Page 12 of 12 FirstFirst ... 2101112
Results 221 to 231 of 231
  1. #221
    Quote Originally Posted by DukieInBrasil View Post
    thanks for your analyses, pretty fascinating stuff and great for putting things in perspective. The one question i have for you is: what about 3FG% and rate? I assume that part of the eFG% and FTR success that Kansas had was due in part to having 3 seriously beefy dudes, even if none of them really destroyed us, and which required lots of attention, which may have opened up space for 3FGs and/or driving. I see that Kansas shot 4-9 (44.4%), which is a great % but not a lot of shots. So where does KU's 3FG% and rate impact your analysis?
    Well, in theory, all that stuff (as you say) would be wrapped up in the eFG%.

    FWIW, Kansas' 3FG% of 44.4% would have been the 5th highest last season, and their %threes of 18.0% would have been the lowest last season. So, I guess that's something else our defense did very well the other night, chasing Kansas off the three-point line.

    Quote Originally Posted by DukieInBrasil View Post
    I think you missed the entire point of Kedsy's argument, that he was analyzing the exact same possession outcomes from last season to this one game, and that this game would have stood as one of our worst in those possessions. While the turnover rate would have been the best, the weakness of the rest of the defensive metrics pointed to meant that the overall defensive performance Tues night was good but not particularly great. It was good enough to get the win, and that's what i care about.
    Kedsy's point though was that the likelihood that the turnover rate being an outlier that might not be approached again is pretty high, and that the rather poor performance of the other defensive metrics leaves room to believe that there remains a good deal of work to be done on the defense. I for one think that K and staff are all over this and will in fact work hard to get the team to buy in to defensive strengths so that the relative weaknesses that Kedsy pointed out will improve.
    Thanks for saying this so well. The bolded part above is exactly why I looked into this. If someone could promise me that for the entire season we'd force turnovers on 35% of opponent possessions, I wouldn't be particularly worried about the other stuff.

    But considering the highest percentage of turnovers Duke has ever forced over an entire season in the Coach K era is 27% (in 1987), I think how we perform in the non-turnover possessions is worth looking at. (Because, for example, if we'd only forced turnovers on 27% of possessions instead of 35%, at the rate we gave up points in non-turnover possessions, Kansas would have had 8 more points and this probably would have been a loss.)

    Quote Originally Posted by HereBeforeCoachK View Post
    I did not miss the point. I know what he was comparing, and my point is that it is not relevant to the game of basketball as it plays out on the hardwood. It's just another way to carve numbers up. It's interesting, but not meaningful.

    A defense that is high risk, playing in the passing lanes, double teaming, pressing, etc, is going to create more turnovers and also give up a higher percentage of possessions where that risk does not pay off. That's a given, born out by every sport on the planet - but to take out the turnovers and think you have something meaningful on a PPP basis is to create a fallacy.
    Last season, in a game against Stetson, Duke forced turnovers on 31.2% of Stetson's possessions. On the possessions that did not end in a turnover, Stetson only scored 0.85 points per possession. That's good defense, all around.

    Against Kansas, we forced turnovers on 34.9% of Kansas's possessions. On the possessions that did not end in a turnover Kansas scored 1.27 points per possession.

    Obviously Kansas and Stetson can't be fairly compared when it comes to strength of opponent, but there's not that big a difference between 31% and 35% and there's a huuuge difference between 0.85 ppp and 1.27 ppp. You can keep saying, over and over, that it's not meaningful, but I don't think that many people would agree with you.

    Quote Originally Posted by HereBeforeCoachK View Post
    And another thing that is key: a turnover wipes out 1.0 possessions. Shooting percentage dips only wipe out a fraction of a possession. The only stat that wipes out an entire possession (the only exception being if a team shoots, misses, gets their rebound, and then has a TO - which is a very small percentage of plays) is turnovers. Thus for offense and defense, it's very likely the most important stat. This is very basic math.
    Basic math, possibly. Incomplete math, almost certainly. Because, for example, most teams in most games take 3 to 5 times as many field goal attempts as they have turnovers. Meaning you are understating the value of shooting (vis-a-vis turnovers) by 3x to 5x.

  2. #222
    Jordan Sperber had a mini-breakdown of Duke's approach to doubling the post against Kansas. In short, the rotations were really good and also Kansas was super predictable.

    https://twitter.com/hoopvision68/sta...02315844657157

    Also, there's a quick clip of Cassius Stanley's weakside defense, which was very good for a freshman in his first game.

    My biggest question heading into the season was interior defense (now it is more on the offensive side, ha!). Kansas and their emphasis on post entry was giving me all sorts of anxiety. But none of the Kansas big men were capable to getting anything going. The defensive effort from Duke and the play of Carey, Hurt, White, and Stanley on that end of the court was super encouraging stuff. It will be interesting to see how this shapes up against a different opponent that doesn't methodically try to pound the ball down to the post every trip down the court.

  3. #223
    Join Date
    Oct 2009
    Location
    Durham
    Quote Originally Posted by Kedsy View Post
    Well, in theory, all that stuff (as you say) would be wrapped up in the eFG%.

    FWIW, Kansas' 3FG% of 44.4% would have been the 5th highest last season, and their %threes of 18.0% would have been the lowest last season. So, I guess that's something else our defense did very well the other night, chasing Kansas off the three-point line.



    Thanks for saying this so well. The bolded part above is exactly why I looked into this. If someone could promise me that for the entire season we'd force turnovers on 35% of opponent possessions, I wouldn't be particularly worried about the other stuff.
    I think the mitigating factor is whether we attempted to force turnovers specifically because it was early. Maybe thought K would be especially susceptible to turnovers, and took as much advantage of it as possible while compromising other things. That would mean in later games we wouldn't force as many turnovers, but also wouldn't play defense as poorly. It's somewhat of a stretch, but not out of the realm of possibility.

    We will see.
    1200. DDMF.

  4. #224
    Quote Originally Posted by uh_no View Post
    I think the mitigating factor is whether we attempted to force turnovers specifically because it was early. Maybe thought K would be especially susceptible to turnovers, and took as much advantage of it as possible while compromising other things. That would mean in later games we wouldn't force as many turnovers, but also wouldn't play defense as poorly. It's somewhat of a stretch, but not out of the realm of possibility.

    We will see.
    I think it much more likely that the 34% TO rate was an extreme outlier due to KU having 1 guard and them being unusually careless with the ball. It's not like last year's team wasn't really good/aggressive on defense--recall the extraordinary steals figures Zion and Tre were generating for much of the season--and we only ended up with a 19% TO rate for the year. I understand part of this was probably a function of Tre's and Zion's injuries, but not entirely.

    My anecdotal sense is it's simply become more difficult to generate lots of turnovers than it used to be -- that with the AAU world of more guys growing up handling the ball, plus tighter whistles, players are less likely to commit TOs now than in the 80s. This is compounded for Duke by the increase in one and done players who take time to learn to be effective team defenders. We haven't exceeded a 20% TO rate for a season since 2013.

    Even teams that fully sell out to pressing and have led the NCAA in TO rate in the last 10 or so years (e.g., Press Virginia, Auburn, Stephen F. Austin, VCU) have topped out around 25-28%. No one playing regular half-court MTM like we do can hope to being close to that.

    We can't reasonably expect to be much more than 20% on TO rate. So, I think Kedsey is right, we absolutely need to tighten up the performance on non-TO possessions.

  5. #225
    Join Date
    Feb 2007
    Quote Originally Posted by uh_no View Post
    I think the mitigating factor is whether we attempted to force turnovers specifically because it was early. Maybe thought K would be especially susceptible to turnovers, and took as much advantage of it as possible while compromising other things. That would mean in later games we wouldn't force as many turnovers, but also wouldn't play defense as poorly. It's somewhat of a stretch, but not out of the realm of possibility.
    We just mostly played our standard defense -- ball pressure, ball denial, icing screens, containment drops by bigs, switching small/small -- and added aggressive post doubles as an opponent-specific adjustment.

    Forcing turnovers is just part of our standard defense.

  6. #226
    Join Date
    Feb 2007
    Location
    Richmond, VA

    New stat

    The new stat I like the most is calculated as follows:

    X = ((Duke 2 pt FG% * # of Duke 2 pt FG) - (Opp 2pt FG% * # of Opp 2 pt FG))*2 + ((Duke 3 pt Fg% * # of Duke 3 pt FG) - (Opp 3 pt FG * # of Opp 2 pt FG))*3 + ((Duke FT% * # of Duke FT) - (Opp FT% * # of Opp FT))

    You need to put this stat in a conditional equation : If (X>1) = W else L.

    This is summed over the season and

    if the total W is greater than the total W for the other ACC teams during the regular season than the overall stat = ACC regular season champion.
    if the total W is greater than the total W for the other ACC teams during the acc tournament than the overall stat = ACC tournament champion.
    if the total W is greater than the total W for the other NCAA teams during the NCAA tournament than the overall stat = NCAA national champion.

  7. #227
    Quote Originally Posted by MarkD83 View Post
    The new stat I like the most is calculated as follows:

    X = ((Duke 2 pt FG% * # of Duke 2 pt FG) - (Opp 2pt FG% * # of Opp 2 pt FG))*2 + ((Duke 3 pt Fg% * # of Duke 3 pt FG) - (Opp 3 pt FG * # of Opp 2 pt FG))*3 + ((Duke FT% * # of Duke FT) - (Opp FT% * # of Opp FT))

    You need to put this stat in a conditional equation : If (X>1) = W else L.

    This is summed over the season and

    if the total W is greater than the total W for the other ACC teams during the regular season than the overall stat = ACC regular season champion.
    if the total W is greater than the total W for the other ACC teams during the acc tournament than the overall stat = ACC tournament champion.
    if the total W is greater than the total W for the other NCAA teams during the NCAA tournament than the overall stat = NCAA national champion.
    What is all that translated into common English?

  8. #228
    Quote Originally Posted by Indoor66 View Post
    What is all that translated into common English?
    Just win, baby!

  9. #229
    Join Date
    Feb 2007
    Location
    Raleigh
    Quote Originally Posted by Indoor66 View Post
    What is all that translated into common English?
    9F 9F

    Because one 9F is only two characters so it's worth repeating.

    [redacted] them and the horses they rode in on.

  10. #230
    Join Date
    Feb 2013
    Location
    Cambridge, MA
    Quote Originally Posted by HereBeforeCoachK View Post
    I did not miss the point. I know what he was comparing, and my point is that it is not relevant to the game of basketball as it plays out on the hardwood. It's just another way to carve numbers up. It's interesting, but not meaningful.

    A defense that is high risk, playing in the passing lanes, double teaming, pressing, etc, is going to create more turnovers and also give up a higher percentage of possessions where that risk does not pay off. That's a given, born out by every sport on the planet - but to take out the turnovers and think you have something meaningful on a PPP basis is to create a fallacy. This is like removing sack plays from a defense and seeing how well they did when they blitzed but did NOT get to the QB. It's interesting, but it's not how the game is played. I'm sure Kedsy' stats are correct, but the game is not scored on PPP outside of turnovers. It's scored on all possessions. FG defense, blocked shots, defensive rebounds, fouls committed all are important - but again, the only thing that wipes out 100% of a possession is a turnover.
    I completely agree that aggressively pursuing turnovers is likely have negative ramifications for possessions that don't end in turnovers. However, I don't think that "defensive performance in non-turnover possessions" is entirely meaningless. It isn't a commonly discussed stat, so I personally don't have a good context for it, but I don't know if that makes it meaningless.

    Take your football analogy. I suspect that well-coached NFL teams who blitz frequently pay lots of attention to what happens when they don't get to the quarterback. Surely, an unsuccessful blitz that ends up with an incomplete pass or short gain is a lot better than one that ends up with a touchdown.

    Quote Originally Posted by HereBeforeCoachK View Post
    And another thing that is key: a turnover wipes out 1.0 possessions. Shooting percentage dips only wipe out a fraction of a possession. The only stat that wipes out an entire possession (the only exception being if a team shoots, misses, gets their rebound, and then has a TO - which is a very small percentage of plays) is turnovers. Thus for offense and defense, it's very likely the most important stat. This is very basic math.
    I am not sure I follow. I agree that a turnover is an excellent outcome because it ends a possession without points, but so does a missed shot and defensive rebound. A live ball turnover is probably a bit better than a defensive rebound because it may lead to a high percentage transition shot, but I don't see how it is any better at preventing points.

    Put another way, all other things being equal,would you prefer to have a defense that

    a) leads the country in turnover percentage, but is average in effective field goal percentage, or
    b) leads the country in effective field goal percentage, but is average in turnover percentage?


    #####################


    For what it is worth, there doesn't seem to be a strong correlation between turnover percentage and effective field goal percentage in college basketball.* There are teams who force a lot of turnovers, but allow their opponents to shoot a high percentage when they don't turn them over. There are also a roughly equal number of teams who force a lot of turnovers AND hold their opponents to a low effective FG%. You can probably guess which group tends to better defensively overall. After one game, it remains to be seen which type of team Duke will be in 2019-20.

    One potential cause for optimism is that Duke's effective FG% of 50% against Kansas MAY not have been as mediocre as Kedsy originally suggested. While an eFG% of 50% is pretty mediocre on the surface (the average eFG% across D1 was 50.7% last season), we should probably consider that KU is not an average opponent. While it remains to be seen how good KU will be offensively this year, they have consistently been one of the top teams with respect to eFG% in recent years. In the past 4 years, they have been in the top 12 in the country 3 times with an average eFG% of 55.5% over this period. If the the 2019-20 KU team turns out to have an eFG% around 55%, holding them to 50% is actually pretty solid.

    On the other hand, Duke's defensive rebounding rate and foul rate were pretty poor, even after adjusting for the level of competition.




    *I plotted turnover percentage vs eFG% for all high major teams in the 2019 season and the r-squared value is 0.003. To me, that implies that there is little correlation between turnover percentage and eFG%. I can show the plot if anyone is interested.

  11. #231
    Quote Originally Posted by House P View Post
    I am not sure I follow. I agree that a turnover is an excellent outcome because it ends a possession without points, but so does a missed shot and defensive rebound. A live ball turnover is probably a bit better than a defensive rebound because it may lead to a high percentage transition shot, but I don't see how it is any better at preventing points.
    Some very interesting comments....with respect to the one comment above....a turnover is almost always before ANY shot is taken....and a turnover is one event. A missed shot and a D rebound are two separate events. Therefore, it takes two events to equal the one turnover creation - which means neither has the impact of the single event that can end a possession by itself.

Similar Threads

  1. MBB: Kansas 85, Duke 81 (NCAA Qtrs) Post-Game Thread
    By JBDuke in forum Elizabeth King Forum
    Replies: 427
    Last Post: 04-15-2018, 09:08 PM
  2. Replies: 335
    Last Post: 03-25-2018, 07:30 PM
  3. MBB: Kansas 77, Duke 75 Post-Game Thread
    By JBDuke in forum Elizabeth King Forum
    Replies: 193
    Last Post: 11-18-2016, 10:19 PM
  4. MBB: Duke 68, Kansas 61 Post-Game Thread
    By JBDuke in forum Elizabeth King Forum
    Replies: 164
    Last Post: 11-27-2011, 10:38 PM
  5. MBB: Duke 82, Kansas State 68 Post Game Thread
    By Bob Green in forum Elizabeth King Forum
    Replies: 135
    Last Post: 11-27-2010, 11:08 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •