Why do advanced statistics hate kobe?

Moderator: Doctor MJ

User avatar
dockingsched
Retired Mod
Retired Mod
Posts: 51,716
And1: 10,687
Joined: Aug 02, 2005
     

Why do advanced statistics hate kobe? 

Post#1 » by dockingsched » Thu Nov 12, 2009 8:59 pm

Part One

Part Two


nothing groundbreaking, but a pretty easy to read take on why kobe, widely regarded as one of the best players in the nba, doesn't stack up to wade/lebron/paul/howard when it comes to PER.

thoughts?
"We must try not to sink beneath our anguish, Harry, but battle on." - Dumbledore
User avatar
rrravenred
Retired Mod
Retired Mod
Posts: 5,901
And1: 416
Joined: Feb 24, 2006
Location: Pulling at the loose threads of arguments since 2006

Re: Why do advanced statistics hate kobe? 

Post#2 » by rrravenred » Fri Nov 13, 2009 12:44 am

Interesting. He does basically say that defence takes up a lot of the slack (and/or is misrepresented in the improved PER of others), but also says that this isn't a big enough factor to change the overall outcome significantly.

Not, of course, that I necessarily favour PER as the One, True advanced metric. :)
ElGee wrote:You, my friend, have shoved those words into my mouth, which is OK because I'm hungry.


Got fallacy?
User avatar
dockingsched
Retired Mod
Retired Mod
Posts: 51,716
And1: 10,687
Joined: Aug 02, 2005
     

Re: Why do advanced statistics hate kobe? 

Post#3 » by dockingsched » Fri Nov 13, 2009 1:28 am

both of the conclusions he comes to are pretty well known, that kobe isn't in the help defending role as much as wade/etc, and kobe plays in the triangle. he uses the yr kobe played under rudy t to show how much kobe's improved assist numbers might have improved his PER...

the one thing that immediately came to mind, and lemme know if its off base, but how much is jordan's PER negatively affected by that? kinda brings more light into player's having "jordan like" PER seasons.
"We must try not to sink beneath our anguish, Harry, but battle on." - Dumbledore
User avatar
rrravenred
Retired Mod
Retired Mod
Posts: 5,901
And1: 416
Joined: Feb 24, 2006
Location: Pulling at the loose threads of arguments since 2006

Re: Why do advanced statistics hate kobe? 

Post#4 » by rrravenred » Fri Nov 13, 2009 1:49 am

Well...

Jordan's PER:

88-89 (under Collins): 31.1
89-90 (1st year under Phil) 31.2
90-91 (2nd year under Phil) 31.6

Pretty static. Yeah, this is a standardised stat and covers the ages 25-27 (the period by which which most wing players approach their peak), but unless you think Jordan markedly improved his game in a way which was suppressed by the triangle, I wouldn't say there's anything conclusive there.
ElGee wrote:You, my friend, have shoved those words into my mouth, which is OK because I'm hungry.


Got fallacy?
User avatar
CellarDoor
Retired Mod
Retired Mod
Posts: 11,141
And1: 927
Joined: May 11, 2008
         

Re: Why do advanced statistics hate kobe? 

Post#5 » by CellarDoor » Fri Nov 13, 2009 5:25 am

First of all, he simply doesn't HAVE to do what they do anymore, so he doesn't.
Second of all, he's got a team full of great passers, so he's not going to get gaudy assist numbers, he's also got three guys that are better rebounders than anyone on the Cavs or Heat, so they're siphoning numbers there.
Third, he's a better man defender than help defender. While he's certainly no slouch there either, the other three are getting a lot of support there.


In short, look at the 2nd scoring options:
David West: he's getting a LARGE amount of his points on assisted field goals this year.
Mo Williams: he's the benefactor of a lot of the doubles thrown at Lebron, so the FGs are assisted, and he's not as consistent a 2nd option as West
Beasley? Whoever it is, they're simply not very good.
Gasol: best of the bunch both at scoring and as a playmaker himself.

Kobe's still on that level, but the team around him is on another level than the other superstars.
tsherkin wrote:You can run away if you like, but I'm not done with this nonsense, I'm going rip apart everything you've said so everyone else here knows that you're completely lacking in basic basketball knowledge...
mysticbb
Banned User
Posts: 8,205
And1: 707
Joined: May 28, 2007
Contact:
   

Re: Why do advanced statistics hate kobe? 

Post#6 » by mysticbb » Fri Nov 13, 2009 2:35 pm

Matter of fact is that not only PER is "hating" Bryant, but every other advanced metric too. If we take PER, Win Shares and Net +/- for the last 7 seasons, Bryant was able to be a Top3 player in those categories only 3 times. He never lead the league in any of those advanced stats. I looked that up and gave the Leader 3 points, the 2nd 2 points and the 3rd 1 point and added those numbers up.

Garnett, Nowitzki: 20
James: 19
Duncan: 13
.
.
.
Bryant: 3

Well, that shows that it is not only PER (well, in fact he was twice 3rd in PER during those years). Overall Bryant is 3rd in Win Shares in those 7 years, behind Nowitzki and Garnett. Add James and you will find three players who had an higher Win Shares per Game value than Bryant.
Bryant is 5th in PER behind Garnett, James, Duncan and Nowitzki for all players with at least 13000 minutes playing time (around half of the possible minutes during that timespan).
Bryant is also behind Garnett, Duncan, Nowitzki and James in adjusted +/- in the last 10 years.

This also shows that is not only related to PER. Statistically speaking Bryant was never the best player in the league nor is a good candidate for the best player for the decade. It seems like he is just overrated by casual fans (and maybe even by GMs due to his ability to sell tickets and jerseys, something a GM also thinks about, I guess).
User avatar
TyCobb
Forum Mod - Lakers
Forum Mod - Lakers
Posts: 37,166
And1: 8,715
Joined: Apr 17, 2005
Location: Pitcher's Mound
     

Re: Why do advanced statistics hate kobe? 

Post#7 » by TyCobb » Fri Nov 13, 2009 8:09 pm

mysticbb wrote:This also shows that is not only related to PER. Statistically speaking Bryant was never the best player in the league nor is a good candidate for the best player for the decade. It seems like he is just overrated by casual fans (and maybe even by GMs due to his ability to sell tickets and jerseys, something a GM also thinks about, I guess).


His stats make him underrated. He is capable of doing so much more, but he chooses not to because he's a winner first.
Read more, learn more, change your posts.
User avatar
rrravenred
Retired Mod
Retired Mod
Posts: 5,901
And1: 416
Joined: Feb 24, 2006
Location: Pulling at the loose threads of arguments since 2006

Re: Why do advanced statistics hate kobe? 

Post#8 » by rrravenred » Fri Nov 13, 2009 11:35 pm

What, like Tim "Half-the-season-mailing-it-in-and-I'm-killing-it-in-the-playoffs" Duncan?

Can't see it myself, although since we're talking about a perception issue "he could do more", then we'll have to disagree on his capacity to positively affect the game.
ElGee wrote:You, my friend, have shoved those words into my mouth, which is OK because I'm hungry.


Got fallacy?
Miller4ever
Retired Mod
Retired Mod
Posts: 8,597
And1: 283
Joined: Jun 24, 2005
Location: Location: Location:

Re: Why do advanced statistics hate kobe? 

Post#9 » by Miller4ever » Sat Nov 14, 2009 6:12 am

Anybody know the PER of the players on the USA Redeem Team over the course of the Olympics? Would be interesting to see how everyone's numbers are affected by playing with superstars.
mysticbb
Banned User
Posts: 8,205
And1: 707
Joined: May 28, 2007
Contact:
   

Re: Why do advanced statistics hate kobe? 

Post#10 » by mysticbb » Sat Nov 14, 2009 8:55 am

TyCobb wrote:His stats make him underrated. He is capable of doing so much more, but he chooses not to because he's a winner first.


Seriously, NO! Why in hell would Bryant want to have a worse +/- number? This would even contradict your claim that he wants to win first. ;) Bryant would need to know how those advanced metrics are working to effect those number negatively. But you are a perfect example of a fan who wants to believe in something.
mysticbb
Banned User
Posts: 8,205
And1: 707
Joined: May 28, 2007
Contact:
   

Re: Why do advanced statistics hate kobe? 

Post#11 » by mysticbb » Sat Nov 14, 2009 9:03 am

Miller4ever wrote:Anybody know the PER of the players on the USA Redeem Team over the course of the Olympics? Would be interesting to see how everyone's numbers are affected by playing with superstars.


I don't have the PER, but I applied my own model to th raw data, which is in a way comparable to PER. I use 10 as the average instead of 15. It is also minutes- and pace-adjusted. I call it just PRA (player rating).

Code: Select all

Dwyane Wade       23.88
LeBron James      20.02
Chris Bosh        19.89
Dwight Howard     18.59
Chris Paul        16.21
Carmelo Anthony   14.86
Kobe Bryant       13.14
Carlos Boozer     13.02
Tayshaun Prince   11.64
Deron Williams    10.92
Jason Kidd        10.04
Michael Redd       8.3
giberish
RealGM
Posts: 11,511
And1: 2,443
Joined: Mar 30, 2006
Location: Whereever you go - there you are

Re: Why do advanced statistics hate kobe? 

Post#12 » by giberish » Mon Nov 16, 2009 3:52 am

It's not so much that advanced stats hate Kobe, but that he's generally been top 5-7 good but not the top 1-3 level player that he's hyped to be.

Part of it is playing in LA (more media exposure). Part is generally having better teammates and therefore winning more (Kobe doesn't win 66 games with the supporting cast LeBron had last year).

Part is that he looks prettier than most other top players (he's got a better jumper than LeBron but people don't see how much better LeBron is at getting easy baskets, Wade and Paul aren't as smooth, KG and Duncan are methodical big men, Dirk is often awkward looking), so people make the 'just look at him - he's clearly the best' arguments.
azuresou1
Head Coach
Posts: 7,157
And1: 737
Joined: Jun 15, 2009
   

Re: Why do advanced statistics hate kobe? 

Post#13 » by azuresou1 » Mon Nov 16, 2009 5:51 am

This is a ridiculous thread.

When every player in the NBA for the past few years has been saying that Kobe is hands down the best player in the league, with only a few people beginning to change their votes to LeBron or Wade or CP3, I'm inclined to believe them over some cobbled-together metrics that all have flaws.
User avatar
John Doe [MIN]
Bench Warmer
Posts: 1,281
And1: 9
Joined: Jul 15, 2009

Re: Why do advanced statistics hate kobe? 

Post#14 » by John Doe [MIN] » Mon Nov 16, 2009 11:13 am

giberish wrote:It's not so much that advanced stats hate Kobe, but that he's generally been top 5-7 good but not the top 1-3 level player that he's hyped to be.

This.

I have to say, I don't like this thread title. It seems to imply that all the advanced statistics met in an alley somewhere and decided to blackball Kobe. That we the fans are the unbiased ones. I don't think the stats "hate" Kobe. They consistently rate him as between the 4th and 10th best player in the league, which is fair. A better question would be "Why do idiot fans still think he's the best?".
azuresou1
Head Coach
Posts: 7,157
And1: 737
Joined: Jun 15, 2009
   

Re: Why do advanced statistics hate kobe? 

Post#15 » by azuresou1 » Mon Nov 16, 2009 2:15 pm

John Doe [MN] wrote:
giberish wrote:It's not so much that advanced stats hate Kobe, but that he's generally been top 5-7 good but not the top 1-3 level player that he's hyped to be.

This.

I have to say, I don't like this thread title. It seems to imply that all the advanced statistics met in an alley somewhere and decided to blackball Kobe. That we the fans are the unbiased ones. I don't think the stats "hate" Kobe. They consistently rate him as between the 4th and 10th best player in the league, which is fair. A better question would be "Why do idiot fans still think he's the best?".


'Still' think he's the best, or that he was the best 2-5 years ago? Because many people in this topic seem to think that Kobe was never the best, and that's a far, far, FAR cry from the truth.

A better question would be, "Do heavy statnerds even watch the games?" Not everyone who looks at stats is Daryl Morey.
User avatar
hoopdata
Freshman
Posts: 96
And1: 0
Joined: Sep 20, 2009
Location: Long Island, NY
Contact:

Re: Why do advanced statistics hate kobe? 

Post#16 » by hoopdata » Mon Nov 16, 2009 3:40 pm

I've come up with a way for PER to actually account for assisted and unassisted field goals as opposed to using estimates based on team percentages. Link on the explanation is here:

http://www.hoopdata.com/recent.aspx?aid=39

The new APER can be found here for the past four seasons:

http://hoopdata.com/advancedstats.aspx

For what it's worth, after accounting for the fact that Kobe is assisted on only 39% of his field goals or so the past four seasons compared to the league average of 55% or so, he jumps from rank 6 to 5 in 09, 8 to 4 in 08, and 5 to 3 in 07. I think this makes PER a lot more accurate and fixes one of its biggest flaws in not assigning proper credit for shot creation. No longer does the formula assume Kyle Korver and Jason Kapono create as many shots as Dwyane Wade and Lebron James.
http://www.hoopdata.com - NBA Statistics and Analysis - Has a sortable player/team stat database that includes offensive shot locations, defensive shot locations, assisted field goals, assist locations, and1s, charges, times blocked, and much more.
mysticbb
Banned User
Posts: 8,205
And1: 707
Joined: May 28, 2007
Contact:
   

Re: Why do advanced statistics hate kobe? 

Post#17 » by mysticbb » Mon Nov 16, 2009 10:52 pm

hoopdata wrote:I think this makes PER a lot more accurate and fixes one of its biggest flaws in not assigning proper credit for shot creation. No longer does the formula assume Kyle Korver and Jason Kapono create as many shots as Dwyane Wade and Lebron James.


Well, overall what you are saying is, that someone who uses a pick to get open and makes the basket while having the ball in his hand is more valuable to his team (PER wants to say something about winning, keep that in mind) than someone who is using a pick&roll or pick&pop situation. In the end in both cases the help for the open shot is coming from a teammate, but only in the latter scene the teammate will get an assist for this. PER already overrates inefficient scorers to a certain degree, you made it even worse, which I doubt was your basic intention.

You can use the Ast% as an indicator for the ability to create shots, but that is in no way an accurate value for that (leave alone that especially for assists the subjective view of the scorekeeper can make a huge difference).
User avatar
rrravenred
Retired Mod
Retired Mod
Posts: 5,901
And1: 416
Joined: Feb 24, 2006
Location: Pulling at the loose threads of arguments since 2006

Re: Why do advanced statistics hate kobe? 

Post#18 » by rrravenred » Mon Nov 16, 2009 10:58 pm

azuresou1 wrote:A better question would be, "Do heavy statnerds even watch the games?" Not everyone who looks at stats is Daryl Morey.


Have you noticed the title of this forum? It proves quite descriptive of what actually gets discussed inside. Have you anything to add with regards to that?
ElGee wrote:You, my friend, have shoved those words into my mouth, which is OK because I'm hungry.


Got fallacy?
mysticbb
Banned User
Posts: 8,205
And1: 707
Joined: May 28, 2007
Contact:
   

Re: Why do advanced statistics hate kobe? 

Post#19 » by mysticbb » Mon Nov 16, 2009 11:00 pm

azuresou1 wrote:A better question would be, "Do heavy statnerds even watch the games?"


Of course not, "statnerds" hate basketball at all. ;)

Anyway, in the last regular season 59466 minutes of basketball were played, if you were able to watch every second of your favourite team, you watched around 3964 minutes, that means you basically missed over 93% of the time. Well, think about that ...

rrravenred wrote:Have you noticed the title of this forum? It proves quite descriptive of what actually gets discussed inside. Have you anything to add with regards to that?


The thread title has the word "Kobe" in, I guess that has to count for something.
azuresou1
Head Coach
Posts: 7,157
And1: 737
Joined: Jun 15, 2009
   

Re: Why do advanced statistics hate kobe? 

Post#20 » by azuresou1 » Tue Nov 17, 2009 5:09 pm

There's a significant difference between interpreting stats used in conjunction with game footage, and using several stats, all of which are flawed and tell incomplete pictures, to say that Kobe was only a top 7 player a few years ago. Every single advanced stat is flawed, other than perhaps TOV%, which is why teams like the Rockets still have scouts instead of just crunching numbers and then coming to bad conclusions, such as that Kobe Bryant being only the 7th best player in 05-06, or not taking Kevin Durant for free on your team because he has an atrocious +/-.

Return to Statistical Analysis