Duke 66, UVa 65 - Postgame thread

[SavDukeGrad]

I just rewatched the game. Our defense was certainly far from perfect. But we held UVa scoreless for the final 3:10, when it mattered. Love that we slapped the floor, then backed it up. We forced them into tough shots, and thankfully, they missed.

I have to admit, I’m still feeling a little bit giddy about this win, even today. Now I hope we can re-focus for Syracuse and continue our winning ways. And, of course, give Jimmy B. the big fat L he deserves!

[hustleplays]

Yeah, what uh_no said.



On what else besides scoring can you ultimately judge defense? Even if you say "we played good defense, but they played great offense," doesn't that just mean our D wasn't good enough to stop them?



Except in the "classic" days, we didn't often let our opponents shoot an eFG% of 55%.



Clemson, NCSU, GaTech, Bellarmine, maybe Michigan State. Just because some of those games weren't close and none of those teams were as good as Virginia doesn't mean we didn't play better D.



This is true. Our first half adjusted dRating (0.98 ppp) was the equivalent of the #92 defense in the country. Our second half adjusted dRating (0.87 ppp) was the equivalent of the #1 defense in the country. But, as you allude, both halves count when you're talking about performance in a game.



Putting aside that I don't completely agree with the conclusions of that Pomeroy article, I do agree with your comment here. Duke did a great job keeping Virginia from shooting threes, especially Huff and Murphy. As I mentioned in my initial post in this thread, that one element was probably what won us the game. That said, just because we did one thing great on defense doesn't mean our defense was great.





If an opponent consistently hits a particular type of shot but we keep letting them take it, it's not good defense. If a team "just doesn't miss" over a length of time, it ceases to be "luck."



Yes. Maybe not "bad," but certainly mediocre.

I'll also say that my "eye test" said the same thing in the first half. Matthew was consistently out of position on defense, allowing Hauser to do whatever he wanted. Jaemyn missed several assignments and was generally not where he was supposed to be, at least when he wasn't letting people blow by him. Mark got caught in no-mans-land a few times. Jeremy and DJ were a little late with their switchbacks. We (especially Jaemyn) did play much better D in the second half, but that doesn't erase the poor defensive play in the first.



I agree. But there are four factors to defensive play. We did well in two of them (foul rate and turnovers) and not-so-well in the other two (opposing shooting and rebounding). To me, that says good but not great defense, not best D of the season.



The adjusted ratings attempt to take into account how good your opponent is, so we can at least try to compare apples-to-apples. Unadjusted, this was our 6th-worst opposing points per possession of the season. Because Virginia is so good at offense, it obviously wasn't our 6th-worst defensive performance. It wasn't our best, either.

Which is better D, if you hold a team that should score 80 to 75 or if you hold a team that should score 70 to 50 (assuming the same number of possessions in the game)? Even though the first opponent is a much stronger offensive team, I'd argue the second defense was much better.



Maybe. The way I look at it is our eyes can't see everything. I know what I think after watching the game and then I go to the stats to see if I was right or if my eyes deceived me. If the stats disagree with my eyes, I consider if there's a reason our play was better or worse than I thought. Unless I see a compelling reason, I trust the stats over my eyes. YMMV.



What I find insufferable is people who think they know best and won't brook any dissent, yet provide absolutely no evidence for their assertions.

Your eyes aren't the ultimate arbiter. Or do you think David Copperfield really made the Empire State Building disappear?

Kedsy, I greatly appreciate your steady stream of illuminating stats. Our board would be much poorer without them. Thank you! Your stats often challenge my “eyeballs.” I put eyeballs in quotes because when we say eyeballs we aren’t only reporting specific instances, we are also asserting generalizations (such as “we played great defense!”) that we form based on our own inferential, synthetic judgment as to the totality of our observations. I find that stats, and in this context, YOUR stats, to be helpful — and very interesting — checks on my synthetic judgments, not on what I know my eyeballs saw.

I think that most reasonable people know that “eyeballs” (I.e., what we see AND our synthetic judgments) contribute perspectives that stats may miss, but they also know that “eyeballs” are far from infallible.

Baseball’s recent history with analytics is compelling to me. Billy Bean and co have dispositively shown that stats add an indispensable element to understanding what individual players and teams do. The baseball analytics also highlight the significant role that our psychological and ideological lenses play when we humans observe things. I’m pretty sure that I am not immune.

Kedsy, one of my “eyeball” observations is that over the recent years, you have become more hospitable toward affirming “eyeballs” as playing a legitimate role. If true, I am happy about this and appreciate it. Please keep giving us great stats.

[jimsumner]

Like you I generally put low stock in autobackpatting about the collective intellect of the group here, but I did think it relevant to the issue at hand, this time.

Language prescriptivism is itself one of the most intense areas of elitism in American life, and I've spent my whole adult life combating the isms on which linguistic prejudice is based... but that gets us into territory categorized by the bosses as "public policy" so,

You know you're on DBR when someone uses "Language prescriptivism" in a post designed to be anti-elitist.

And for the record, I think we're the most smartest message board on the whole interweb thingee.

[weezie]

You know you're on DBR when someone uses "Language prescriptivism" in a post designed to be anti-elitist.

And for the record, I think we're the most smartest message board on the whole interweb thingee.

Unless those Mars Perseverance Rover folks have one...they're something else.

[Bay Area Duke Fan]

Quote Originally Posted by Bay Area Duke Fan View Post
When comparing your stats from game to game, is the strength of the opponent taken into account in any way?

Yes.

Please explain how it's reflected.

[-jk]

Please explain how it's reflected.

Uh, oh...

-jk

[uh_no]

Please explain how it's reflected.

KP stats evaluate a strength of each teams offense and defense to minimize the difference between the assigned ratings and the observed outcomes. this method is applied to all stats (shooting %, TOs, FT%, boards). and for each, it's like one giant system of linear equations, and we find the coefficients for each team that minimizes the total error.

[Nugget]

KP stats evaluate a strength of each teams offense and defense to minimize the difference between the assigned ratings and the observed outcomes. this method is applied to all stats (shooting %, TOs, FT%, boards). and for each, it's like one giant system of linear equations, and we find the coefficients for each team that minimizes the total error.

Is an easier way of understanding this that the “adjusted” numbers Kedsy cites for points per possession scored or given up in any given game mean they are “adjusted” for the opponent’s relative quality?

[Kedsy]

Is an easier way of understanding this that the “adjusted” numbers Kedsy cites for points per possession scored or given up in any given game mean they are “adjusted” for the opponent’s relative quality?

Yes. I use KenPom's offensive and defensive efficiency numbers for our opponent and adjust our "raw" points per possession numbers accordingly.

[uh_no]

Is an easier way of understanding this that the “adjusted” numbers Kedsy cites for points per possession scored or given up in any given game mean they are “adjusted” for the opponent’s relative quality?

yes.

The kenpom-provided numbers are what you would be expected to do vs an "average" opponent. So if it says our offense is 120, that means we'll score 120 points in 100 possessions against an average opponent. If an opponents defense is an 90, it means they're expected to give up 90 points in 100 possessions. The average is usually "around" 100, so in that case, if a 120 offense (20 better than average) went up against an 95 defense (5 better than average), you'd expect the offense to score 115 points (20 - 5 = 15 better than average)

The adjustment works backwards, too, allowing us to take an observed result and figure out what numbers WOULD have been needed for that outcome to be expected. If we scored 105 (5 better than average) against a team that has a defense of 90 (10 better than average), we would expect that team to score 115 (10 + 5 = 15 better than average) against an average team.

All KPs stats are computed roughly this way, and thus we can apply the same principles in reverse on the observed game data.


Now, none of this means raw performance isn't important or eye test is useless. The eye test says we played great stretches for long portions of the game. The numbers say we need to do a couple things better more consistently throughout the game, and not just for stretches. They're both right.

[BandAlum83]

After going 9-23 from 3 last night (39.1%), Duke is now 151-421 from 3 this season as a team which is good for 35.9%.

Last night, Matthew Hurt shot a blistering 5-8 from 3 (62.5%). For the season, Matthew Hurt is 45-96 from 3 which is good for 46.9%!

Hurt’s shot isn’t all that pretty, but his stats sure are!

And although loonardi’s opinion doesn’t matter, it does demonstrate that Duke is still in the hunt. That means when we make it in, it won’t come out of left field making people say “Duke’s only in for the name!” (I know, they will anyway)

[elvis14]

Awesome win, refs swallowed the whistles all game but we toughed it out. High scoring first half, low scoring second half, it didn't matter we just did what we needed to get the W. We'll have time to break down the stats later but for now this win was all about toughness, executing for 40 full minutes, and making plays. Let's GOOOOOOOO!!!!!!

I was gobsmacked at how that elbow from Hauser wasn’t a flagrant when it’s been a flagrant all year.

Those are our turnovers the last 3 games. That’s called playing smarter, harder and more disciplined.

If Huff had a Duke uniform on...he’d have 5 fouls called on him before warmups. Me no likely his game.

Coleman...Brake...seizing the moment. Kudos to you, young men.

They didn’t even review the screen by Huff that laid out Moore. Definitely leaned his shoulder in.

UVA has mastered how to dupe the officials into thinking that they don’t foul because they keep their arms up while moving their bodies into offensive players.

Some observations: I believe that's one area of officiating that needs cleaning up. Some teams seem immune to the whistle while others don't.

Late to posting about this game. Really happy with the win, we really needed it optically to improve our chances of getting into the NCAA tournament. Lots of guys contributed to this win and I thought it was a great team effort. I think our defense is getting better over time and we will keep winning if that trend continues. IMHO, not having Jalen improves the defense. One of the best things about this win was that even though we played well and won the game, it's obvious that there is still tons of room for improvement and if we can make the tournament we will have a few weeks go continue our growth.

As you can guess from the quoted posts above, I took exception to the way the game was called. Every single time I watch UVa play, I have the same reaction: it's not that they don't foul, it's that the fouls go uncalled. This game was no exception. When I want them to win it's fun to watch, like having a cheat code. When they are playing Duke, it makes me want to throw stuff at my TV. For years, I've noticed that they bump the man with the ball when he comes off a screen. This is a blocking call against every other team. It's a huge advantage to be allowed to do this without being called. Every time Jay Bilas would repeat that UVa doesn't foul much I'd correct him with "UVa doesn't get called for fouls much". They ugly up the game on an off the ball (don't even get me started with the off the ball contact). It's effective and I commend them for their results but it's time to remove the cheat code and get back to hoops that flow. Even when the game is called the same on both sides, when it's called in a way that matches one teams style of play more than the other, it helps that team.

I think we need a separate thread for the stats vs eye_test holy war. Someplace that whenever these posts show up in post game threads, they can be removed to clean up the post game. I still remember the first time I became aware of these arguments, reading a Cincinnati Reds forum in the mid-90's. My take on it is simple, I use and enjoy both. Often, for me, they are in agreement and when they aren't it's interesting sometimes to see why (I've rewatched games before because my eye test was so far off from what the stats were showing and I wanted to try and figure out why).