Most important advanced NBA stats; W/S on perception v. reality with Rondo

[elements1985]

What are the most important advanced NBA stats?

I've been looking into win shares on Basketball-Reference -- are they similar to WAR in baseball (in terms of importance)?

And do Rondo's win shares suggest he is NOT an elite player? Rose v. Rondo isn't even close....

Are assists overrated?

[Corey]

Win shares are pretty useless.

You might enjoy looking this page over:

http://www.82games.com/1213/12BOS1.HTM

[se247042]

Win shares are pretty useless.

You might enjoy looking this page over:

http://www.82games.com/1213/12BOS1.HTM

just looking off data shown here, we really should try a Garnett-Sullinger-Pierce-Green-Lee type lineup

good share though

[elements1985]

Win shares are pretty useless.

You might enjoy looking this page over:

http://www.82games.com/1213/12BOS1.HTM

Those floor statistics at the top are interesting, but supporting cast seems introduce more variables (and thus a greater dilemma in evaluating individual performance, systematically ... same applies to assists, IMO). Or is isolating player performance more a baseball thing?

Either way -- and perhaps I'm reading them wrong -- those numbers don't take to kindly to RR. His figures aren't even close to Rose's.

That's an informative site, though. Lot of good categories on there. Thanks for sharing.

[elements1985]

This is interesting: http://www.82games.com/1213/ROLRTG8.HTM

[Opalmerr]

Stats...

They simply cannot account for intangibles that present a myriad of variables.

Trust your eyes.

[leprechaun5]

Stats...

They simply cannot account for intangibles that present a myriad of variables.

Trust your eyes.

since when you have become hustlenomics best friend ?

[Opalmerr]

since when you have become hustlenomics best friend ?

There has come to be over emphasis and over analysis of stats with the advent of Sabermetrics. The whole advanced stats concept works better in some sports than others. There can be value in them but one must understand what they actually "say", know how to properly interpret them and appreciate they are not necessarily the bottom line of interpreted performance. There are not suddenly a lot of "new" statistics available to the sports world. They are just being spun differently. And stats can be spun so many different ways to support a realm of varying opinions. In the end, it still comes down to the same handful of stats that have always mattered... which is what James (and others like him) has made bundles on by repackaging in different forms.

[elements1985]

There has come to be over emphasis and over analysis of stats with the advent of Sabermetrics. The whole advanced stats concept works better in some sports than others. There can be value in them but one must understand what they actually "say", know how to properly interpret them and appreciate they are not necessarily the bottom line of interpreted performance. There are not suddenly a lot of "new" statistics available to the sports world. They are just being spun differently. And stats can be spun so many different ways to support a realm of varying opinions. In the end, it still comes down to the same handful of stats that have always mattered... which is what James (and others like him) has made bundles on by repackaging in different forms.

I don't, nor do I pretend to possess really level of expertise in advanced stats; but there seems to be some wild notion in the sports and business worlds that people ought not be evaluated. If you're a GM and about to pay X player Y dollars for Z years, wouldn't various methods of analysis be useful in determining the player's value? Wouldn't a different take help before handing out a sizeable contract?

Why is it okay to complain about a certain player's contract, yet nitpicky to invoke advanced stats in assessing his value? Why is it okay to play GM, but trivial when introducing the methods of analysis used by actual professionals in the sports industry?

Personally, I see advanced stats as a form of business analysis....

[xk4]

Advanced stats should be used for high usage players, volume scorers, ball-handlers. TS%, win shares, PER, they might tell you when a player is helping or hurting the team.

Basic stats...if a spot up shooter is hitting on 2/5 open 3s that's all you need to know. If a guy gets a lot of rebounds, or blocks or steals, he's probably hustling and helping the team.

Your eyes and brain should used to evaluate role players and energy guys. There's no stat to measure:

bbiq: orchestrating offense, defense, calling out assignments
setting good screens
playing good on-ball, post, or help defense without getting a steal, block or rebound
taking a charge/ diving for loose balls
passing to an assist
tipping a rebound to a teammate
boxing out
taking very few shots/playing on a team with a ball hog

That's my take.

[Hustlenomics]

since when you have become hustlenomics best friend ?

swing and a miss

[Opalmerr]

I don't, nor do I pretend to possess really level of expertise in advanced stats; but there seems to be some wild notion in the sports and business worlds that people ought not be evaluated. If you're a GM and about to pay X player Y dollars for Z years, wouldn't various methods of analysis be useful in determining the player's value? Wouldn't a different take help before handing out a sizeable contract?

Why is it okay to complain about a certain player's contract, yet nitpicky to invoke advanced stats in assessing his value? Why is it okay to play GM, but trivial when introducing the methods of analysis used by actual professionals in the sports industry?

Personally, I see advanced stats as a form of business analysis....

Statistics have value.

Advanced statistics have value.

However, at best statistics provide a partial picture. There are too many intangibles and variables associated with sport statistics to take them as absolutes. Advanced statistics seek to account for some of these intangibles and variables, but cannot account for them all.

I have had the pleasure of conversing with Bill James on the topic of Sabermetrics. Great stuff, sabermetrics. In a nutshell what he does is take traditional statistics and combines them in innovative ways to create "new" advanced statistics. Now, the purpose of advanced statistics was not only to measure historical production in regards to winning and losing but also to provide a means of predicting furture performance. Bill James Sabermetrics for baseball is the most well known of these although they do exist for basketball, football, soccer and other sports.

Baseball provides an easily understood and excellent example of what advanced statistics attempt to do. The traditional Fielding % statistics used to help determine Gold Glove winners measures the percentage of cleanly handled chances a player has at a given position. The "new" advanced statistic called Range Factor places emphasis on the total number of outs a player participates in from that same position. The difference being Range Factor in it's simpliest form doesn't account for mishandled chances (errors) concentrating on the postive bottom line of outs participated in. Both statistics have value but can give contradictory feedback on a given player.

So, it takes a practiced eye to properly interpret statistics, especially advanced statistics. For the average Joe Poster on Internet forums such as this, we are as likely to misinterpret an advanced statistic as we are to apply it correctly. One simple reason for this is many advanced statistics are not meant to be "stand alone" statistics but are meant to be viewed in conjunction with other statistics, advanced and / or traditional, to form a more overall picture.

One of the more fascinating sites for me shows NFL advanced statistics that utilizes innovative statistics and formulas to "grade" offensive linemen among other positions. Fascinating to me anyways.

Statistics tell part of the story, but certainly not a complete story. xk4 points out some of the intangibles that advanced statistics for basketball have a difficult time accounting for. Critics of advanced statistics often point to "things that cannot be found in boxscores" but even some things included in boxscores can be skewed by how they are interpreted.

In the end, trust your eyes... at least as much as you do any statistics.

FWIW, maybe not much at all, I only keep a bare minimum of statistics for the teams I coach and do so only for award purposes. I rarely look at statistics in season. Have had 1 losing season in 30 years. I trust my eyes, my mind, my experience. I do like statistics though. Productivity in relation to minutes played with this Celtics team, shot chart and assists to made baskets ratio are good predictors of overall TEAM performance with this group. Rebounding, not so much.

Would it be fair to say some Celtic players would be "better" rebounders on teams that have a different philosphy on transitioning between attack and defending? If so, it is also fair to say a great rebounder on another team might struggle to rebound as well in the Celtics system, is it not?

This is where they create innovative formulas to measure rebounding in percentages and opportunities...which only serves to introduce other sets of variable and intangibles. At some point it becomes complicated to a point as to actually disguise what was originally sought to be measured. Sometimes the KISS method really is the best method.

[69centers]

In the end, trust your eyes... at least as much as you do any statistics.

Good point. Too bad there are some who only look at stats.

[Corey]

And some that fail to have the capacity to comprehend them.

[69centers]

And those who immediately say "you don't understand them" if you don't use them.

And those who think they rule the world (and PSD forums) if they use them.