Archived

This topic is now archived and is closed to further replies.

The Flaws of Defensive Win Shares


14 posts in this topic

Posted · Report post

I always suspected that Win Shares's extremely flawed (especially the defensive one), so i try to make some research and only got my suspicion stronger.On the same time, i got interested to make an article about basketball, so i made this one.

Please, ignore some weak writing and the mistakes.

Many NBA fans who’s really care about it’s game and the history are using Offensive and Defensive Win Shares to measured a specific player’s impact on his team.Via this method, they can see at how, for example, LeBron carries the Heat team in his shoulder in his last championship runs, or how great Bill Russell’s defense impact in Boston Celtics’s Dynasty as well.

Unfortunately, Defensive Win Shares, while showing several players in the place that they deserves, it is probably a system that’s not the best in term of showing on how great defensively a player, or how big his impact on the team as well.It placed players that was considered to be average defender, or even the poor one, on a very high place (and vice versa).Bob Cousy, for example, currently placed in the 39 position of the DWS, despite being considered by his contemporaries and himself to be a player that was not even cared for defense.People also wondering whether Dirk Nowitzky has larger impact than, let’s say, more ferocious defenders like Alonzo Mourning and Mark Eaton whose got placed below him.Meanwhile, some great defender like Bill Sharman didn’t even make to the top 250 players...

Did the system’s really flawed?This article’s dedicated to showing at how defensive win shares counting players on eras with counted defensive stats (post 1972-1973 season) or without them.

First, let’s see how this system counting defensive win shares prior to 1973-1974 season (where blocks, Defensive Rebounds and steals still not counted).Based on the process to counted defensive share provided by basketball-reference.com, here’s the All-Defensive teams in the 1955-1960 based on defensive shares.

1955-1960 (1st Team / 2nd Team)

C - Bill Russell (29.4) / Maurice Stokes (17.7)

F - Bob Pettit (18.0) / Red Kerr (15.6)

F - Dolph Schayes (19.5) / Tom Heinsohn (17.0)

G - Tom Gola (14.7) / Frank Ramsey (11.8)

G - Bob Cousy (23.3) / Bill Sharman (13.6)

Based on this alone, we can already see something wrong’s going on here.Bob Cousy, like i said before, was a poor defender by his own account (especially in 50’s).Yet he’s ended as the first defensive team G!Meanwhile, Bill Sharman got far behind Bob Cousy in the defensive shares, despite had reputation as an excellent defender in contrast to Cousy.Tom Heinshon being in the second team also disputable, as he considered himself to be a not so elite defender (although he also said that he and Russell were the one that provided defense for Celtics).

How about the next five years?Here’s the first/second defensive teams.

1960-1965 (1st Team / 2nd Team)

C - Bill Russell (65.9) / Wilt Chamberlain (33.5)

F - Tom Heinsohn (24.7) / Bob Pettit (21.6)

F - Tom Sanders (23.7) / Red Kerr (16.3)

G - Sam Jones (24.2) / Guy Rodgers (19.6)

G - K.C. Jones (25.4) / Bob Cousy (17.0)

While Bob Cousy’s not the best defensive guard in the league anymore, and began to cared for defense more, he’s still not by any mean the best in the league, yet he’s still included in the defensive team in his twilight years.What’s making it sounds even more ridiculous is that he’s already retired in the 1963, yet he’s still made more Defensive Win Shares than anyone not named K.C. Jones!Sam Jones and Guy Rodgers position are also disputed, as there are clear cut superior defenders like Jerry West, Hal Greer and Al Attles.

We can already concluded at how well the defensive win shares work in this era, but i’ll included another next 5 years defensive teams for more prove.

1965-1970 (1st Team / 2nd Team)

C - Bill Russell (38.3) / Wilt Chamberlain (32.5)

F - Dave Debusschere (19.7) / Bailey Howell (17.5)

F - John Havlicek (23.7) / Billy Cunningham (18.6)

G - Jeff Mullins (13.3) / Sam Jones (13.2)

G - Hal Greer (17.3) / Jerry Sloan (12.8)

Still no Jerry West instead of Sam Jones here, and Bailey Howell’s position is completely absurd, as he was considered to be a bad defender in his prime, and he would retired after 1971 season.

Based on these data and facts, we can easily say that DWS is a bad system to calculating at how great player was in defense in the era without counted blocks and steals, as well as not being the best in measured a player’s defensive impact on his team as it giving some bad defenders high win shares.

How about the era with counted block, steals and separated rebounds?Let’s see at how the defensive win share work compared to the stats provided by mysynergysports.com in the present NBA.

First, let’s see a comparison between player in the same team and same position, but with big minutes difference (in this case, LeBron to Battier) by DWS and mysynergysports’s stats.

Note:The Data is accurate as of 1-3-2013

LeBron’s DWS: 1,5

Stats:

MP PPP FGM FGA PPM

1151 0,86 85 239 0,207

%TO %Score

9% 35,6%

Battier’s DWS: 0,4

Stats:

MP PPP FGM FGA PPM

692 0,87 82 216 0,312

%TO %Score

5,8% 38,4%

Note:

MP- Minutes Played

PPP – Points Per Defensive Play.

PPM- Defensive Plays per minute

%TO – Percent Turnover of Defensive Plays.

%Score – Percent Score of Defensive Plays.

FGA can also be considered as defensive plays, as it also included defensive plays that made the offensive player committed turnover or got stolen away.

So, while LeBron looks better in both system, mysynergysports’s stats doesn’t make Battier got nearly as humilliated as the DWS showed.However, it should be noted that Battier’s Defensive Plays per minute suggested that he’s alot more focused in his defense than LeBron (although mysynergysports doesn’t give the stats to a helping defender for plays like isolation.But in isolation, both players looks similar, as LeBron’s involved in 39 plays to Battier 36, so it may neutralized the effect.)

Now, let’s use another comparison that has similar minutes on the same team but with different position (in this case, Dwight Howard and Kobe Bryant).

Dwight Howard’s DWS: 1,8

Stats:

MP PPP FGM FGA PPM

1116 0,63 43 141 0,126

%TO %Score

7,8% 31,7%

Kobe’s DWS:1

Stats:

MP PPP FGM FGA PPM

1202 0,75 97 296 0,246

%TO %Score

7,6% 33,6%

This one made both system contrast each other.Kobe with much lower DWS’s actually much more involved in defensive plays than Dwight Howard!The facts that both players are playing at similar MPG and Howard got critized for being soft this year also made any chances to redeem Howard got nullified.

So, while not as obvious as the pre- counted defensive stats, DWS is still very flawed to the root for current NBA with recorded and refined defensive stats.

For me, these facts already debunked the ultimate fact:Defensive Win Shares is a flawed system for measuring players impact defensively, and even more flawed for measuring players defensive capacity.Personally, I won’t even considered this system seriously anymore, but you?

Share this post


Link to post
Share on other sites

Posted · Report post

Every stat is flawed to some degree. All of them aren't taken in consideration when we're talking about the greats of all time in this league. That's where the eye test, footage, articles and so on comes in. If we look the game based on stats and stats only, Wilt'd be by far the GOAT basketball player ever and there would be no debate... at all.

Share this post


Link to post
Share on other sites

Posted · Report post

Very interesting post. Thanks for sharing.

Now...
Me and probably everyone who has used DWS, have noticed from time to time some crazy numbers for players that weren't good on defense. But on the other hand, most of the great defenders, have very high DWS numbers. I mean, for some players, DWS may picture their real defensive abilities in contrast with some others.

Something that I noticed reading that, is that the Celtics had too many players in those all-defense teams. That probably means that DWS are directly related with team defense.

 

I believe that the WS numbers overall though are much more accurate than the DWS.

Share this post


Link to post
Share on other sites

Posted · Report post

I want to say that not everyone who has high stats in some category is considered great at something. For example, Ice has a ton of amount of blocked shots but no one considers him an excellent defender.. actually far from that.

Share this post


Link to post
Share on other sites

Posted · Report post

Very interesting post. Thanks for sharing.

Now...

Me and probably everyone who has used DWS, have noticed from time to time some crazy numbers for players that weren't good on defense. But on the other hand, most of the great defenders, have very high DWS numbers. I mean, for some players, DWS may picture their real defensive abilities in contrast with some others.

Something that I noticed reading that, is that the Celtics had too many players in those all-defense teams. That probably means that DWS are directly related with team defense.

I believe that the WS numbers overall though are much more accurate than the DWS.

One thing that's also directly involved with WS is playing minutes.That's why Bob Cousy has very high Winning Share.But, before someone gave that crazy number for Cousy, he should've at least searching whether Bob Cousy is at least decent at best on defense, LOL.

Share this post


Link to post
Share on other sites

Posted · Report post

One thing that's also directly involved with WS is playing minutes.That's why Bob Cousy has very high Winning Share.But, before someone gave that crazy number for Cousy, he should've at least searching whether Bob Cousy is at least decent at best, LOL.

Bob Cousy is one of the best PG's of all time.

Share this post


Link to post
Share on other sites

Posted · Report post

Bob Cousy is one of the best PG's of all time.

And he is because of his offense, his impact on offense and being just one of those guys who revolutionized the league with their play, in his particular case, flashiness. As good as he was offensively, he was 10x times worse defensively. We penalize Nash for his D, but at least, he tries... Bob never did and he admitted it. That's why I don't understand people that put him in their all time lists so so high.

Share this post


Link to post
Share on other sites

Posted · Report post

And he is because of his offense, his impact on offense and being just one of those guys who revolutionized the league with their play, in his particular case, flashiness. As good as he was offensively, he was 10x times worse defensively. We penalize Nash for his D, but at least, he tries... Bob never did and he admitted it. That's why I don't understand people that put him in their all time lists so so high.

Well, he is a lot higher than "At least decent". The post made it seem like he wasn't a top 50 PG or something. He is top 10-15 without a doubt and if this is going off legacy and awards and what the player did to the game then top 5 is understandable

Share this post


Link to post
Share on other sites

Posted · Report post

Well, he is a lot higher than "At least decent". The post made it seem like he wasn't a top 50 PG or something. He is top 10-15 without a doubt and if this is going off legacy and awards and what the player did to the game then top 5 is understandable

I've seen people put Cousy as #10 on their all time lists... I never understood that. This guy played at one side of the court and all of his major success came after '56... after Russell came along. That what was my comment about.

ISOMELO likes this

Share this post


Link to post
Share on other sites

Posted · Report post

I've seen people put Cousy as #10 on their all time lists... I never understood that. This guy played at one side of the court and all of his major success came after '56... after Russell came along. That what was my comment about.

Yeah, he isn't a top 10 player of all time, but he is a top 10 PG of all time and that is what my original post was about. The fact that guy said he isn't even decent is just wrong lol

Share this post


Link to post
Share on other sites

Posted · Report post

Yeah, he isn't a top 10 player of all time, but he is a top 10 PG of all time and that is what my original post was about. The fact that guy said he isn't even decent is just wrong lol

I respected Bob Cousy.He's a legit top 5 Offensive Point Guard.But he's a very bad defender by many accounts including himself.What i'm talking about not even decent is his defense.

Sorry, for the misunderstanding.LOL.

Share this post


Link to post
Share on other sites

Posted · Report post

Good thread.

I'm not at all surprised by Cousy being ranked undeservingly by this worthless metric. Carlos Boozer led the Bulls in defensive win shares last season. That speaks volumes about the metrics credibility considering the fact his coach benches him in the fourth quarter of games because of his poor play defensively.

Share this post


Link to post
Share on other sites

Posted · Report post

Good thread.

I'm not at all surprised by Cousy being ranked undeservingly by this worthless metric. Carlos Boozer led the Bulls in defensive win shares last season. That speaks volumes about the metrics credibility considering the fact his coach benches him in the fourth quarter of games because of his poor play defensively.

Thanks for the visit and comment.You just add another reason of why this metric's suck.

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.