How Great is the 2017 draft class?

[Hoop Head]

If Bledsoe is on the block you'd have to think his value is at least a late lotto. I'm a Bledsoe fan, but with the crop of pgs in the upcoming draft and us being at least a year or two away from competing in the west I'm all in on trading him if value is found. Knight, Chandler, and Tucker need to be shipped out as well.

McD got a lottery pick out of Markieff last year, surely Bledsoe could net more than that. I don't think trading him is the answer, he'd probably have more value during the draft or the offseason. Trading him to a team that is in the lottery now could be problematic, he might push a team to a better record or into the playoffs if they're on the cusp of them now.

I doubt any teams in the lottery would be interested in Chandler but I read that Portland was interested in him, he might be able to sure up their defense enough to push them back into the 8th seed. They need a rim protector and they're looking for someone to help immediately rather than a younger player. Chandler for their 1st is probably the best deal Phoenix could get.

 

[AzStevenCal]

If we're picking clear cut talent sure but in the middle of the draft it's something else. I've had enough combo guards for a while.

Not me. Saying "no combo guards" in today's NBA is tantamount to saying I don't want to win. Every player on the perimeter has to be a threat to shoot and if you're handling the ball you have to be a real threat to score from distance and be able to penetrate too.

I WANT a combo guard running this team. Which means, I also want him to be able to generate offense for other players. But if he isn't a threat to shoot and score, it's almost impossible for him to be fully effective as a playmaker.

Maybe the league will once again change the rules that protect the perimeter player but until then, playing a 90's style PG puts your team at a disadvantage. And since the NBA loves the fact that these rules have mostly negated the advantages created by today's zone defense rules, I doubt it's ever changing back.

 

[elindholm]

Wouldn't you just need to establish the standard dev from the mean of their shooting? Those with higher std dev's would be streakier.

You mean FG% on a game-by-game basis? You'd have to have a way to normalize the variance in attempts, because otherwise you'd get a conflating effect with players whose attempts also vary a lot. If you have a player who sometimes has 20 FGA/game and sometimes has 5, the games with 5 are more likely to produce percentages that look like outliers (Wow, 4 of 5, he shot 80%!).

As an example, even a very bad shooter like Brandon Knight winds up with relatively few games in which he's below 30%, because one way or another he figures out a way to jack up enough attempts, and it's pretty hard to go 4/15 if you have any shooting ability at all. (To put it another way, once he's started off 1/8, he keeps shooting.) And yet most people describe Knight as streaky.

 

[DWKB]

You mean FG% on a game-by-game basis? You'd have to have a way to normalize the variance in attempts, because otherwise you'd get a conflating effect with players whose attempts also vary a lot. If you have a player who sometimes has 20 FGA/game and sometimes has 5, the games with 5 are more likely to produce percentages that look like outliers (Wow, 4 of 5, he shot 80%!).

As an example, even a very bad shooter like Brandon Knight winds up with relatively few games in which he's below 30%, because one way or another he figures out a way to jack up enough attempts, and it's pretty hard to go 4/15 if you have any shooting ability at all. (To put it another way, once he's started off 1/8, he keeps shooting.) And yet most people describe Knight as streaky.

No, I meant 3P% instead of FG%. One, I think "shooter" is defined better by a 3P% as FG% can be high without having a jump shot perse and has much more variance in how you attain your attempts and where they occur on the court as opposed to 3P%.

Take Gordon and Harden this year to date:

Both have 353 3PA, Gordon shooting a better 3P% than Harden .411 to .354, however, Gordon's STDDEV is higher that Hardens, so while Gordon is shooting at a higher 3P%, Harden is the "more consistent" shooter.

Just a quick example of what I'm getting at.

 

[Ouchie-Z-Clown]

Not me. Saying "no combo guards" in today's NBA is tantamount to saying I don't want to win. Every player on the perimeter has to be a threat to shoot and if you're handling the ball you have to be a real threat to score from distance and be able to penetrate too.

I WANT a combo guard running this team. Which means, I also want him to be able to generate offense for other players. But if he isn't a threat to shoot and score, it's almost impossible for him to be fully effective as a playmaker.

Maybe the league will once again change the rules that protect the perimeter player but until then, playing a 90's style PG puts your team at a disadvantage. And since the NBA loves the fact that these rules have mostly negated the advantages created by today's zone defense rules, I doubt it's ever changing back.

"90s style pg????"

You mean like:

KJ
Derek Harper
John Stockton
Tim hardaway
Penny hardaway
Mark price
Magic Johnson
Gary Payton
Allen iverson
Stephon Marbury

And that's just odd top of my head without thinking hard. You think having these lead guards would put your team at a disadvantage today?!?

 

[elindholm]

No, I meant 3P% instead of FG%. One, I think "shooter" is defined better by a 3P% as FG% can be high without having a jump shot perse and has much more variance in how you attain your attempts and where they occur on the court as opposed to 3P%.

Take Gordon and Harden this year to date:

Both have 353 3PA, Gordon shooting a better 3P% than Harden .411 to .354, however, Gordon's STDDEV is higher that Hardens, so while Gordon is shooting at a higher 3P%, Harden is the "more consistent" shooter.

Just a quick example of what I'm getting at.

How are you measuring standard deviation?

 

[DWKB]

How are you measuring standard deviation?

You mean what is my data set or what is my equation?

I'm using individual games 3P% for each player.

 

[AzStevenCal]

"90s style pg????"

You mean like:

KJ
Derek Harper
John Stockton
Tim hardaway
Penny hardaway
Mark price
Magic Johnson
Gary Payton
Allen iverson
Stephon Marbury

And that's just odd top of my head without thinking hard. You think having these lead guards would put your team at a disadvantage today?!?

I was thinking a little later than much of that group but no, Marbury and Iverson would be much more dangerous in today's NBA. But they were knocked for being shooting guards pretending to be point guards. Today, they'd be hailed instead of over analyzed. But Jason Kidd (early in his career) and Rajon Rondo for example would have had less value under modern rules/interpretations.

Most of the players you mentioned played prior to the adoption of the league's current zone defense rules and the development of zone defenses eventually led to the rules changes that now favor the scoring guard over the pure passer.

 

[elindholm]

I'm using individual games 3P% for each player.

Right, okay, so that's what I was getting at. If an 0-3 game gets weighted the same as a 6-16 one, you're getting a very distorted picture. (Think of what an 0-1 game from the line would do to a 90% free-throw shooter.) Gordon has an 0-3 game and an 0-4 game, whereas Harden has at least 5 attempts in every game.

Also, if Gordon's overall make rate is higher, his standard deviation will be higher as a raw number. How much higher is his standard deviation as a percentage of his average make rate? If he's .411 +/- .120 and Harden is .354 +/- .105, then they're actually about the same.

I'm not saying it's impossible to establish streakiness through data analysis, only that it's more complicated than it appears.

 

[DWKB]

Right, okay, so that's what I was getting at. If an 0-3 game gets weighted the same as a 6-16 one, you're getting a very distorted picture. (Think of what an 0-1 game from the line would do to a 90% free-throw shooter.) Gordon has an 0-3 game and an 0-4 game, whereas Harden has at least 5 attempts in every game.

Also, if Gordon's overall make rate is higher, his standard deviation will be higher as a raw number. How much higher is his standard deviation as a percentage of his average make rate? If he's .411 +/- .120 and Harden is .354 +/- .105, then they're actually about the same.

I'm not saying it's impossible to establish streakiness through data analysis, only that it's more complicated than it appears.

Gordon was .190 and Harden was .150. Not sure why the std dev would expect to be higher for a higher mean though.

I was just saying this is a start and would get you close enough with 3PT shooter comparisons in consistency. Much better than saying "so and so shoots at a x % overall so he's consistent". No, that's not true.

 

[hcsilla]

Watch tape of kevin Martin, the shot is similar although he starts it lower than martin did

So you are telling me there is a chance.

xc_hide_links_from_guests_guests_error_hide_media

Yes, but among every 500 NBA players there is one Kevin Martin. Ball might become the next, but I wouldn't hold my breath.

 

[elindholm]

Gordon was .190 and Harden was .150. Not sure why the std dev would expect to be higher for a higher mean though.

Because the range is compressed for the lower-percentage shooter. You can't get lower than zero in an individual game, so someone who shoots less well has less room in the low end of the curve to broaden the distribution. Extreme example, if you have someone who shoots only .150 from three-point range, it will be almost impossible for him to have a standard deviation greater than Harden's .150, because the lowest he can record in a given game is zero. Similarly, you'd expect a 90% FT shooter to have a smaller scatter than an 80% one.

I was just saying this is a start and would get you close enough with 3PT shooter comparisons in consistency. Much better than saying "so and so shoots at a x % overall so he's consistent". No, that's not true.

Another way would be to go shot-by-shot and assess whether making the last shot has any impact on the likelihood of making the next one. Last I read, those studies had come up empty, but there are confounding factors with those too. Some players might indulge in low-percentage heat-check shots if they feel like they've earned them, while others might have teammates who work harder to get them really good looks if they're on a roll.

 

[DWKB]

Because the range is compressed for the lower-percentage shooter. You can't get lower than zero in an individual game, so someone who shoots less well has less room in the low end of the curve to broaden the distribution. Extreme example, if you have someone who shoots only .150 from three-point range, it will be almost impossible for him to have a standard deviation greater than Harden's .150, because the lowest he can record in a given game is zero. Similarly, you'd expect a 90% FT shooter to have a smaller scatter than an 80% one.

I don't think this is correct. The range for both of them is 0% to 100% on any given night. Regardless of overall avg, the "streakiest" of shooters would only have nights of 0-fer or perfect behind the arc and have large std devs. Higher overall avg would mean more 100% than 0% and lower avg would mean the opposite.

Another way would be to go shot-by-shot and assess whether making the last shot has any impact on the likelihood of making the next one. Last I read, those studies had come up empty, but there are confounding factors with those too. Some players might indulge in low-percentage heat-check shots if they feel like they've earned them, while others might have teammates who work harder to get them really good looks if they're on a roll.

Yeah, I think you can get too close with this idea. That's why I took a game by game approach instead of a shot by shot approach.

 

[Russ Smith]

Because the range is compressed for the lower-percentage shooter. You can't get lower than zero in an individual game, so someone who shoots less well has less room in the low end of the curve to broaden the distribution. Extreme example, if you have someone who shoots only .150 from three-point range, it will be almost impossible for him to have a standard deviation greater than Harden's .150, because the lowest he can record in a given game is zero. Similarly, you'd expect a 90% FT shooter to have a smaller scatter than an 80% one.



Another way would be to go shot-by-shot and assess whether making the last shot has any impact on the likelihood of making the next one. Last I read, those studies had come up empty, but there are confounding factors with those too. Some players might indulge in low-percentage heat-check shots if they feel like they've earned them, while others might have teammates who work harder to get them really good looks if they're on a roll.

And there's defense involved too.

Lots of times I think hot shooter really means finding the mismatch and going after it which NBA teams do really well. If they find a mismatch they will go to it every time until you either change the matchup or it stops working.

I was watching USC and Cal the other night and Cal had their 2 7 footers taking turns at C and guarding USC's Chimezie Metu who is a knockdown midrange shooter. At one point he hit 5 or 6 in a row and was trash talking Ivan Rabb who he was guarding at the other end. When Cal went smaller and suddenly Rabb was on Metu, he stopped scoring. It wasn't that he wasn't hot anymore, it was that unlike the 2 7 footers he's quick enough to guard Metu out there without getting blown by so he contested the shots they were giving him. It was one of main keys to Cal winning the game they changed the matchup and Metu stopped scoring.

 

[elindholm]

I don't think this is correct. The range for both of them is 0% to 100% on any given night. Regardless of overall avg, the "streakiest" of shooters would only have nights of 0-fer or perfect behind the arc and have large std devs. Higher overall avg would mean more 100% than 0% and lower avg would mean the opposite.

I don't have the formulas to back it up, but I can appeal to intuition. A 0% shooter would have no variance. Neither would a 100% shooter. A 1% or 99% shooter would have very little variance. So either (a) the expected variance depends on the overall average, with a peak at 50%, or (b) in spite of what happens at either tail of the distribution, there exists some middle area in which the expected variance is the same. It's hard to imagine the kind of math that would be necessary for (b) to be true.

 

[Errntknght]

Wouldn't you just need to establish the standard dev from the mean of their shooting? Those with higher std dev's would be streakier.

I think almost everyone is streaky from beyond the arc. Maybe a few guys who shoot them only when they're completely uncontested are fairly consistent.

To get some idea of streakiness, I computed the std dev for Kyle Korver FG% and 3P% on a yearly basis, excluding his rookie season. FG% avg .448 std .022; 3P% avg .433 std .044. On a yearly basis the sample sizes are large and he shot very similar averages on 2's and 3's yet the std dev doubled. Also the 3's were counted in FG%, causing it to increase from 2's alone. I'd guess the std dev of 2's alone would be about a half that of FG% so the ratio for std dev for 3's relative to 2's would be 4:1.

If you computed the std dev using individual game percentages that would introduce so much noise that it may not be very meaningful.

Statisticians (more than one) have declared that there is no evidence of streakiness in shooting in the NBA but I've never read an account of how they arrived at that seemingly silly result - I mean, we all know a hot hand when we see it! I presume they looked at sequences of misses and makes by an individual from actual games and found that the probability of a make following a make is the same as probability of a make following a miss. It is true that if players have hot and cold streaks there should be a significant difference. If there really is no such thing as streakiness what it means is that that the normal outcome of random makes and misses appears streaky to us.

I doubt if they considered year long hot and cold streaks, which Kyle seems to have. Very likely they'd explain that the appearance of those is due to changes in the circumstances in which Kyle shot from year to year. Differences in the sets and plays, different players around him, opponents keying on him more or less,
et cetera. Three's have more variance because they are more delicate shots and more prone to disturbance by defenders. Something like that.

 

[Ouchie-Z-Clown]

I don't think this is correct. The range for both of them is 0% to 100% on any given night. Regardless of overall avg, the "streakiest" of shooters would only have nights of 0-fer or perfect behind the arc and have large std devs. Higher overall avg would mean more 100% than 0% and lower avg would mean the opposite.




Yeah, I think you can get too close with this idea. That's why I took a game by game approach instead of a shot by shot approach.

NERDS!!!!

 

[leclerc]

Hey if the guard can run an offence, get the ball moving, identify the hot hand, have a minimum of BBIQ (point guard skills) I'm okay that he also has the size to play/defend SG and not needing to control the ball to contribute out there. I want a point guard. Having two small shooting guards who have limited PG skills doesn't cut it. Perhaps today's game is all about isolation plays and barreling into the lane so the referee must make the call. That's not the game I fell in love with.

 

[Ronin]

Someone to keep an eye on.

Jeff Eisenberg– ‏@JeffEisenberg
The list of point guards I'd rather have this season than Notre Dame's Matt Farrell is very short. Such a good decision maker.

You must be registered for see images attach

 

[DWKB]

Someone to keep an eye on.

Jeff Eisenberg– ‏@JeffEisenberg
The list of point guards I'd rather have this season than Notre Dame's Matt Farrell is very short. Such a good decision maker.

You must be registered for see images attach

If you're going to draft an upperclassman PG then why not draft Frank Mason?

 

[GatorAZ]

If you're going to draft an upperclassman PG then why not draft Frank Mason?

Mason is definitely a sleeper at PG. He wasn't highly rated out of HS was he?

 

[devilalum]

Suns draft steategy is simple. BPA!!! The Suns need 2 pick the guy with the most star potential period. They have young prospects at every level but no obvious stars.

from phone

 

[Mainstreet]

Suns draft steategy is simple. BPA!!! The Suns need 2 pick the guy with the most star potential period. They have young prospects at every level but no obvious stars.

I'm with you about BPA. However, I don't watch college basketball enough to know who that might be right now. I wouldn't be adverse to drafting another big man if he has excellent skills.

PGs seem easier to come by unless your looking at the Suns PGs.

I'm guessing the Suns will be drafting around #6.

 

[overseascardfan]

If PHX could acquire 3 first rounders for Bledsoe/Knight/Chandler I would hope for 3 of any combination of these 6 players:
Markele Fultz / Lonzo Ball
Jason Tatum / Josh Jackson
Lauri Markkenan / Johnathan Motley

 

[AzStevenCal]

If PHX could acquire 3 first rounders for Bledsoe/Knight/Chandler I would hope for 3 of any combination of these 6 players:
Markele Fultz / Lonzo Ball
Jason Tatum / Josh Jackson
Lauri Markkenan / Johnathan Motley

I'm not familiar with Motley but yeah, I'd love to get multiple players from that list. I'd add Tucker to that list of assets that might net us a pick this season. There are also some player that will likely fall to the second round that hold my interest such as a couple of players in Oregon. Anyway, this is the season to own lots of picks.