How legit is win shares?

[tpols]

Good luck. Having a conversation about advanced statistical analysis with a bunch of kids who are failing 9th grade Geometry is wasted effort.

LOL at people here who think "its too team oriented" and "its too complicated". The reason it was created was so that a players impact on winning could be measured. Its basis is on the team. If you don't care how the team did don't use the metric. Too complicated? Any good stat should be. PPG, etc are vastly overrated because guys who jack it up at 36% get rewarded for jacking it. You have to measure efficiency.

You're an idiot who cant recognize real basketball ability.. stats can be accumulated in a 100 different ways in basketball. There are a million guys in the league that have averaged 20ppg on decent efficiency that have never been the scorer of say.. Ray Allen. Andrew Bynum averages 15/10 on great percentages but Paul Millsap can average the same thing and notW be considered nearly as good because he cant dictate games like Bynum has at points.

There's more to basketball games than raw scoring/passing efficiency.. momentum, making the right plays, effort, off the ball positioning, defense.. they all cant be accurately measured. If you have a guy that can take over games at will and make amazing momentum swinging plays, that guy may be more important to your team winning than a big man who just scoops a lot of offensive rebounds and gets a lot of putbacks even if that big man has a higher FG.. or a guard that constantly gets a lot of his points when he's fed open jumpers or gets a lot of opportunities on the fast break.

I perfectly understand PER.. and TS.. and eFG. TS and eFG at least have concrete concepts behind them. They are VERY simple formulas. Winshares and PER? They are very complicated formulas that stretch for paragraphs that apply different SUBJECTIVE weights to the stats of their creator's preference. PER isn't that bad in my eyes though because even if you altered some of it's formulas, you'd still generally have the guy with the best/most efficient statline be on top. Of course efficiency is not the most important thing in basketball.. much more to the game when it comes to winning ball games than efficiency.

Anyone thats ever played even remotely competitive basketball knows the emotions and natural swings of momentum that occur often dictate where the game will end. Guys and teams that can control that momentum the best are the ones that win.. even if they dont post 40+ PERs.

[Glide2keva]

He doesn't understand any stat that he deems "unworthy." He chooses which ones he follows in a completely arbitrary manner (for instance, he credits "Blocks" but not "Shots altered" or "Passes tipped").

Make shit up much?

When did we have a conversation about passes tipped and shots altered? I'd love to know this because to my recollection, we've never had that conversation.

Secondly, shots altered and passes tipped are things that actually happen on the floor and can give you a feel for the nuances of the flow of the game, like FG, 3P, FT, blocks, steals, rebounds.

Third, I don't throw out stats because I don't understand them. I understand them just fine, that's why I say they are useless. Besides, you can't rely on stats too heavily in basketball, because each game has it's one ebb and flow. Baseball on the other hand has the same feel no matter where it's played.

You can break the whole game down in stats because it generates so many things that can happen, but tendencies always prevail. Basketball? Not so much.

And I used to be a stat junkie. When I worked at U.S Cellular Field (then Comiskey Park), the old Chicago Stadium, and the United Center, the first thing I would do is get stat sheets from the floor. So miss me with that.

[Glide2keva]

You're an idiot who cant recognize real basketball ability.. stats can be accumulated in a 100 different ways in basketball. There are a million guys in the league that have averaged 20ppg on decent efficiency that have never been the scorer of say.. Ray Allen. Andrew Bynum averages 15/10 on great percentages but Paul Millsap can average the same thing and notW be considered nearly as good because he cant dictate games like Bynum has at points.

There's more to basketball games than raw scoring/passing efficiency.. momentum, making the right plays, effort, off the ball positioning, defense.. they all cant be accurately measured. If you have a guy that can take over games at will and make amazing momentum swinging plays, that guy may be more important to your team winning than a big man who just scoops a lot of offensive rebounds and gets a lot of putbacks even if that big man has a higher FG.. or a guard that constantly gets a lot of his points when he's fed open jumpers or gets a lot of opportunities on the fast break.

I perfectly understand PER.. and TS.. and eFG. TS and eFG at least have concrete concepts behind them. They are VERY simple formulas. Winshares and PER? They are very complicated formulas that stretch for paragraphs that apply different SUBJECTIVE weights to the stats of their creator's preference. PER isn't that bad in my eyes though because even if you altered some of it's formulas, you'd still generally have the guy with the best/most efficient statline be on top. Of course efficiency is not the most important thing in basketball.. much more to the game when it comes to winning ball games than efficiency.

Anyone thats ever played even remotely competitive basketball knows the emotions and natural swings of momentum that occur often dictate where the game will end. Guys and teams that can control that momentum the best are the ones that win.. even if they dont post 40+ PERs.

Epic post is epic.

[HurricaneKid]

You're an idiot who cant recognize real basketball ability.. stats can be accumulated in a 100 different ways in basketball. There are a million guys in the league that have averaged 20ppg on decent efficiency that have never been the scorer of say.. Ray Allen. Andrew Bynum averages 15/10 on great percentages but Paul Millsap can average the same thing and notW be considered nearly as good because he cant dictate games like Bynum has at points.

There's more to basketball games than raw scoring/passing efficiency.. momentum, making the right plays, effort, off the ball positioning, defense.. they all cant be accurately measured. If you have a guy that can take over games at will and make amazing momentum swinging plays, that guy may be more important to your team winning than a big man who just scoops a lot of offensive rebounds and gets a lot of putbacks even if that big man has a higher FG.. or a guard that constantly gets a lot of his points when he's fed open jumpers or gets a lot of opportunities on the fast break.

I perfectly understand PER.. and TS.. and eFG. TS and eFG at least have concrete concepts behind them. They are VERY simple formulas. Winshares and PER? They are very complicated formulas that stretch for paragraphs that apply different SUBJECTIVE weights to the stats of their creator's preference. PER isn't that bad in my eyes though because even if you altered some of it's formulas, you'd still generally have the guy with the best/most efficient statline be on top. Of course efficiency is not the most important thing in basketball.. much more to the game when it comes to winning ball games than efficiency.

Anyone thats ever played even remotely competitive basketball knows the emotions and natural swings of momentum that occur often dictate where the game will end. Guys and teams that can control that momentum the best are the ones that win.. even if they dont post 40+ PERs.

You don't know me son. I've played at levels you couldn't reach on a ladder. I've been paid to coach players far better than yourself. You are certainly right that there are 100s of unmeasurables in basketball. I just don't trust you to know, understand, or value those subjects. I'll gladly spend an hour dissecting a set with you and why it was good/bad or how to make it better. But I'm certain any advanced conversation about basketball would fly over your head faster than a "complicated" formula.

The truth is known statistics are far more accurate than almost all analysis. They are far from foolproof, but they certainly outweigh fools like yourself.

[chips93]

They are just useless stats that people use as a crutch to make a player look like he played better than what he did. Then when you look at real stats like FT%, 3-point%, FG% they tell the truth.

but they dont.

last year ray allen shot better than daquan cook from 3pt, from the line, and from the field, but cook had a better TS%.

ray allen shot better from everywhere, but cook took more of the efficient shots (free throws and 3s).

[StateOfMind12]

You're an idiot who cant recognize real basketball ability.. stats can be accumulated in a 100 different ways in basketball. There are a million guys in the league that have averaged 20ppg on decent efficiency that have never been the scorer of say.. Ray Allen. Andrew Bynum averages 15/10 on great percentages but Paul Millsap can average the same thing and notW be considered nearly as good because he cant dictate games like Bynum has at points.

There's more to basketball games than raw scoring/passing efficiency.. momentum, making the right plays, effort, off the ball positioning, defense.. they all cant be accurately measured. If you have a guy that can take over games at will and make amazing momentum swinging plays, that guy may be more important to your team winning than a big man who just scoops a lot of offensive rebounds and gets a lot of putbacks even if that big man has a higher FG.. or a guard that constantly gets a lot of his points when he's fed open jumpers or gets a lot of opportunities on the fast break.

I perfectly understand PER.. and TS.. and eFG. TS and eFG at least have concrete concepts behind them. They are VERY simple formulas. Winshares and PER? They are very complicated formulas that stretch for paragraphs that apply different SUBJECTIVE weights to the stats of their creator's preference. PER isn't that bad in my eyes though because even if you altered some of it's formulas, you'd still generally have the guy with the best/most efficient statline be on top. Of course efficiency is not the most important thing in basketball.. much more to the game when it comes to winning ball games than efficiency.

Anyone thats ever played even remotely competitive basketball knows the emotions and natural swings of momentum that occur often dictate where the game will end. Guys and teams that can control that momentum the best are the ones that win.. even if they dont post 40+ PERs.

[Dasher]

Your ball on the roof bro. Dont be mad.

Nah they made the same huff and puff anti-advanced statistics argument that is always made on ISH. The game is art. People who use advanced statistics don't watch games. Hakeem Olajuwon 1995 season, which actually does more to prove his regular season dominance and the viability of win shares than anything. He was 11th in the NBA in win shares after missing ten games and playing for a team that only won 47 games. The numbers say Hakeem was playing on an otherworldly level. So how in that instance does that season invalidate it? The real point of advanced stats is that they are supposed to remove the bias that the traditional basketball narrative that relies too much on the story and sportswriter cliches to allow front office personnel to make better decisions. Advance stats and information technology has been shown to be successful in soccer, a sport a lot like basketball, and with time it will revolutionize the game of basketball.

[chips93]

seems to me like team success plays way too big a part in this metric.

lots of great players, who have been consistently great, have their winshares fluctuate a ton, often correlating with how well their team is playing.

how much does defense come into win shares?

hypothetical:
say superstar player A is on a team where he puts up great offensive stats, but he has little help on offense, and their defense sucks. they lose a lot of games, so his winshares are pretty low. then next year, they hire a great defensive coach, they improve their defense vastly, but their offense remains mediocre, with superstar player A continuing to carry them offensively, but they win way more games, mostly on account of their defensive improvement. wont superstar player A's winshares go way up even though hes playing virtually the same?

[StateOfMind12]

Nah they made the same huff and puff anti-advanced statistics argument that is always made on ISH. The game is art. People who use advanced statistics don't watch games. Hakeem Olajuwon 1995 season, which actually does more to prove his regular season dominance and the viability of win shares than anything. He was 11th in the NBA in win shares after missing ten games and playing for a team that only won 47 games. The numbers say Hakeem was playing on an otherworldly level. So how in that instance does that season invalidate it.

Will this invalidate it then?

Terry Porter and Reggie Miller both had higher win shares than Hakeem Olajuwon did in the 89'-90' season. Oh and you want to know something more interesting? Reggie Miller and Hakeem Olajuwon both played the same amount of games. The only difference was the Reggie Miller won just ONE game more than Hakeem did. The Pacers were 42-40 and the Rockets were 41-41. If such a stat penalizes someone and shows someone is inferior than another player just because he had one less win than the other player then the stat sucks. Well, the stat win shares sucks and is useless.

You want another one? Ok...

How come Brent Barry in the '01-'02 season had higher/more win shares than Jason Kidd did? That was the season where a lot of people argued that Kidd should have been the MVP.

What is even more funny about that is that Brent Barry played one less game than Kidd did and played on a team that finished with a worse record than Kidd did. Barry played 81 games in the season and his Sonics finished the season with a 45-37 record while Kidd played all 82 games and his Nets finished the season with a 52-30 record.


The stat sucks....

[Duncan21formvp]

One of the best stats around, just like in Baseball which is called Wins above replacement.

http://www.baseball-reference.com/le...R_career.shtml

[ZenMaster]

I under stand them just fine and they are easily manipulated.

They put too much emphasis on FT's and that's why guys like Durant usually has a high TS%, because he gets a conga line to the FT line and it can hide the fact that he may have shot poorly from the floor.

They are just useless stats that people use as a crutch to make a player look like he played better than what he did. Then when you look at real stats like FT%, 3-point%, FG% they tell the truth.

TS% is "easily manipulated" because points from the free throw line for some reason don't count as much as other points.

Got it.

So if you coached basketball would you tell your players to take semi-open jumpers instead of trying to force fouls? And would the reasoning behind it be that they shoot a better FG%?