How much (percentage wise) do you believe you can accurately describe a player's value to a team based upon their stats? And by stats, I'm including everything from the basic pretty stats like PPG and RPG to advanced measures like PER, TS% and win shares.
PPG is ineffective until you look at FG%. RPG depends on so many factors, size, opponent FG%, and I like to look at O-boards to judge a players hustle that way.
I have no problems with stats, but I think you need to look past them when defining a player. Is there a stat that records the energy that a player brings to a team? How about a "leadership" stat? None of that is recorded, yet both contribute greatly to any team. So, I don't like just looking at PPG, RPG etc and determining how great a player is. Other factors need to be taken in consideration, too.
Stats are a foundation for evaluationg a player, but the intangibles a player brings to a team can only be appreciated by watching someone play. For example. Zach Randolph has great stats, but he's one of those players who disrupts the flow of a teams offense because he doesn't pass/rotate the ball. Once he's got it, he's putting it up and it hurts the rest of the team. Sure he's close to being a 20/10 guy and that's impressive, but he doesn't actually contribute much to a team being a winner. PJ Brown has average stats, but his solid positional defense and clutch playmaking in close games cannot be appreciated by stats, you have to see him play to realise he brings a lot to a team.
Statistics need to be put in context when evaluating a player. That's really all I have to say about it. I know people keep trying to develop more and more precise forms of statistical analysis, but at some point you just have to watch the game to better understand how/why a player is putting up those numbers.
I think if one has various "advanced" (Let's say, TS%, PER, etc.) stats on a player, most of the time they have a good gist of what tier that person is at offensively. Evaluating Defense is a bit trickier though.
<div class='quotetop'>QUOTE (cpawfan @ Jul 23 2008, 10:21 PM) <{POST_SNAPBACK}></div><div class='quotemain'>How much (percentage wise) do you believe you can accurately describe a player's value to a team based upon their stats?</div> If by stats you means numbers that can be determined via the box-score, then depending on the player it can be a pretty good indicator to a lousy one. But I wouldn't put any limit on what "stats", in principle, can tell you. If someone recorded every minute detail of every play in a database, including extensive contextual information, they organized all that information in a very clever manner, and they figured out a super-intelligent way to extract relevant information and make predictions from it -- it might tell us quite a lot about a player. Maybe even more than what scouts could tell you. The difficulty in basketball is separating team effects from the individual. If player A is scoring a bunch of points on a particular team, to what extent is that really helping his team win? And then .. how would scoring ability translate on another team? Very difficult questions to answer.
I think that in general, the average fan highly overrates stats. I'm not one to put a ton of stock into them, especially in a discussion or comparison. I think team stats are far more telling than individual numbers.
I think highly of using stats to evaluate players, but you have to consider them all together to get as whole a picture as possible. In the late 1970's, sabermatrician Bill James started a trend of new-age statistics for baseball and now the discipline is apparent in most sports. When REAL GM's and baseball managers started paying attention to the new stats and tailoring their teams and lineups accordingly, they became winners with players who you might not think were winners. Similar stories in the NBA front offices and benches, as well. Though it's also true that many teams pay no attention to them (and sign guys like ZBo to big contracts).
<div class='quotetop'>QUOTE (Chutney @ Jul 24 2008, 10:58 AM) <{POST_SNAPBACK}></div><div class='quotemain'>Statistics need to be put in context when evaluating a player. That's really all I have to say about it. I know people keep trying to develop more and more precise forms of statistical analysis, but at some point you just have to watch the game to better understand how/why a player is putting up those numbers.</div> This guy got it right.
The only real stats I pay attention to are FG%'s, eFG%, etc because they measure efficiency. And the more efficient a player is, the more likely it is that something good will happen when the ball is in that players hands. Also, things like Assist/TO ratio to measure a PG's ability in taking care of the ball, etc. Otherwise, you need to watch the games. Things like boxing out, reading passing lanes effectively, making the extra pass to the open man, etc can never be measured in numbers.
To make a general (for the whole NBA) assumption about this is very difficult because it varies from player to player... But that's too generic. A better way to make this comparison is to say it varies greatly from the TYPE of player you are. For example, for a defensive specialist like Battier or Bowen stats tell 40% of the story. For an offensive only player such as JR Smith the value would be closer to 70% In general, for all of the NBA, I'd take a shot in the dark and say about 65%.
Agree with those that say stats need to be put in perspective. I also think you learn a lot more about players by actually observing them play then you do by reading and manipulating box scores. Analyzing stats will never be a substitute for good old fashioned scouting, but as time goes by and statistics evolve, they are becoming more and more reliable as an indicator of a player's true value.