in Shump's defense..

[GYK]

I saw this reading how Melo should go back to the four. The article was simply showing how advanced stats wise Melo is the 3rd best player in the league at the four spot.

Anyway I see that Shump's defense is a huge difference maker for our team.

https://mobile.twitter.com/HerringWSJ/s ... 80/photo/1

It reads the 9.3 dropoff without Shump is the equivalent difference between the Bulls defense and 6ers.

[JSmooth93]

Scary.

But advanced stats like per 36 and per 100 poss. don't hold water in my opinion.

Basketball is a volatile and unpredictable. It's silly to extrapolate what a player does in a set amount of minutes and say that you can apply it to 36 minutes or 100 possessions because every single second on the court is completely unique due to variables such as player fatigue, focus, awareness, positioning, and luck.

Without these stats I'm sure many of us are well aware of Shump's value on defense even in his off year last season.

[GYK]

I totally agree. I wasn't really trying to defend Shump. Title was just wordplay.

[suicidedeuce]

But advanced stats like per 36 and per 100 poss. don't hold water in my opinion.

They aren't advanced stats. They are simplied stats. They breakdown what's impossible to quantify by site over 48 mins a night/82 nights a season.

Basketball is a volatile and unpredictable. It's silly to extrapolate what a player does in a set amount of minutes and say that you can apply it to 36 minutes or 100 possessions because every single second on the court is completely unique due to variables such as player fatigue, focus, awareness, positioning, and luck.

Every single event or small sample size of events is volatile and unpredictable. This is true of EVERYTHING, from sports to science.

That's what's statistics are. Its the recording of volatile and unpredictable events over the course of long periods of time to find if there is any commonality and predictability within the volatility and seemingly randomness.

Those stats will tell you NOTHING for one game or a small group of games. But over whole months, and whole seasons, the volatility and randomness gets rounded, and legitimate patterns emerge.

Just like point differential is one of the most reliable ways to judge performance, the stats you're attempting to dismiss weed out their randomness over large sample sizes.

Close observation over time universally produces useful, relevant information. This is NO different for point differential, +/-, per 36, per 100, per 48.

1 or 2 games - practically useless.

20 game - starting to become useful.

82 games - pretty much told you exactly what happened and why.

[E-Balla]

Scary.

But advanced stats like per 36 and per 100 poss. don't hold water in my opinion.

Basketball is a volatile and unpredictable. It's silly to extrapolate what a player does in a set amount of minutes and say that you can apply it to 36 minutes or 100 possessions because every single second on the court is completely unique due to variables such as player fatigue, focus, awareness, positioning, and luck.

Without these stats I'm sure many of us are well aware of Shump's value on defense even in his off year last season.

That's not how those stats are used though. If someone scores 15 ppg in 18 minutes most will say 15 ppg isn't really high. 30 pp36 is very high and it shows that player is a high scorer despite averaging 18 minutes a night.

And many weren't aware of Shump's impact and many other's impact without the numbers. Did people realize when Mike Conley got good or was it about 2 years after he got good that people caught up with what the on/off was saying for years.

[MaseInYourFace]

I always find it funny that when the team plays really bad defense as a team you then start hearing all this talk about Shumpert being overrated defensively or just average. The long term eye test and the stats back up that he is a very good defender and is pretty close to being considered among the top perimeter defenders.

[PMFJB]

But advanced stats like per 36 and per 100 poss. don't hold water in my opinion.

They aren't advanced stats. They are simplied stats. They breakdown what's impossible to quantify by site over 48 mins a night/82 nights a season.

Basketball is a volatile and unpredictable. It's silly to extrapolate what a player does in a set amount of minutes and say that you can apply it to 36 minutes or 100 possessions because every single second on the court is completely unique due to variables such as player fatigue, focus, awareness, positioning, and luck.

Every single event or small sample size of events is volatile and unpredictable. This is true of EVERYTHING, from sports to science.

That's what's statistics are. Its the recording of volatile and unpredictable events over the course of long periods of time to find if there is any commonality and predictability within the volatility and seemingly randomness.

Those stats will tell you NOTHING for one game or a small group of games. But over whole months, and whole seasons, the volatility and randomness gets rounded, and legitimate patterns emerge.

Just like point differential is one of the most reliable ways to judge performance, the stats you're attempting to dismiss weed out their randomness over large sample sizes.

Close observation over time universally produces useful, relevant information. This is NO different for point differential, +/-, per 36, per 100, per 48.

1 or 2 games - practically useless.

20 game - starting to become useful.

82 games - pretty much told you exactly what happened and why.

Yes and no. This depends on factors not taken into account by these stats. For example, if a player only plays 20 mins a game and scores 10pts a game. Those silly per 36/per 48 stats say that you can extrapulate this data and make and assumption of a player given more minutes . So given twice the time he can possibly score twice the points or close to it (I am simplifying things computer so don't come at me with OMG THAT"S NOT HOW IT WORKS).

However these stats DO NOT take into account that if a player is only playing 20 mins a game he is usually doing it vs bench competition. Further, the opponents gameplan IS NOT keying in on him because he is a bench player. But give that same player starter minutes and he now faces tougher competition and defensive schemes that could or could not key in on said player which would null and void any extrapulation possible because now the variables have changed.

Which is why I find it silly that advanced stats try to quantify things in a vacuum without taking these things into account. Your turn computer.

[KnicksGod]

Bottom line: My hypothesis is that if you play Shump 30+ minutes and make sure either Cole or Dally is on the court at all times, the Knicks WOULD struggle to score BUT would eke out enough points to be at least a .500 team.

That is much better than losing by 10 points at home to even crappy teams because said crappy teams are in rhythm all -- night -- long.

Building a tough, disciplined, defensive-minded squad that moves the ball is all we need to accomplish this year. Trying to outscore people and overcome terrible, idiotic mistakes is no way to exist or to build anything.

After you've rigorously applied the system and have established team basketball, then maybe you can introduce A LITTLE crazy and low-IQ players like JR because then the system will already be in place. The system will be there to contain the player some. But if you try to build the offensive / defensive system with dumb players at the helm, it'll never take.

It'll be a wasted year.

So start small (though play big lineups). It's not about finding some magical lineup combination. It's about reducing the rotation to dedicated players with smarts rather than talented players without discipline. Use those guys to establish a real team and a defined system FIRST.

I believe Cole is just the kind of guy who can help build such a team. He should not be out of the rotation.

[suicidedeuce]

Yes and no. This depends on factors not taken into account by these stats. For example, if a player only plays 20 mins a game and scores 10pts a game. Those silly per 36/per 48 stats say that you can extrapulate this data and make and assumption of a player given more minutes . So given twice the time he can possibly score twice the points or close to it (I am simplifying things computer so don't come at me with OMG THAT"S NOT HOW IT WORKS).

No, it isn't the silly stats that say that, it's the silly people using them that either want to say that in a silly way or want to criticize them in a silly way.

Per stats is just a tool to look at production in different ways. You of course brings YOUR acknowledge to the equation when using the stats.

Per 48 was NOT created to show what a player would score under the silly premise he plays every minutes of every game. it just provides context to what happens while he's in the game.

Its all about context and critics want to dismiss it on context nobody in their right mind would ever give it.

However these stats DO NOT take into account that if a player is only playing 20 mins a game he is usually doing it vs bench competition.

This is a silly distinction. Starters play the majority of minutes, 60-66% percent, plus the idea that there are first and second units is mostly bunk. There is bleed over so starter are playing with bench players all the time.

[quote[ Further, the opponents gameplan IS NOT keying in on him because he is a bench player. But give that same player starter minutes and he now faces tougher competition and defensive schemes that could or could not key in on said player which would null and void any extrapulation possible because now the variables have changed.

Which is why I find it silly that advanced stats try to quantify things in a vacuum without taking these things into account. Your turn computer.[/quote]

You missed the point. Sample size means all.

if you record every minute every player of every game, all the anecdotal things you cite are rounded. If one considered what is being record gather than considered ways to dismiss what's being recorded, this would become clear.

You cannot dismiss simple written descriptions of THINGS. THAT. HAPPENED has irrelevent. Because they happened.

The kNicks were a better defensive team when Shumpert was in the game, period. Same way they were a LOT worse defensive team when Tim Hardaway Jr. was in the game. Period.

[blueNorange]

every stat shows that the team plays better with shumpert on the court

but let's ship him away for a 2nd rounder and give heavy minutes to thj who still doesn't know how to defend or rebound

[suicidedeuce]

I believe Cole is just the kind of guy who can help build such a team. He should not be out of the rotation.

I'm ready to dismiss the character or acumen of Derek Fisher all that quickly.

Derek Fisher has played against and I believe with (Thunder?) Aldrich.

I'm willing to bet he's watched every min of film on his mins last year like we all did.

He coached him in summer league and coached the games we all saw on TV.

He's coached him all training camp, an insight we do not have.

I'm not saying Derek Fisher is infallible, but he HAS more information on Aldrich than we all do.

Is it possible he's making a vast mistake in judgment, AND jackson is doing nothing privately to try to get him to reconsider? I acknowledge that's possible.

At the moment, however, I think it's more like Fisher sees things we don't, mainly because we don't have access to seeing them.

[PMFJB]

Yes and no. This depends on factors not taken into account by these stats. For example, if a player only plays 20 mins a game and scores 10pts a game. Those silly per 36/per 48 stats say that you can extrapulate this data and make and assumption of a player given more minutes . So given twice the time he can possibly score twice the points or close to it (I am simplifying things computer so don't come at me with OMG THAT"S NOT HOW IT WORKS).

Per stats is just a tool to look at production in different ways. You of course brings YOUR acknowledge to the equation when using the stats.

Per 48 was NOT created to show what a player would score under the silly premise he plays every minutes of every game. it just provides context to what happens while he's in the game.

Its all about context and critics want to dismiss it on context nobody in their right mind would ever give it.

However these stats DO NOT take into account that if a player is only playing 20 mins a game he is usually doing it vs bench competition.

This is a silly distinction. Starters play the majority of minutes, 60-66% percent, plus the idea that there are first and second units is mostly bunk. There is bleed over so starter are playing with bench players all the time.

[quote[ Further, the opponents gameplan IS NOT keying in on him because he is a bench player. But give that same player starter minutes and he now faces tougher competition and defensive schemes that could or could not key in on said player which would null and void any extrapulation possible because now the variables have changed.

Which is why I find it silly that advanced stats try to quantify things in a vacuum without taking these things into account. Your turn computer.

You missed the point. Sample size means all.

if you record every minute every player of every game, all the anecdotal things you cite are rounded. If one considered what is being record gather than considered ways to dismiss what's being recorded, this would become clear.

You cannot dismiss simple written descriptions of THINGS. THAT. HAPPENED has irrelevent. Because they happened.

The kNicks were a better defensive team when Shumpert was in the game, period. Same way they were a LOT worse defensive team when Tim Hardaway Jr. was in the game. Period.[/quote]


First off bold is not english so I skipped.

2nd Per minute stats try to quantify the quality of play of player per minute.

Here is a link with the all mighty hollinger explaining it. http://www.sportingcharts.com/dictionar ... stics.aspx

Basketball statistician John Hollinger has said of per-minute statistics, "It's a pretty simple concept, but one that has largely escaped most NBA front offices: the idea that what a player does on a per-minute basis is far more important than his per-game stats. The latter tend to be influenced more by playing time than by quality of play, yet remain the most common metric of player performance."

But at the same time it tries to extrapulate that break down of per minute into what would happen if said player played more. WHICH IT CAN'T. Variables are different. You can't make and assumption on 12 minutes of data and say it correlates to 48 minutes.

You can't break down someone on a per minute basis who plays 12 minutes under different variables than a guy who plays 30 under different circumstances.

Fatigue is another factor you haven't considered in your whole 1st squad bleeds into 2nd squad.
A player who starts and plays a full 8 mins in the first quarter is the same as a player who enters in the 8 minute mark? So fatigue doesn't play a role in anything?

[suicidedeuce]

2nd Per minute stats try to quantify the quality of play of player per minute.

Right.

Here is a link with the all mighty hollinger explaining it. http://www.sportingcharts.com/dictionar ... stics.aspx

Basketball statistician John Hollinger has said of per-minute statistics, "It's a pretty simple concept, but one that has largely escaped most NBA front offices: the idea that what a player does on a per-minute basis is far more important than his per-game stats. The latter tend to be influenced more by playing time than by quality of play, yet remain the most common metric of player performance."

Okay, since you're quoting what part of that says if you double the minutes you double the production? Can you find me any quote that suggests that's the intended use of the stat?

Hollinger is saying here per is a better way of measuring a players performance in terms of production than cumulative stats might. No one is going to be impressed by 7 pts and 4 rebounds. But if that is done consistently in 20 mins of play, then it simply illustrates that a player is likely having a positive effect on the game when he's in the game.

The whole "if you made him a starter the stats would extropolate exactly is a silly criticism invented by people highly motivated to dismiss them.

Its measuring contribution, rather than accumulation.

But at the same time it tries to extrapulate that break down of per minute into what would happen if said player played more. WHICH IT CAN'T.Variables are different. You can't make and assumption on 12 minutes of data and say it correlates to 48 minutes.

Who but you is trying?

You can't break down someone on a per minute basis who plays 12 minutes under different variables than a guy who plays 30 under different circumstances.

Sure you can, for the reason you can compare two guys who each play 12 mins.

Fatigue is another factor you haven't considered in your whole 1st squad bleeds into 2nd squad.
A player who starts and plays a full 8 mins in the first quarter is the same as a player who enters in the 8 minute mark? So fatigue doesn't play a role in anything?

How about flow and rhythm? How about the experience playing more minutes gives you?

I'm willing to have an objective conversation about this but you can't frame this issue like playing more mins is exclusively a detriment to per min production. That's be a ludicrous, ignorant assertion.

[blueNorange]

forget stats, it's the eye test

Knicks team defense is better with shumpert on the court

Knicks are terrible with amare on the court

smith is just awful

[KnicksGod]

I believe Cole is just the kind of guy who can help build such a team. He should not be out of the rotation.

I'm ready to dismiss the character or acumen of Derek Fisher all that quickly.

Derek Fisher has played against and I believe with (Thunder?) Aldrich.

I'm willing to bet he's watched every min of film on his mins last year like we all did.

He coached him in summer league and coached the games we all saw on TV.

He's coached him all training camp, an insight we do not have.

I'm not saying Derek Fisher is infallible, but he HAS more information on Aldrich than we all do.

Is it possible he's making a vast mistake in judgment, AND jackson is doing nothing privately to try to get him to reconsider? I acknowledge that's possible.

At the moment, however, I think it's more like Fisher sees things we don't, mainly because we don't have access to seeing them.

I'm willing to give him this benefit of the doubt for a month or two. If the team is just a new version of the same crap come the new year, then all bets are off. I pick up the pitchfork again.

Aldrich does a lot of things that this team needs and we already know his limitations. They are limitations we can live with.

Statistically, eye test, whatever the measurement tool -- we shouldn't play guys who allow the other team to get into rhythm. NBA teams score at will once they're comfortable. All of them, even the worst. I love THJ's potential and a young guy is always worth an effort to improve his weaknesses, but he too should sit if he's off the gameplan or not playing D.

I'm saying that if the team is losing 2/3rds of its games in a couple of months, there's no excuse to play guys on big contracts with big names in the name of "offense." You sit every single one of them, 100% of the time, and play guys who are on the plan. It'll create controversy but great leadership requires difficult decisions.

Fisher has a honeymoon but that doesn't mean we should bless all his choices either. Frankly he should do the above from Day 1. If it takes him a couple of months to figure it out, I won't be happy but obviously he's not going to be fired early. He should make this choice from Day 1 however, and it's a mistake not to.

If he thinks the team is worse with Aldrich defensively, which I find very hard to believe, that would be his only excuse for not playing him. If he thinks the team is worse offensively with Aldrich, and that's why he benches him, then he's wrong.

If Amar'e leads to easy scores from the other team, then he simply should not see the light of day.

[Deeeez Knicks]

I am sure being replaced by Tim Hardaway and even Amare had some effect on the stats.

Shump is definetly a positive for our defense and I fully support him starting, but +/- stats really do not tell the entire story.

[suicidedeuce]

I'm willing to give him this benefit of the doubt for a month or two. If the team is just a new version of the same crap come the new year, then all bets are off. I pick up the pitchfork again.

Am I allowed to pick up the pitchfork of my choosing like you, or do you get to decide what I can or can't pick up?

I'm saying that if the team is losing 2/3rds of its games in a couple of months, there's no excuse to play guys on big contracts with big names in the name of "offense." You sit every single one of them, 100% of the time, and play guys who are on the plan. It'll create controversy but great leadership requires difficult decisions.

I doubt you'll encounter much of an opposition to that viewpoint/

Fisher has a honeymoon but that doesn't mean we should bless all his choices either.

That's fine. But we should also empirically acknowledge he has significantly more insight than we all do.

That should be the easiest thing in the world to agree to. Yes, he sees a TON more than we do.

If he thinks the team is worse with Aldrich defensively, which I find very hard to believe, that would be his only excuse for not playing him. If he thinks the team is worse offensively with Aldrich, and that's why he benches him, then he's wrong.

I don't get why offense and defense is separated. The equation should be, who gives the the Knicks the best chance to outscore their opponents, be that on whatever end of the spectrum. If that's Aldrich, then okay.

[suicidedeuce]

I am sure being replaced by Tim Hardaway and even Amare had some effect on the stats.

Shump is definetly a positive for our defense and I fully support him starting, but +/- stats really do not tell the entire story.

I'll say this.

Something(s) DOES tell the whole story. You CAN measure and conclude why things happen in a basketball game.

The idea that there is some unrecordable thing that you can only know in your head and not make a record of is ridiculous.

Energy should be spent on finding better ways or combination of ways to demonstrate what happens and what you can do to improve the results, rather than on dismissing the idea it can be done.

Of course it can.

[sasso]

I've been saying this forever. Iman's defense helps us win. And in general, plugging in more defensive players into the lineup have helped us win and increased our +/-. People tend to dismiss average to below average offensive players that play great defense because the product doesn't look pretty on the court. But in reality, we win more games with defensive players on the court. It was just a preseason game, but I guarantee you if the lineup last night consisted of this we would have won:

Calderon / Prigioni
Shump / JR
Melo / Early
Acy / Smith
Dalembert / Aldrich

With smarter and better defensive players down low, the paint is better-protected, which allows the guards and wings to guard the 3-point line more aggressively so we don't get burned like last night.

[j4remi]

I'm not a fan of +/- for individuals, but it's interesting to see +/- for five man line-ups (Pretty sure I was using 82 games for these). Stats are really useful, but you need more than one to tell a story and you need the eye test to help out too. Over-reliance on box scores is just as bad as people who refuse to accept statistics they don't agree with.