meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Bonn1997 wrote:Fivethirtyeight's method was off by an average of 6 wins last year, for example. So the stat gives you an idea of future performance but clearly there's a lot of uncertainty going into any off-season projections. I'd be willing to guess most of our fans have been off by more than 6 games most years, though. We're usually off by almost double that. 538 is just one source. I'm still looking up others. I see that using win shares, SI predicted us to win 28 games last year, which wasn't that far off.

Last year the Knicks actual talent was better than the 32 wins they ended up with. Minus injuries and with better coaching they easily win more than 32 games. And that team wasn't very good. This team has far more potential than 28 or 38 wins. This I say despite what the Win Share indications from last season might suggest.

A lot of the Knicks failures were made worse by the front office or coaching. I think this time the FO and coach are in sync. I think that makes a huge difference and it's getting overlooked.

Bonn1997 wrote:

nixluva wrote:

Bonn1997 wrote:Fivethirtyeight's method was off by an average of 6 wins last year, for example. So the stat gives you an idea of future performance but clearly there's a lot of uncertainty going into any off-season projections. I'd be willing to guess most of our fans have been off by more than 6 games most years, though. We're usually off by almost double that. 538 is just one source. I'm still looking up others. I see that using win shares, SI predicted us to win 28 games last year, which wasn't that far off.

Last year the Knicks actual talent was better than the 32 wins they ended up with. Minus injuries and with better coaching they easily win more than 32 games. And that team wasn't very good. This team has far more potential than 28 or 38 wins. This I say despite what the Win Share indications from last season might suggest.

A lot of the Knicks failures were made worse by the front office or coaching. I think this time the FO and coach are in sync. I think that makes a huge difference and it's getting overlooked.

The story you're describing does not seem likely. You're saying the team won 32 games, and coincidentally that's right where the metrics thought they'd be. The metrics and the results both just happened to underestimate the team to the same degree. You're saying an optimistic fan can actually more accurately assess the team's skill than the metrics can and can tell the team's talent better the win-loss record at the end of the season can. I have a tough time buying that.

nixluva wrote:

Bonn1997 wrote:

Bonn1997 wrote:

nixluva wrote:

Bonn1997 wrote:Fivethirtyeight's method was off by an average of 6 wins last year, for example. So the stat gives you an idea of future performance but clearly there's a lot of uncertainty going into any off-season projections. I'd be willing to guess most of our fans have been off by more than 6 games most years, though. We're usually off by almost double that. 538 is just one source. I'm still looking up others. I see that using win shares, SI predicted us to win 28 games last year, which wasn't that far off.

Last year the Knicks actual talent was better than the 32 wins they ended up with. Minus injuries and with better coaching they easily win more than 32 games. And that team wasn't very good. This team has far more potential than 28 or 38 wins. This I say despite what the Win Share indications from last season might suggest.

A lot of the Knicks failures were made worse by the front office or coaching. I think this time the FO and coach are in sync. I think that makes a huge difference and it's getting overlooked.

The story you're describing does not seem likely. You're saying the team won 32 games, and coincidentally that's right where the metrics thought they'd be. The metrics and the results both just happened to underestimate the team to the same degree. You're saying an optimistic fan can actually more accurately assess the team's skill than the metrics can and can tell the team's talent better the win-loss record at the end of the season can. I have a tough time buying that.

although your use of the word potential may be what's throwing me off. The metrics are predicting the most likely result, not potential. Almost half the teams end up winning more and half less than the predicted amount since only a few win the exact number predicted. So of course they have the potential to be better or worse than predicted.

I have stated for years that my predictions are based on what I think the potential of the team is and not the most likely outcome statistically. I'm saying that at their best this is what I believe this team is capable of.

thats fair but then you would agree that the outcome is not likely to match the full potential in most cases and therefore can't question much lower estimations.

Bonn1997 wrote:

meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.

Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26

Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.

Cartman718 wrote:

Bonn1997 wrote:

meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.

Bleacher Report 2015-16
= within 5 wins/losses of prediction
= more than 5 wins/losses of prediction

Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26

Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.

meloshouldgo wrote:

Cartman718 wrote:

Bonn1997 wrote:

meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.

Bleacher Report 2015-16
= within 5 wins/losses of prediction
= more than 5 wins/losses of prediction

Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26

Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.

Interesting. Go right ahead and predict the win loss record for every team this year and let's see you beat whatever this year's win shares based predictions is on accuracy. It's s simple enough position to put to test.

The rest of us can take bets on how good or bad you are. This can be a fun season long activity.

that would be fun!

Outgoing Players: Arron Afflalo, Lou Amundson, Cleanthony Early, Langston Galloway, Kevin Seraphin, Derrick Williams

Incoming Players: Willy Hernangomez, Justin Holiday, Brandon Jennings, Mindaugas Kuzminskas, Courtney Lee, Maurice Ndour, Joakim Noah, Marshall Plumlee, Derrick Rose

Re-Signed Players: Lance Thomas, Sasha Vujacic

Projected Starting Five: Derrick Rose (PG), Courtney Lee (SG), Carmelo Anthony (SF), Kristaps Porzingis (PF), Joakim Noah (C)

"They're high," Derrick Rose told NBA.com's Lang Whitaker when asked about the New York Knicks' expectations for 2016-17. "I mean, with these teams right now, they're saying us and Golden State are the superteams, and they're trying not to build that many superteams, and Adam Silver came out with the statement and this and that. And the expectations I think of us, we just want to win."

Not to be too harsh... but that's delusional. No one—well, no one outside of brainwashed fans bragging on Twitter—has called the Knicks a superteam. Nor should they. They're still not even a lock to make the playoffs.

Did New York add talent to the roster? Absolutely. Rose, Joakim Noah, Brandon Jennings, Courtney Lee, Justin Holiday and Marshall Plumlee lend some level of legitimacy, even if none currently plays at a star level.

But this is an outfit composed of mismatched pieces and players who have struggled in recent years. Rose hasn't been the same since he started earning the injury-prone label, though he did look better down the stretch of 2015-16. Noah is coming off the worst year of his career. Jennings can't hit 40 percent of his shots from the field.

The Knicks have more star power on the roster, and they're earning some excitement in the Big Apple. The depth chart looks like it's filled with legitimate NBA players.

But they're not contenders. Not even in the Eastern Conference.

Grade: B-

Anyone who says the Knicks roster is full of mismatched pieces doesn't really know very much about basketball. The reason to hedge on the Knicks is injury concerns. Their starting five makes perfect basketball sense.

meloshouldgo wrote:

Cartman718 wrote:

Bonn1997 wrote:

meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.

Bleacher Report 2015-16
= within 5 wins/losses of prediction
= more than 5 wins/losses of prediction

Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26

Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.

Interesting. Go right ahead and predict the win loss record for every team this year and let's see you beat whatever this year's win shares based predictions is on accuracy. It's s simple enough position to put to test.

The rest of us can take bets on how good or bad you are. This can be a fun season long activity.

Um...I am not the one claiming that predictions of win/loss records before the season even begins is a worthwhile exercise, my whole post was to refute exactly that. The so called "experts" in NBA predictions that you and Bonn and maybe mreinman are touting as almost gospel in this thread... they don't know $h|t.... So perhaps you and Bonn and mreinman should go ahead and make the predictions? I can bet you that your predictions will have the same level of accuracy as the folks you are agreeing with. In other words, they don't know any better or worse than you before the season begins.

Why should I waste my time with that...you guys think it's a worthwhile exercise...so YOU do it.

meloshouldgo wrote:

Cartman718 wrote:

meloshouldgo wrote:

Cartman718 wrote:

Bonn1997 wrote:

meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.

Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.

Bleacher Report 2015-16
= within 5 wins/losses of prediction
= more than 5 wins/losses of prediction

Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26

Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.

Interesting. Go right ahead and predict the win loss record for every team this year and let's see you beat whatever this year's win shares based predictions is on accuracy. It's s simple enough position to put to test.

The rest of us can take bets on how good or bad you are. This can be a fun season long activity.

Um...I am not the one claiming that predictions of win/loss records before the season even begins is a worthwhile exercise, my whole post was to refute exactly that. The so called "experts" in NBA predictions that you and Bonn and maybe mreinman are touting as almost gospel in this thread... they don't know $h|t.... So perhaps you and Bonn and mreinman should go ahead and make the predictions? I can bet you that your predictions will have the same level of accuracy as the folks you are agreeing with. In other words, they don't know any better or worse than you before the season begins.

Why should I waste my time with that...you guys think it's a worthwhile exercise...so YOU do it.

No one said anything about predictions being gospel. Instead I called you out to back up your very real statement that you can make predctiona matching statistical models in your sleep. Appears you are not man enough to accept the challenge so you need to hide behind this level of double talk. SHOCKING.

I am confused...are you agreeing with Bonn that these predictions are a worthwhile exercise or not? That's where we started. Answer this.
As far as being man enough to accept the challenge, I don't accept challenges that would be a waste of my time. No double talk.