Win Shares is the most worthless stat ever

[iElusive]

Seriously..... I always laugh when people bring up win shares in an argument especially since 99% of the time the person doesn't even know what it means. It is just some bogus way for a person to defend the player they are arguing for in a comparison.

Here is why it is worthless....

-Detlef Schrempf, Dana Barros, Reggie Miller, and Stockton were all ahead of Hakeem in win shares in 1995. The year he won his second ring and was at worst...the second best player in the NBA(in reality...the best). Hakeem wasnt even top 10. Not sure how far he fell but it amuses me to know numbers can be bent that far.

-In 1990 Hakeem put up 24/14/5 blocks 2 steals and 3 assists a game. But his teams record was bad(well...average) so he was behind Terry Porter, Reggie, and so on. And odd thing is Reggies team had almost the exact same record as Hakeems.

-Brent Barry ended up having more win shares than Kidd in 2002. And that was in 02 when midseason Kidd likely would have been voted MVP.

-Oh and also that year in 2002...Elton Brand over Dirk, Shaq(3rd ring year) and KG somehow.

Some of the results mock themselves just by existing.

[SideshowBob]

I think Win Shares is a decent stat when used in proper context and not relied upon as an end-all-be-all stat, kinda like Hollinger does with PER.

As you said, it's better to use it if you know what it's telling you, rather than throwing it out there and saying X has more win shares than Y, therefore X>Y

[iElusive]

I think Win Shares is a decent stat when used in proper context and not relied upon as an end-all-be-all stat, kinda like Hollinger does with PER.

As you said, it's better to use it if you know what it's telling you, rather than throwing it out there and saying X has more win shares than Y, therefore X>Y

Here is the problem.....99% of the people don't understand win shares and don't know how it works. I don't either and I think it's pointless altogether. I don't know how you take win shares into context either. It's just random jibberish thrown into a digital.

[Dr Aki]

OP's right.

its all about the losses with HCA

[Atmanne]

Kevin Love having more win shares than the Wolves had wins at certain points last year got a good chuckle out of me.

I agree, it's a grossly overused stat.

[Wavy Q]

JordansBulls disagrees

[iElusive]

I would rather know how much Gatorade a player drank per game than what his win shares were for a season.

[Dupp]

Kobe cant drink as much gatorade as he use to, he is done.

[Doormatt]

WIn Shares is more useful than PER imo, and obviously every stat is dependent on context. there is no stat or argument that should be made or used without context when analyzing players.

also, i dont think its fair to throw something out the window because you dont agree with it. unless you can find a serious flaw, then i think all stats should be considered. even something as turrible as Wins Produced can help by giving you a different perspective. again, its important to know what a stat is telling you and why it is giving you that "answer," and as long as you know why and how, then every stat can be useful to a degree.

[iElusive]

WIn Shares is more useful than PER imo, and obviously every stat is dependent on context. there is no stat or argument that should be made or used without context when analyzing players.

also, i dont think its fair to throw something out the window because you dont agree with it. unless you can find a serious flaw, then i think all stats should be considered. even something as turrible as Wins Produced can help by giving you a different perspective. again, its important to know what a stat is telling you and why it is giving you that "answer," and as long as you know why and how, then every stat can be useful to a degree.

No it's not more useful than PER, go ahead and explain what win shares is and how it is used. I don't even get the formula or calculation. It just seems like a bunch of drivel complied of numbers.

[bhartsell]

Yeah but what's the context and when is the right time to use it, etc? I think PER is the best because it doesn't make any pretense about what it is and isn't. Win Shares tries to include defense and these things that really aren't quantifiable. PER doesn't try to tell you anything about defense and I like that. Just weight box score stats into one compact number. You can add in your subjective value of defense to PER if you want to. Win Shares confines itself to no real room for interpretation.

[Mahoney]

WIn Shares is more useful than PER imo, and obviously every stat is dependent on context. there is no stat or argument that should be made or used without context when analyzing players.

also, i dont think its fair to throw something out the window because you dont agree with it. unless you can find a serious flaw, then i think all stats should be considered. even something as turrible as Wins Produced can help by giving you a different perspective. again, its important to know what a stat is telling you and why it is giving you that "answer," and as long as you know why and how, then every stat can be useful to a degree.

No it's not more useful than PER, go ahead and explain what win shares is and how it is used. I don't even get the formula or calculation. It just seems like a bunch of drivel complied of numbers.

And because you can't grasp it means the measure is useless and hogwash?

[DreDay]

All stats are worthless IMO. Even if you put them all together, you still won't know the true impact of a player, you will just have a general idea that has a good chance to be wrong. I'll probably get blasted for saying that though.

[iElusive]

WIn Shares is more useful than PER imo, and obviously every stat is dependent on context. there is no stat or argument that should be made or used without context when analyzing players.

also, i dont think its fair to throw something out the window because you dont agree with it. unless you can find a serious flaw, then i think all stats should be considered. even something as turrible as Wins Produced can help by giving you a different perspective. again, its important to know what a stat is telling you and why it is giving you that "answer," and as long as you know why and how, then every stat can be useful to a degree.

No it's not more useful than PER, go ahead and explain what win shares is and how it is used. I don't even get the formula or calculation. It just seems like a bunch of drivel complied of numbers.

And because you can't grasp it means the measure is useless and hogwash?

I'm letting you guys explain what it means, or is it because none of you people who actually use it know what it means?

[coldgrip1]

At this point, is a terrible misleading stat IMO.

[aurareturn1]

Offensive rating and defensive rating are almost as worthless.

The moment you see scrubs/bad offensive players/bad defensive players in the top 10, they are worthless.

[Sabzi]

All advanced statistics or sabermetrics are quite accurate if you know the proper context to put them in. The problem with the general masses is too much emphasis is placed on scoring.

[RolloTommasi]

All advanced statistics or sabermetrics are quite accurate if you know the proper context to put them in. The problem with the general masses is too much emphasis is placed on scoring.

No the problem is that Basket is not Baseball

[abark]

From what I understand win shares is similar to per, in that it's 1 number that represents a players overall value. The main difference is that it meant to correlate with the amount of wins a team would hypothetically get by having that player. I don't know why, but the number is actually 3 or 4 times higher than the team wins the number is supposedly representing (This might explain the Kevin Love thing).

Also, you cannot have a negative WS, so it's not an accurate average. If you create a false floor, then you are not penalizing players enough for consistently playing below that point. I'm not sure how often a player scores below the 0 mark, but it's hard to put credibility in a stat that cannot differentiate between having a very off night and being straight terrible.

I'm no expert on these stats, so if anything I said about WS is wrong, let me know.

[DJDanji]

From what I understand win shares is similar to per, in that it's 1 number that represents a players overall value. The main difference is that it meant to correlate with the amount of wins a team would hypothetically get by having that player. I don't know why, but the number is actually 3 or 4 times higher than the team wins the number is supposedly representing (This might explain the Kevin Love thing).

Also, you cannot have a negative WS, so it's not an accurate average. If you create a false floor, then you are not penalizing players enough for consistently playing below that point. I'm not sure how often a player scores below the 0 mark, but it's hard to put credibility in a stat that cannot differentiate between having a very off night and being straight terrible.

I'm no expert on these stats, so if anything I said about WS is wrong, let me know.

If we are talkin WS as tracked on basketball-reference, there is the possibility of getting negative WS, and the amount of WS correlate with the amount of wins the Team will propably get. Plus it directly leads to the teams overall record (not multiplied by 3 or smth.).
http://www.basketball-reference.com/about/ws.html -> pls visit if u want to know how it is calculated, cant really explain it better then that.

As for usefulness of the stat:
you should be aware, that WS displays the amount of Wins a single player is responsible for. With that said, it should be obvious that a team with a better expected record has more wins to share between its players. Same goes for teams with only a few good players: take Lebron in his Cleveland years -> this team had a pretty good record for how bad most of these players were => leading to a high amount of WS for their by far best player LBJ.
WS more or less show how much each player of a team is responsible for the amount of wins this specific team will propably get (the amount of failure if u add up all WS of one team is pretty low, it's like a difference of ~3 wins per season if i remember correctly).
I wouldnt use it to find the leagues best player, but more to compare the players of a certain team regarding their effect on the overall record.

@OP
With that said, it's not too surprising that for example Detlef Schrempf had a higher WS than Hakeem, as his team had a significantly better record.