outside these bounds, but these are my best guesses as to what they could reasonably be. I give the scenario, the stat, and the NCAA rank (or projected rank based off of 2010’s numbers. I’m also going to concentrate on the offense here, as this is what the RTC article is about, and this is a long enough post already. I’ll cover the defense in a new thread later because it deserves its own discussion.

OFFENSE

eFG% (NCAA Rank)
2010: 50.5% (92)
2011 worst: 50.5 (92)
2011 best: 55.5 (7)

Effective field goal percentage is the most important of the four factors, and Duke won last year in spite of a mediocre shooting number. For those who have followed Pete Tiernan’s Bracketscience.com, which uses a large number of different computer projections to fill out brackets every year, you know that about half of his systems picked Duke to win it all, which caused skepticism among his followers (and Pete himself) until it turned out to be right. One of the big reasons that Duke wasn’t picked in more of his computers is that Duke’s raw shooting was so poor (though 3pt% makes up for some it in eFG). The worst case scenario presented by RTC is that Duke simply replaces its offensive production. The Plumlees may get more looks inside than Zoubs and Lance, but they won’t as efficient with them (Zoubs had an offensive rating of over 120, second on the team only to Jon). Curry and Irving are both good shooters, but combined, they can only equal Jon’s shooting, which leaves us a square one. The best case scenario is that Curry is a sniper, even beyond NBA range. The team actually manages to improve on its top ten 3pt% from last year, and with more fast break opportunities, and inside looks, the proportion of mid-range jumpers to layups/dunks goes way down, and the shooting numbers go way up as a result.

TO% (rank)
2010: 16.4% (15)
2011 worst: 20.0% (150)
2011 best: 16.4% (15)

Duke’s raw turnover numbers will go up this year. That is a given based on the fact that Duke will play at a faster pace, and more possessions mean’s more turnovers. The worst case scenario is that Duke turns into a very average turnover team, and not in a good way. Jon’s turnover numbers were off the charts, and it is nearly impossible for anyone, let alone a freshman playing in a fast system, to match them. Duke’s turnover numbers were also kept down last year by the fact that the ball was always in the hands of perimeter players. Big guys just don’t handle the ball as well, and with more touches by the big men, there will be more turnovers. The best case scenario is that while Duke has more raw turnovers, its tempo-free numbers somehow stay the same. How can this happen? First, many of Duke’s extra possessions will come with fast breaks, and other than offensive fouls, it’s really tough to see Duke committing many turnovers on those. Irving has shown so far that he can take care of the ball, and the senior experience of Kyle and Nolan will really help keep the TO% down. With better athletes on the floor, Duke is better equipped to handle high pressure defenses in the half court, which gave them trouble at times last year. At the end of the day, this one is the most likely to not approach the two extremes.

OR% (rank)
2010: 40.3% (7)
2010 worst: 33.3% (151)
2010 best: 38.0% (25)

Duke’s offensive rebounding will go down this year, like it or not. Just based on style, Duke will likely not be as aggressive on the offensive boards, and Zoubs was head and shoulders above the rest of the COUNTRY last year in this category. The only other person on the same planet as Zoubs in OR% was Demarcus Cousins. Everbody else couldn’t even get close. Throw this in with some time spent in a smaller line-up and it just won’t be the same. The worst case scenario as presented by RTC is that Duke will lose Zoubs, and that the other guys won’t significantly improve their numbers. This doesn’t seem very likely to me as every tempo free guy will tell you that the biggest leap happens from freshman to sophomore year. Miles’ numbers last year were actually pretty good, and Mason was just ok. Kelly was horrible, but keep in mind that he barely played, and when he did, he sometimes came in for Kyle and played significant minutes at the 3 (remember those four forward sets), which would hurt his rebounding numbers. Add in his new found muscle, and he is going to get better. The best case scenario is that Miles, Mason, and Ryan all improve significantly. Even if they can’t get to Zoubs type numbers, they can get above the 10% that Lance added, which would significantly mitigate the loss of Zoubs and keep Duke in contact with its 2010 numbers.

FT Rate: FTA/FGA (rank)
2010: 37.9 (158)
2011 worst: 42.0 (75)
2011 best: 50.0 (5)

Let’s keep this in mind first. The factor in question is FT rate, which has nothing to do with FT%. Sometimes FTM/FGA is used, which does factor it in (and is my personal preference), but FTA/FGA has become more common, so we'll use that for now. The thinking behind this is that getting to the line is the most important part of FT shooting. You can’t make them if you don’t take them. Even if Duke drops off to just below 70% where it has been much of the past 5 years, it will have no bearing on FT rate. Even if they did drop off in FT%, Duke is not about to turn into Clemson or anything, so we can evaluate this for what it is. The fact is that by the standards of elite teams, Duke was horrible at getting to the line in 2010. They only managed to have just about the D-I average, when they got a lot of late game FTs in close games (think @BC, GTech in the ACC tourney, and Baylor in the Elite 8). During regular game play, Duke just didn’t get to the line and it was because there wasn’t a ton of penetration. That will change this year, the question is how much. This factor was largely ignored by RTC (a big omission despite being the least important factor), but the worst case scenario is still a lot better than last year. Despite playing bad opponents, Duke has not gotten to the line much in the first two games. This will improve, but maybe not to elite levels if things don’t go right. Duke has a lot of good shooters. There was really only one game last year (Gtown) where all three of the perimeter guys were cold. That is even less likely to happen with 5 guys to work with. Shooting will be a part of this team, and that doesn’t lead to free throws, even if it is better than last year. Duke does have the potential to be a top ten team in FT rate if we go to a smaller lineup, which could put 4 guys on the floor all capable of beating their man off the dribble (5 when Mason is in the game, but it’s really awkward watching him beat a center off the dribble). Once you get past the first guy, the defense is scrambling, which often leads to a trip to the line. The added fast breaks will help here, too.

gus

11-17-2010, 02:21 PM

Yes, but would it matter? Efficiency is much more important than raw scoring.

Are you being facetious, or did you really just argue that how many points you score doesn't matter?

I would argue that nothing matters more, except perhaps how many points your opponent scores.

MChambers

11-17-2010, 02:58 PM

Are you being facetious, or did you really just argue that how many points you score doesn't matter?

I would argue that nothing matters more, except perhaps how many points your opponent scores.
Increasing possessions gives you more chances to score, but it also gives your opponent more chances to score. So yes, you'll score more, but so will your opponent.

Duvall

11-17-2010, 03:05 PM

Are you being facetious, or did you really just argue that how many points you score doesn't matter?

I would argue that nothing matters more, except perhaps how many points your opponent scores.

Well, of course it's important to score more points than the other team. But there's no inherent value to scoring those points at a higher tempo. Scoring 60 points in a 60-possession game isn't automatically worse than scoring 75 points in a 75-possession game.

pfrduke

11-17-2010, 03:06 PM

Increasing possessions gives you more chances to score, but it also gives your opponent more chances to score. So yes, you'll score more, but so will your opponent.

I generally agree, with one caveat. If you play to form, the more possessions you play, the wider your margin for error will be. Maintaining an offensive rating of 115 over 75 possessions may be more valuable than maintaining one of 120 over 60 possessions because you are putting more pressure on the other team to continually score to keep up with you. I think better teams are (as a general matter) better served by higher possession games because the more possessions you play, the more likely you are to revert to your baseline performance, and the wider the gap on the scoreboard will be.

Kedsy

11-17-2010, 03:33 PM

Are you being facetious, or did you really just argue that how many points you score doesn't matter?

I would argue that nothing matters more, except perhaps how many points your opponent scores.

This is basic math. Let's say your team scores 1.10 points per possession and gives up 0.95 ppp, and you average 60 possessions a game due to your slow tempo. On average, you will win your games by a 66 to 57 score.

Now, say you speed things up to 100 possessions a game but due to turnovers in the fast-paced game you only score 0.93 points per possession. If your defense stays the same you will lose your average game by a score of 95 to 93.

You're scoring LOTS more points in the second example, but which situation would you prefer?

SCMatt33

11-17-2010, 03:35 PM

I generally agree, with one caveat. If you play to form, the more possessions you play, the wider your margin for error will be. Maintaining an offensive rating of 115 over 75 possessions may be more valuable than maintaining one of 120 over 60 possessions because you are putting more pressure on the other team to continually score to keep up with you. I think better teams are (as a general matter) better served by higher possession games because the more possessions you play, the more likely you are to revert to your baseline performance, and the wider the gap on the scoreboard will be.

That's true, but it's all proportional. What's "better"? A 60-40 win, or a 90-60 win. Either way, you beat you're opponents 3-2 on average. Let's put it this way, is your lead safer at 4:00 left if your up 10 and play slow or 15 and play fast? Sure, the 15 point lead looks better, but if there are extra possessions, it's easier to come back. Scoring margin, just like points, is subject to tempo, and a bigger margin doesn't always mean you were better.

airowe

11-17-2010, 03:44 PM

gus

11-17-2010, 04:01 PM

That's true, but it's all proportional. What's "better"? A 60-40 win, or a 90-60 win. Either way, you beat you're opponents 3-2 on average. Let's put it this way, is your lead safer at 4:00 left if your up 10 and play slow or 15 and play fast? Sure, the 15 point lead looks better, but if there are extra possessions, it's easier to come back. Scoring margin, just like points, is subject to tempo, and a bigger margin doesn't always mean you were better.

The end of game strategy you describe is a different scenario. There the plan is to limit the number of possessions to make it unlikely that your opponent can score enough points to take the lead. This has nothing to do with the relative strengths of the two teams. Properly executed this strategy should result in a win even if your team's efficiency is near zero and your opponent's is near perfect.

When the score is tied, a higher number of possessions favors the stronger team. To boil it down to a ridiculous extreme, if Duke were playing the Lakers, would you want the game to last 1 minute, or 48 minutes? I think Duke's odds are far better in the very short game that will have about 2 possessions. Similarly, last night - after two possessions, Duke was losing.

gus

11-17-2010, 04:05 PM

This is basic math. Let's say your team scores 1.10 points per possession and gives up 0.95 ppp, and you average 60 possessions a game due to your slow tempo. On average, you will win your games by a 66 to 57 score.

Now, say you speed things up to 100 possessions a game but due to turnovers in the fast-paced game you only score 0.93 points per possession. If your defense stays the same you will lose your average game by a score of 95 to 93.

You're scoring LOTS more points in the second example, but which situation would you prefer?

You're saying that you can't just look at one statistic alone without considering others. I don't think anyone is arguing differently.

pfrduke

11-17-2010, 04:11 PM

That's true, but it's all proportional. What's "better"? A 60-40 win, or a 90-60 win. Either way, you beat you're opponents 3-2 on average. Let's put it this way, is your lead safer at 4:00 left if your up 10 and play slow or 15 and play fast? Sure, the 15 point lead looks better, but if there are extra possessions, it's easier to come back. Scoring margin, just like points, is subject to tempo, and a bigger margin doesn't always mean you were better.

I agree with everything you say, but I wasn't necessarily talking about margin (even if I inarticulately expressed it that way); I was more talking about taking luck out of the equation, which is basically what gus is saying (much more clearly than I did) below.

When the score is tied, a higher number of possessions favors the stronger team. To boil it down to a ridiculous extreme, if Duke were playing the Lakers, would you want the game to last 1 minute, or 48 minutes? I think Duke's odds are far better in the very short game that will have about 2 possessions. Similarly, last night - after two possessions, Duke was losing.

The more possessions in a game, the less luck is involved, and the less critical perfect execution on every possession becomes. I think that if we play at a faster pace this season, we have room for wiggle room in the offensive efficiency numbers because a) the more possessions we play, the more likely we are to hit our average efficiency; and b) the same is true for our opponents.

CDu

11-17-2010, 04:12 PM

You're saying that you can't just look at one statistic alone without considering others. I don't think anyone is arguing differently.

Well, this whole discussion is stemming from a post that seemed to suggest that a reduction in offensive efficiency would be offset by more possessions resulting in more points, and thus we wouldn't be worse off than last year despite a loss in offensive efficiency. The subsequent posts were just clarifying that more possessions mean more points for both teams, so that logic didn't really hold. You have to consider the other team's efficiency as well, because more possessions means more scoring opportunities for the opponent, too.

SCMatt33

11-17-2010, 04:17 PM

The end of game strategy you describe is a different scenario. There the plan is to limit the number of possessions to make it unlikely that your opponent can score enough points to take the lead. This has nothing to do with the relative strengths of the two teams. Properly executed this strategy should result in a win even if your team's efficiency is near zero and your opponent's is near perfect.

When the score is tied, a higher number of possessions favors the stronger team. To boil it down to a ridiculous extreme, if Duke were playing the Lakers, would you want the game to last 1 minute, or 48 minutes? I think Duke's odds are far better in the very short game that will have about 2 possessions. Similarly, last night - after two possessions, Duke was losing.

That didn't come out clear, but I was assuming that the pace stayed the same. Team A had built the 10 pt lead at a slow pace and continued slow, and team B built it at a fast pace and continued fast. It's tough to argue efficiency within a single game, because if the score is tied, then the teams have played that particular game with the same efficiency up to that point, and assuming that the team keeps playing as well as they had in the last minute as the first 39 minutes, the number of possessions is irrelevant. If Duke were playing the Lakers and managed to be in a tie game with a minute left, I would want the pace to stay exactly the same as what got them there, whether it was fast or slow.

Kedsy

11-17-2010, 04:20 PM

I think you missed the point.

efficiency * possesions = points

blueintheface is arguing that a drop in the first factor will be compensated by an increase in the second. Duke will score more this season than prior season, despite the lower efficiency.

Yes, but would it matter? Efficiency is much more important than raw scoring.

Are you being facetious, or did you really just argue that how many points you score doesn't matter?

I would argue that nothing matters more, except perhaps how many points your opponent scores.

You're saying that you can't just look at one statistic alone without considering others. I don't think anyone is arguing differently.

Actually what I'm saying (or at least trying to say) is that efficiency is more important than total points scored. Your previous posts suggest you are arguing differently, but if I misread them, then we all agree.

MChambers

11-17-2010, 04:28 PM

I agree with everything you say, but I wasn't necessarily talking about margin (even if I inarticulately expressed it that way); I was more talking about taking luck out of the equation, which is basically what gus is saying (much more clearly than I did) below.

The more possessions in a game, the less luck is involved, and the less critical perfect execution on every possession becomes. I think that if we play at a faster pace this season, we have room for wiggle room in the offensive efficiency numbers because a) the more possessions we play, the more likely we are to hit our average efficiency; and b) the same is true for our opponents.
Pomeroy makes the luck point in explaining his ratings. I think he'd say that most observes place way too much emphasis on wins and losses and not enough on point spread and other factors. In any reasonably close matchup, luck will have a lot to do with the outcome.

By adding more possessions, the better team usually has a better chance to win. Note that this wasn't true last year for Duke, because of the lack of perimeter depth and the particular skills of many of the players. This year, we've got perimeter depth and many players who excel at uptempo basketball.

bjornolf

11-17-2010, 04:32 PM

The subsequent posts were just clarifying that more possessions mean more points for both teams, so that logic didn't really hold. You have to consider the other team's efficiency as well, because more possessions means more scoring opportunities for the opponent, too.

Of course, that doesn't take into account Duke's increase in athleticism, allowing for tighter defense resulting in steals and fast break layups and dunks. Fast break layups and dunks, which are high percentage shots, somewhat help make up for loss of overall efficiency in the offense. On top of that, while you're giving your opponent more chances with the ball, increased steals and breakaways thanks to higher defensive pressure from increased athleticism and depth in the back court help make up for some of that. If you give the opponent 10 more chances a game with the ball, but you deny them even a chance to shoot 6 of those times by turning them over, then you're really only giving them 4 more chances, while you might be getting 8 out of 10 more chances, 5 of them being breakaway layups and dunks. You have a delta of 4 extra "legitimate" scoring chances per time period over your opponent.

CDu

11-17-2010, 04:40 PM

Of course, that doesn't take into account Duke's increase in athleticism, allowing for tighter defense resulting in steals and fast break layups and dunks. Fast break layups and dunks, which are high percentage shots, somewhat help make up for loss of overall efficiency in the offense. On top of that, while you're giving your opponent more chances with the ball, increased steals and breakaways thanks to higher defensive pressure from increased athleticism and depth in the back court help make up for some of that. If you give the opponent 10 more chances a game with the ball, but you deny them even a chance to shoot 6 of those times by turning them over, then you're really only giving them 4 more chances, while you might be getting 8 out of 10 more chances, 5 of them being breakaway layups and dunks. You have a delta of 4 extra "legitimate" scoring chances per time period over your opponent.

Yes, and this was discussed in some of the earlier posts as well. Our offensive efficiency may or may not end up being worse than last year. On the one hand, we'll commit more turnovers and get fewer offensive rebounds per possession. On the other hand, we'll likely generate more fast break opportunities and "easy" scoring chances. It is unclear at this point which way our free throw scoring rate will go. But it is certainly possible that our offensive efficiency goes up in spite of fewer offensive rebounds and more turnovers committed.

gus

11-17-2010, 04:54 PM

Actually what I'm saying (or at least trying to say) is that efficiency is more important than total points scored. Your previous posts suggest you are arguing differently, but if I misread them, then we all agree.

I can see how it looks like my posts argue that your team's total points are more important than efficiency. That's a consequence of rapid fire responses and not really taking care to compose reponses properly.

What I reacted to was this assertion:

"If you're less efficient, it doesn't matter how much faster or slower you run, you'll still be worse." followed up by the condescending (and a bit ironic) statement "Lets use the stat right please. "

Ultimately what we care about is whether the team wins or loses, and the discussion (to me) is about the probability of Duke winning games. Claiming a team is "worse" suggests the team has a lower chance of winning.

In this context, offensive efficiency is telling, but it is not the most important thing to consider. Claiming that because Duke has a lower projected efficiency suggests that Duke has a lower probability of winning is very flawed reasoning.

Efficiency is a product of many factors, but let's isolate two: talent and strategy.

Duke's efficiency will decrease this year not because of a significant change in talent (I'd argue that it's actually greater) but because Duke is changing to a strategy that will result in a decrease in efficiency, both for Duke and its opponents. Despite decreasing efficiency, Duke's winning probability should increase, because we are increasing the number of expected possessions. A higher number of possessions favors this team, because its talent (particularly in the back court) will exceed most opponents. Conversely, I think this team could execute the same or a similar low possession count, half-court strategy the team employed last year, and have a higher efficiency as a result. But Duke will have a lower probability of winning.

Nugget

11-17-2010, 04:56 PM

I think whoever made the point that the RTC piece is a bit flawed because it gave substantially disproportionate attention to offensive rebound percentage in assessing offensive efficiency hit it head on. While our OR last year was amazing, the RTC piece fails to consider that our FG% last year (really, our 2 pt FG%) was historically awful, and that this was almost as much of a drag on efficiency as OR was a boon.

I think it highly likely that an anticipated drop in offensive rebounding, which could be expected to be about 5% fewer of our misses turning into offensive rebounds (2 or 3 per game, assuming we take about the same number of shots), and an expected increase in our turnover rate, could be offest by increases in our number of free throws attempted and a significant increase in our percentage of two point FGs made.

RTC overlook just how unusually “bad” last year’s team was in FT attempts and 2 point FG percentage (because we got so few easy shots, due to lack of a running game or much penetration, lack of interior offense from Zoubs/Lance).

Compare last year’s team against some of the other powerhouse Duke teams of the recent past:

2009-2010: 44% (2pt. FG); 22.5 FTA per game.
2001-2002: 57% (2pt. FG); 26 FTA
2000-2001: 55% (2pt. FG); 26 FTA
1998-1999: 57% (2pt. FG); 29 FTA
1997-1998: 53% (2pt FG); 27 FTA.

Leave out tempo changes and just assume we were to take the same number of shots this year as last. If this team is able to match/slightly exceed arguably the most comparable of those past clubs, 1997-1998 (it had Brand in the post, but he was a FR and injured much of the year), that would increase this year’s 2 point FG shooting from 44% to 54%, and raise scoring by 7 points per game, and result in 4.5 more free throws per game, increasing scroing by another 3 points per game.

So, adding 10 ppg in performance just from being more efficient in those two areas, would be enough to offset a “loss” of the points represented by grabbing 2-3 fewer OR and making 2-3 more turnovers per game (we committed 11 per game last year vs. an average of 13.5 in the other years listed above).

Kedsy

11-17-2010, 05:16 PM

Claiming that because Duke has a lower projected efficiency suggests that Duke has a lower probability of winning is very flawed reasoning.

Well, if your defensive efficiency stays the same, then if you have a lower offensive efficiency then you *do* have a lower probability of winning. By definition. One problem with the original article is it assumes our defensive efficiency will be the same or worse, which is not necessarily true (although we were a really good defensive team last year, so it might be true -- we just have to watch and see).

Efficiency is a product of many factors, but let's isolate two: talent and strategy.

Duke's efficiency will decrease this year not because of a significant change in talent (I'd argue that it's actually greater) but because Duke is changing to a strategy that will result in a decrease in efficiency, both for Duke and its opponents. Despite decreasing efficiency, Duke's winning probability should increase, because we are increasing the number of expected possessions. A higher number of possessions favors this team, because its talent (particularly in the back court) will exceed most opponents.

I agree that a higher number of possessions favors this year's Duke team. And I agree that if our offensive efficiency decreases it will be OK as long as our defensive efficiency increases (i.e., our opponent's offensive efficiency decreases) by the same amount or more.

But I don't think you should concede that our offensive efficiency will decrease. As others have pointed out, just because our turnovers increase and our offensive rebounds decrease, we could still be a better (i.e., more efficient) offensive team, due to improvements in eFG% and FT rate.

And, incidentally, offensive efficiency, as it's defined here (and in the article) is a product of four factors -- eFG%, turnover %, offensive rebound %, and FT rate. The things you mention (talent and strategy) may effect one or more of these four factors (e.g., if you speed up your pace your turnover % may increase) but they do not affect efficiency directly.

Conversely, I think this team could execute the same or a similar low possession count, half-court strategy the team employed last year, and have a higher efficiency as a result. But Duke will have a lower probability of winning.

Well, again, if our defensive efficiency remains the same and our offensive efficiency increases, by definition we would have to have a *higher* probability of winning. What I think you mean here is if this year's team plays a lower possession game, our defensive efficiency would decrease more than our offensive efficiency would increase (assuming our offensive efficiency would increase, which is your premise here but I'm not sure is correct). And if our defensive efficiency decreases more than our offensive efficiency increases, then you are right that Duke will have a lower probability of winning.

cato

11-17-2010, 05:26 PM

The subsequent posts were just clarifying that more possessions mean more points for both teams

That is simply not true. More possessions means the opportunity for more points, but if FG% or attempts go down, then those possessions may not lead to more points.

CDu

11-17-2010, 05:38 PM

That is simply not true. More possessions means the opportunity for more points, but if FG% or attempts go down, then those possessions may not lead to more points.

Well, I wouldn't use the term "simply not true," but rather "not always true." In the context of my point (i.e., holding defensive efficiency constant), more possessions mean more points. If you don't change anything else but increase the number of possessions for the opponent, then it is absolutely the case that the opponent gets more points. I apologize for not clarifying the "holding all else constant" part of my point.

BlueintheFace

11-17-2010, 05:42 PM

whoops, misread his premise. I understand efficiency, i swear. It is that darn long division that always kills me

cato

11-17-2010, 05:49 PM

In the context of my point (i.e., holding defensive efficiency constant), more possessions mean more points. If you don't change anything else but increase the number of possessions for the opponent, then it is absolutely the case that the opponent gets more points. I apologize for not clarifying the "holding all else constant" part of my point.

That's what I thought was lacking from most of the early posts. Why would we assume that defensive efficiency would stay constant? If increased tempo leads to decreased offensive efficiency in one team, it probably leads to descreased offensive effeciency in the other. One of the goals of pushing tempo is to force a significant decrease in the other team's effeciency, either by generating turnovers or poor shots.

gus

11-17-2010, 05:49 PM

Well, if your defensive efficiency stays the same, then if you have a lower offensive efficiency then you *do* have a lower probability of winning.

And if your opponents score more points than you, you lose. duh. ;) Offensive efficiency and defensive efficiency are both measures of how talent and strategy are affecting the game, so I think it's nonsensical to have as an assumption no change to defensive efficiency. Which isn't really different than what you go on to say.

One problem with the original article...

full disclosure: I was unable to open and thus haven't read the article.

But I don't think you should concede that our offensive efficiency will decrease.

I think it will decrease, but I agree that the drop cited from the article seems unlikely,

And, incidentally, offensive efficiency, as it's defined here (and in the article) is a product of four factors -- eFG%, turnover %, offensive rebound %, and FT rate. The things you mention (talent and strategy) may effect one or more of these four factors (e.g., if you speed up your pace your turnover % may increase) but they do not affect efficiency directly.

I think you're putting the cart before the donkey. I don't see how one can argue that talent and strategy won't affect offensive efficiency, but I don't understand what you mean by "directly".

What I think you mean here is if this year's team plays a lower possession game, our defensive efficiency would decrease more than our offensive efficiency would increase (assuming our offensive efficiency would increase, which is your premise here but I'm not sure is correct). And if our defensive efficiency decreases more than our offensive efficiency increases, then you are right that Duke will have a lower probability of winning.

I think both the strategy the team is executing will indeed 1) increase defensive efficiency and 2) decrease offensive efficiency. But you're missing my point-- increasing the number of possessions WITHOUT changing the relative offensive and defensive efficiency will increase the probability of a Duke win. 1) and 2) will both happen as a result of employing a strategy that increases Duke's chances of winning, and thus a drop in offensive efficiency is not a predictor for Duke's probability of winning (and this, whether Duke is "worse").

gus

11-17-2010, 06:22 PM

I can't be arsed to do the math, so I just ran an easy model instead to illustrate my point.

assumptions:

Team's offensive efficiency: 45
Team's defensive efficiency: 40

In 10,000 75 possession games, Team A won 75.6% of the games
In 10,000 100 possession games, Team A won 78.4% of the games

Increasing the possessions obviously increased Team A's chances.

Then I decided to run another set of assumptions. What happens if the expected score is unaffected? i.e, if the efficiency of .45 is for a 75 possession game, yielding an expected score of 33.75 (and 30 for the opponent). Increasing to a 100 possession game, and the expected score is still 33.75 to 30. What happens?

In 10,000 75 possession games, team A won 76.2% of the games.
In 10,000 100 possession games, team A won 73.3% of the games.

I didn't expect such a dramatic decrease in team A's chances, so I'm rethinking my assumptions. But I guess this makes sense: it can't be universally true that increasing possessions increases odds of winning for favored teams, if you assume that increasing possessions has an effect of efficiency.

CDu

11-17-2010, 06:29 PM

That's what I thought was lacking from most of the early posts. Why would we assume that defensive efficiency would stay constant? If increased tempo leads to decreased offensive efficiency in one team, it probably leads to descreased offensive effeciency in the other. One of the goals of pushing tempo is to force a significant decrease in the other team's effeciency, either by generating turnovers or poor shots.

I don't think the previous posts (I can say for certainty none of mine at least) were assuming that increased tempo alone leads to decreased offensive efficiency. The entire discussion as I understood it centered around the idea that the offensive efficiency would be decreased due to committing more turnovers (by replacing Jon with a freshman PG, but also us pushing tempo on offense) AND by getting fewer offensive rebounds (without Zoubek). So folks were not arguing that it's the increased number of possessions causing a decreased efficiency, but rather the style of play in which we'd be utilizing our possessions that would result in decreased efficiency.

Now, this is entirely based on the offensive side of the ball, and thus doesn't necessarily have an impact on the style of play of the opponent. If we simply try to run every time we get the ball but play the same style of defense as last year, it's not necessary So it's not impossible to assume an increased possession rate with no impact on the other team's offensive efficiency.

The gist of my point in the post you questioned was simply that you can't ignore the other end of the floor when talking increasing possessions. If you increase possessions you'll likely score more points (even if you reduce efficiency slightly). But if you ignore the other end of the equation (i.e., assume no change in defensive efficiency), then an increase in possessions also increases the opponents' points.

Of course, as I (and others) have noted in this thread, we are changing our defensive approach as well. And our defensive strategy may help boost our offensive efficiency. By pressuring the ball, we can force turnovers (reducing their efficiency) and generate more fast break opportunities (increasing our efficiency).

gus

11-17-2010, 06:40 PM

Okay, so I tinkered a bit more with my model. Increasing the possessions without increasing the expected score isn't realistic at all. I added another factor to it. Increasing the possessions increases the expected score, but not linearly. Using an arbitrary coefficient of .85 instead of .75 (needed to keep expected score flat) Team A's winning percentages for 10,000 games were 75/75.4% 100/75.8%.

So what does this indicate? Well, for one I'm a geek. But it also shows that increasing possessions is a winning strategy as long asit doesn't have a material impact on efficiency. In my model, a 33% increase in possessions with an efficiency drop of 25% (for both teams) is a losing strategy. A 15% drop in efficiency makes it a break-even strategy. A drop of less than 15% is a winning strategy.

loran16

11-17-2010, 07:13 PM

Pomeroy makes the luck point in explaining his ratings. I think he'd say that most observes place way too much emphasis on wins and losses and not enough on point spread and other factors. In any reasonably close matchup, luck will have a lot to do with the outcome.

By adding more possessions, the better team usually has a better chance to win. Note that this wasn't true last year for Duke, because of the lack of perimeter depth and the particular skills of many of the players. This year, we've got perimeter depth and many players who excel at uptempo basketball.

Eh, except that if this was the case, every team would try and play at the fastest tempo possible for their personnel. Which is not the case.

Adding more possessions should theoretically reduce variance. That's true. But the amount extra in a fast tempo team compared to a super slow team isn't enough to reduce it at a significant enough level to matter on a game by game basis.

In other words, Adding 20 more possessions (the difference between the near slowest pace and fastest pace) won't result in reducing occasional bad performances. It's just not a large enough sample size.

Playing Faster =//= better. Score more points sure. But if you're more inefficient, you're a worse team.

loran16

11-17-2010, 07:15 PM

ForkFondler

11-17-2010, 07:26 PM

You're showing a .4% difference per 10,000 games. NCAA BBall seasons are 40. You won't see any difference that is clearly caused by pace in that sample size.

I think the possessions-per-game issue can be illustrated with a simpler model:

With one posession (roll):

Four-sided die (average outcome = 2.5) vs six sided die (average outcome = 3.5) gives

One posession:

4 sided die wins 4/24, or 16.7 % of the

OT 8/24 or 33% of the time

6 sided die wins 12/24

Increasing the possessions (rolls) to 2:

4 sided die wins 12% of the time

OT 8% of the time

6 sided wins 80% of the time

Other thing being equal (they aren't, of course), a high score favors the team with a greater efficiency.

gus

11-17-2010, 08:25 PM

You're showing a .4% difference per 10,000 games. NCAA BBall seasons are 40. You won't see any difference that is clearly caused by pace in that sample size.

That's why I called it my break even point. Maybe if I restate my conclusion a little clearer...

The strategy is to increase the number of possessions in the game, with the hope of improving winning probability. The assumption is that increasing possessions decreases both sides' offensive efficiency.

If the efficiency drops by:

less than 15%, it is a winning strategy
by 15%, it's a wash
by more than 15%, it is a losing strategy

gus

11-17-2010, 08:32 PM

I think the possessions-per-game issue can be illustrated with a simpler model:

With one posession (roll):

Four-sided die (average outcome = 2.5) vs six sided die (average outcome = 3.5) gives

One posession:

4 sided die wins 4/24, or 16.7 % of the

OT 8/24 or 33% of the time

6 sided die wins 12/24

Increasing the possessions (rolls) to 2:

4 sided die wins 12% of the time

OT 8% of the time

6 sided wins 80% of the time

Other thing being equal (they aren't, of course), a high score favors the team with a greater efficiency.

my model wasn't actually that much more complicated. The team with the higher efficiency will *always* be favored. I think your conclusion should be that increasing possessions increases the likelihood that the more efficient team wins.

MChambers

11-17-2010, 08:36 PM

Eh, except that if this was the case, every team would try and play at the fastest tempo possible for their personnel. Which is not the case.

Adding more possessions should theoretically reduce variance. That's true. But the amount extra in a fast tempo team compared to a super slow team isn't enough to reduce it at a significant enough level to matter on a game by game basis.

In other words, Adding 20 more possessions (the difference between the near slowest pace and fastest pace) won't result in reducing occasional bad performances. It's just not a large enough sample size.

Playing Faster =//= better. Score more points sure. But if you're more inefficient, you're a worse team.
I think we're agreeing. All things being equal, a better team will win slightly more often with a faster tempo. Not all things are equal, as K realized last year, with a shallow team that also was more effective at a slower tempo. Good thing Roy hasn't figured this out yet (and probably never will).

ice-9

11-18-2010, 06:07 AM

Guys, it isn't that complicated....

Net impact from last year to this year

OFFENSE
Effective FG% + Kyrie to penetrate and assist + sharp shooters Curry and soph Dawkins
Turnover % - We don't have Jon and faster style = more mistakes
Off Rebounding % - We don't got da beard
FTA/FGA + We'll penetrate more and draw more fouls
Overall Can be better, can be worse

DEFENSE
Effective FG% - We're not as tall, plus over gambling can lead to open buckets
Turnover % + Aggressive D will lead to more fast breaks
Off Rebounding % +/- No Zoubs or Lance but maybe Plumlees & RK can compensate
FTA/FGA +/- Aggressive D can lead to penetration but Plumlees & RK are better blockers
Overall Can be better, can be worse

OVERALL - can be better, can be worse. We just have to see whether the advantages of this year's strategy will outweigh the disadvantages relative to last year.

That doesn't mean of course that we should play like last year. Last year's strategy was perfect for the personnel that we had -- Coach K maxed out that team. This year's strategy is better suited for the team that we do currently have.

So let's just see what happens! Based on Princeton and Miami (OH), so far so good. :)