PDA

View Full Version : "DataBall" vs. Old School (Real World Example Spurs vs. Cavs)



KnicksorBust
02-07-2014, 08:22 AM
http://grantland.com/features/expected-value-possession-nba-analytics/

After Duncanís screen frees up Parker to attack Zeller, EPV actually decreases as Parker penetrates through the midrange closely marked by Zeller. But as he gets close to the basket, the EPV surges to 1.36. Parkerís dribble drive has already elevated that value of the possession by 0.39 points ó but heís not done. He increases the value of the play further when he fires that crazy baseline pass to Leonard, standing open in the corner. EPV accounts for both Leonardís great corner shooting prowess and that he is wide open. As a result, the EPV peaks at 1.75 as Parker throws the game-winning assist. There is a slight decrease in value, to 1.58, as Dion Waiters frantically attempts to close out Leonard, but Waiters is too late.

The article itself could probably spur (no pun intended) another 2-3 threads (like Ricky Rubio being the worst player in the NBA last season by EPV) but for now what jumped out to me was the fact that statistically the article promotes the value of the 3 but personally I would much rather Parker have gone up with the layup.

The Spurs were down 2 with 9.5 seconds left to play.

Should Parker have shot the ball or passed to Leonard?

theducksmuggler
02-07-2014, 08:51 AM
As a coach im telling drive and take the easy 2 if its open but at the same time i say if you drive and get doubled dont be afraid to kick it to the open man...i dont think numbers has to do with much this type of play happens all the time you just take the better more open shot in my opinion...if you have video of the specific play id like to watch it\

EDIT: after watching the play that was a great written play by POP he had 3 options on the play from the get go...
He ran the pick and roll with Duncan they decided to stay on parker while Duncan rolled but he wasnt able to pass it off then the 3rd option is Leonard in the corner so if Waiters gets lost/goes to help on defense which he obviously did then Parker could kick it out for the winner which happened...overall amazing play if you watch it

D-Leethal
02-07-2014, 10:21 AM
Turning every moment of a basketball game into a series of numbers is such a stupid exercise.

JasonJohnHorn
02-07-2014, 11:15 AM
This is interesting.

I say drive at Zeller, because you have a chance to draw a foul that was as well, and a foul + 2 is better than +2 when that +2 can miss either way.

I realize the numbers you give likely account for the foul and additional free-throw, but I value the ACTUAL foul. Especially if you are playing against a team whose big man is crucial (not Zellar).

In those instances, in terms of simply converting a basket, it is essentially up to the player. Some you trust, some you don't. Parker I'd trust. Chris Paul I'd trust. Jose Calderon I trust (I'm not just speaking to elite PGs here), Brandon Jennings.... I DO NOT TRUST!

Some guys have good judgement, some don't.

John Walls Era
02-07-2014, 11:16 AM
Turning every moment of a basketball game into a series of numbers is such a stupid exercise.

Not if it helps you win. But these #s are only decently telling.

D-Leethal
02-07-2014, 11:47 AM
Not if it helps you win. But these #s are only decently telling.

I'm all for teams hiring analytical departments for stuff like this. You have the resources and should have as much information as you can as a half a billion dollar business. As a fan, its incredibly underwhelming to me. I appreciate the ambiguities of the game, I don't want to try and eliminate them.

What I got most out of this, is that even when you break down a possession statistically you are still going to leave out key parts of the play statistically and you are still not properly weighing each part of the play statistically. The Duncan screen was arguably the most important part of the play and Duncan gets zilch, Parkers penetration and collapsing of the defense only gets him an assist in the end etc.

If were going to microanalyze possessions like this, were still a long ways away from properly weighing the importance of each player's contribution to the successful possession.

Rentzias
02-07-2014, 12:51 PM
I'm all for teams hiring analytical departments for stuff like this. You have the resources and should have as much information as you can as a half a billion dollar business. As a fan, its incredibly underwhelming to me. I appreciate the ambiguities of the game, I don't want to try and eliminate them.

What I got most out of this, is that even when you break down a possession statistically you are still going to leave out key parts of the play statistically and you are still not properly weighing each part of the play statistically. The Duncan screen was arguably the most important part of the play and Duncan gets zilch, Parkers penetration and collapsing of the defense only gets him an assist in the end etc.

If were going to microanalyze possessions like this, were still a long ways away from properly weighing the importance of each player's contribution to the successful possession.
I think the article acknowledged as much, noting that they're understaffed, undertrained, and that most of this is still in its infancy.

I don't think of being able to analyze each moment of the game as taking away the ambiguities, I think those still exist since this model looks at probabilities. I mean they broke down the play in the article, and yet here we are still debating the merits of the play.

It just makes for a smarter conversation, more like a discussion like how a poker hand was played, and whether these guys increased probabilities or set up the other team throughout the course of the game, and which plays those were, etc.

To oversimplify it, this is like going from the 8-color crayon box, and working your way up through the 32, the 64, and higher detail color schemes. The discussion gets opened up more, but at the same time is controlled and more productive because you now have less-disputable tools/currency in your debate.

D-Leethal
02-07-2014, 01:15 PM
I think the article acknowledged as much, noting that they're understaffed, undertrained, and that most of this is still in its infancy.

I don't think of being able to analyze each moment of the game as taking away the ambiguities, I think those still exist since this model looks at probabilities. I mean they broke down the play in the article, and yet here we are still debating the merits of the play.

It just makes for a smarter conversation, more like a discussion like how a poker hand was played, and whether these guys increased probabilities or set up the other team throughout the course of the game, and which plays those were, etc.

To oversimplify it, this is like going from the 8-color crayon box, and working your way up through the 32, the 64, and higher detail color schemes. The discussion gets opened up more, but at the same time is controlled and more productive because you now have less-disputable tools/currency in your debate.

"Quantify" or "Explain" was probably a better word than eliminate. I just think its an exercise in futility and impossible to explain the ambiguities of the game. For example, what happens if Parker's pass to Kawai Leonard was off and not directly on the numbers? What is Kawai has to lean down, disrupt his rhythm to catch before shooting the ball? I would imagine that has a major impact on whether or not he hits the shot and thus, whether or not it was the right play to make.

If Parker takes it all the way, what if he is forced to make an extra dribble that disrupts his timing and rhythm? Most attacking players have a certain rhythm once they turn the corner and beat their defender, a certain number or strides and dribbles, what if he is thrown off base, forced to spin, studder or crossover? How do you quantify that? It surely can't be the same expected result as a perfect rhythm stride to the hoop can it?

D-Leethal
02-07-2014, 01:20 PM
I don't think all jump shots are equal even in the same area of the court with the same level of defense on you. All shots at the rim aren't equal. But these stats assume they are. Those are examples of the ambiguities I was referring to for the most part. How do you quantify that? Is it even worth microanalyzing it THAT deep?

Some guys like the passes low, some like the passes high, some guys like to catch the ball at their right shoulder, some like it at their left shoulder, guys are going to respond differently to the passes they prefer and the different pass should have a different expected outcome.

I guess we will need to wait and see what the 64 crayon case looks like before knowing if we will ever get that far.

KniCks4LiFe
02-07-2014, 01:52 PM
Depends on the players. That pass is to Leonard and Kwhahi is stroking it at 36% from that spot. You trust that decision by Tony b/c it's a team designed play.

But if it's Patty Mills and the recipient is say Nando, um..take the lay up. :laugh2:

Shammyguy3
02-07-2014, 03:42 PM
I think this is a fantastic debate. One thing that I think has to be taken into consideration is that the Spurs are playing the Cavs. If you extend the game into overtime, there's 5 more minutes to be played. Who would pick in a 5 minute game is what it comes down to. I would lean towards Parker taking the layup, because with the game tied at let's say 8 seconds i'm confident that the Spurs won't let the Cavs score, sending the game into overtime where I trust the Spurs far more than Cleveland.

D-Leethal
02-07-2014, 03:49 PM
To actually get on topic, I don't think you should ever give up guaranteed tie game for a chance at the win. You take the guarantee (or as close to a guarantee as you can get barring a missed layup). I think the OT point made above is a good one and say, your best player is fouled out, momentum has been against you all quarter, etc., sometimes you just know your not going to win in OT. Thats more of a gut feeling than anything and sometimes you need to try and hit the dagger and avoid tie game/OT.

I surf, and my dad always told me "you don't leave great waves to go search for better waves because you might not surf any waves at all". You don't give up a great outcome for a chance at a better outcome 99% of the time in my book because you might strike out completely in search of that better outcome.

KnicksorBust
02-07-2014, 04:20 PM
We are early in the voting but I am surprised people would rather take a lower percentage shot than drive for the layup.

Joshtd1
02-07-2014, 04:33 PM
On a play like that, you take the easiest shot available. If Waiters doesn't come down, Tony easily takes that layup. However Waiters did come down, and Tony felt the easier play would be to an open Kawhi for an open 3. Pretty simple IMO. I think that's an instinct play. Why force up something tougher, when you can get someone else an easier shot on a designed play.

kdspurman
02-07-2014, 04:59 PM
The Spurs are a little different than other teams. Pop simply hates going into overtime. All it does is force guys to play longer when he wants to play them as little as possible as it is. So in the regular season, Pop is willing to take risks. Plus, it's a confidence builder for Kawhi. Not to mention when you look @ what he was shooting from the corner from around that time of the year, it was pretty impressive.

http://i59.tinypic.com/nponpj.jpg

He trusts TP to make the right call. If Leonard misses, I'm sure Pop is happy the play was executed properly and they got the shot they wanted.

IndyRealist
02-07-2014, 05:27 PM
I don't think all jump shots are equal even in the same area of the court with the same level of defense on you. All shots at the rim aren't equal. But these stats assume they are. Those are examples of the ambiguities I was referring to for the most part. How do you quantify that? Is it even worth microanalyzing it THAT deep?

Some guys like the passes low, some like the passes high, some guys like to catch the ball at their right shoulder, some like it at their left shoulder, guys are going to respond differently to the passes they prefer and the different pass should have a different expected outcome.

I guess we will need to wait and see what the 64 crayon case looks like before knowing if we will ever get that far.
Actually, you are assuming that it sees all shots at the rim as equal. As you can see with Kawhi's shot, the rating changes based on range, proximity of the defender, whether the defense is actually in position or not, etc. It's confusing that you think you see the potential flaws in a system, but know with absolute certainty that professional economists who have advanced degrees and do this for a living would fail to consider it.

Chronz
02-07-2014, 06:58 PM
Great step in the right direction, allows team to lok for players who make high% plays, whether they go in or not. Shows when players are taking the best shots on the court, even if they are chucker.s

Pacerlive
02-07-2014, 07:17 PM
One thing I thought of is that EPV probably changes late in games when refs may swallow their whistle on a drive to the lane from a pg.

KnicksorBust
02-09-2014, 10:29 AM
The Spurs are a little different than other teams. Pop simply hates going into overtime. All it does is force guys to play longer when he wants to play them as little as possible as it is. So in the regular season, Pop is willing to take risks. Plus, it's a confidence builder for Kawhi. Not to mention when you look @ what he was shooting from the corner from around that time of the year, it was pretty impressive.

http://i59.tinypic.com/nponpj.jpg

He trusts TP to make the right call. If Leonard misses, I'm sure Pop is happy the play was executed properly and they got the shot they wanted.

You dont actually think that was designed for Kawhi do you?

He is probably 3rd option at best.

Jamiecballer
02-09-2014, 11:50 AM
i don't see anything automatic about that layup :shrug:

seems like he went with his instincts and they were good on that one.

Raps18-19 Champ
02-09-2014, 11:56 AM
I didn't think Parker had a clear shot for the layup unless he went around.

kdspurman
02-09-2014, 01:28 PM
You dont actually think that was designed for Kawhi do you?

He is probably 3rd option at best.

It's sort of like a read option. Parker has a couple different options once the play was executed. I have seen the Spurs out of time outs or towards the end of games get a wide open corner 3 many times. So I absolutely think Kawhi was part of the play. They love the corner 3, and like I mentioned Pop is not a fan of OT games. So if Parker saw the defense was sucked in enough, he was going to pass to Kawhi the whole time.

They've been running that kind of play for years now. Some aren't the same exact scenario as this one, but pretty close. All end of game situations. There's some other ones that I can't seem to locate, but it's definitely something Pop believes in towards the end of games.

http://www.youtube.com/watch?v=G4NbhRb862k (skip to 1:10 or so) TP probably could've went @ Shaq or dropped it off to TD for a layup/dunk

http://www.youtube.com/watch?v=cIO8-61XmVU

http://hangtime.blogs.nba.com/2012/01/28/film-study-spurs-from-the-corner/

KnicksorBust
02-09-2014, 02:00 PM
If he extended with his off arm its an easy deuce. If he took an extra dribble he reverse is cake. Waiters wasnt even in help.

PatsSoxKnicks
02-09-2014, 02:18 PM
"Quantify" or "Explain" was probably a better word than eliminate. I just think its an exercise in futility and impossible to explain the ambiguities of the game. For example, what happens if Parker's pass to Kawai Leonard was off and not directly on the numbers? What is Kawai has to lean down, disrupt his rhythm to catch before shooting the ball? I would imagine that has a major impact on whether or not he hits the shot and thus, whether or not it was the right play to make.

If Parker takes it all the way, what if he is forced to make an extra dribble that disrupts his timing and rhythm? Most attacking players have a certain rhythm once they turn the corner and beat their defender, a certain number or strides and dribbles, what if he is thrown off base, forced to spin, studder or crossover? How do you quantify that? It surely can't be the same expected result as a perfect rhythm stride to the hoop can it?

Um, you can quantify all of that. You're just not privy to the amount of detail these teams get in their stats. And you can't imagine a team getting that amount of data, but they can.

KnicksorBust
03-24-2014, 11:24 AM
i don't see anything automatic about that layup :shrug:

seems like he went with his instincts and they were good on that one.


I didn't think Parker had a clear shot for the layup unless he went around.

Pause at 0:07 seconds. He's a step ahead and can easily extend to finish the play.

torocan
03-24-2014, 11:43 AM
As a general rule...

Open lay up > open 3 > open mid range > contested lay up > contested 3 > contested jump shot

This is of course subject to a TON of variables. The skill of the person contesting, the skill of the person shooting, whether they're good at drawing fouls or not, their risk of turning the ball over, whether you have a big man in position to rebound, whether you're at a rebounding advantage or disadvantage, etc, etc.

Open shots are nearly *always* better than contested shots, simply because shot percentages dive through the floor as soon as someone has a hand up within 5 feet (20% drop in efficiency).

Also, the time on the clock has an impact. 3 pointers are higher volatility shots (high reward, but higher risk - ie, streaky), so shooting a ton of 3's during the early/middle of the game is less dangerous than in the closing minutes.

That said if you're taking the occasional 3 pointer instead of a lay up to the rim, you're not going to go TOO far wrong in terms of efficiency, as long as your team shoots decently from range and can finish at the rim. As long as you're not falling in love with the 15-20 foot jump shot (guys like Dirk and Melo excluded of course).

Finally, per the OP, I would have preferred he takes the 2 if he's got a good run at the rim. Even if he misses, there's the chance for a short rebound or a foul that puts him on the line to tie.

Ebbs
03-24-2014, 11:54 AM
I read this a month ago it's interesting. And there is value to it.

Guppyfighter
03-24-2014, 11:43 PM
Turning every moment of a basketball game into a series of numbers is such a stupid exercise.

And the franchises who care about winning will be doing exactly that.