from Netsdaily: <div class='quotetop'>QUOTE </div><div class='quotemain'>ESPN’s John Hollinger is nothing if not consistent. After ranking Jason Kidd #34 in his annual PER (Player Efficiency Ratings) last season, he has dropped the Nets’ captain 22 places after his near triple-double season. At #56, Kidd is right behind Utah’s Paul Millsap. Some other Nets’ rankings: Vince Carter at #29; Nenad Krstic at #74; Richard Jefferson at #118; Josh Boone at #148; Boki Nachbar at #214; Marcus Williams at #216; Jamaal Magloire at #266; and Jason Collins at #328 and dead last in the league.</div> link</p> Personally, I don't really care about this stuff. Interesting to see Collins at #328, last in league. pretty funny.</p> </p>
Blah blah blah.</p> According to his stats, last year Chuck Hayes and Eddy Curry were supposed to have better seasons than Kidd. To him basketball is numbers, he should have been a baseball analyst.</p>
<p style="margin-bottom: 0in;">OK, I'm going to try and stand up for Hollinger a bit here.</p> <p style="margin-bottom: 0in;">Firstly, there's no way that Johnny would ever stand up and say that because one player has a higher PER than another, that they are qualitatively a “better” player. All PER does is measure the statistical output of a given player.</p> <p style="margin-bottom: 0in;">PER is a HUGE improvement over most other metrics out there for ranking players. It is pace-adjusted, minutes adjusted, and is probably the only successful single statistic out there at roughly describing the total offensive contribution of an NBA player.</p> <p style="margin-bottom: 0in;">Think about it – is PPG a better stat for overall ranking? Points per 48 minutes? Shooting percentage? Adjusted shooting percentages? They all have their drawbacks. And yet even if this system does have its quirks (Milsap over Kidd is a good example), it's better than any other calculated number as a ranking system.</p> <p style="margin-bottom: 0in;">I guess my point is, PER comes out of a formula. It's objective. And by that standard, it's better than any other single objective number out there.</p>
I object to the very concept of the PER; its a subjective statistic. That's like me going "James White is the most prolific scorer in the NBA," and then defending it by spending some time coming up with some huuuuuuge formula (and the PER is ridiculous) that just so happens to list James White ahead of everyone. To that end, its circular reasoning. Hollinger's opinion is the evidential support for his statistic, while statistics should be proper evidence to support an opinion. Its all backwards, man.</p>
<div class='quotetop'>QUOTE (GMJigga)</div><div class='quotemain'></p> I object to the very concept of the PER; its a subjective statistic. That's like me going "James White is the most prolific scorer in the NBA," and then defending it by spending some time coming up with some huuuuuuge formula (and the PER is ridiculous) that just so happens to list James White ahead of everyone. To that end, its circular reasoning. Hollinger's opinion is the evidential support for his statistic, while statistics should be proper evidence to support an opinion. Its all backwards, man.</p> </div></p> </p> Not only Hollinger's opinion, but ESPN's opinion. To make him the coverboy for stats is big support that they agree with him, and I think it's a bit foolish to do so. Just stick with normal stats and a bunch of intelligent sports analysts to judge rankings of players, it's so much easier that way.</p>
Yeah I mean could you imagine if regular statistics were based on the same concept of the PER? There'd be a guy at every game going</p> "Hmmm I don't think that shot by Kidd went in. I'll count it as a miss" "Tough luck that Lebron's shot didn't go in, but since it rimmed in-and-out, he was close enough. I'll give him the bucket"</p> And then when that guy gets to give his opinion if Lebron or Kidd is a good player, he gets to use his own opinion to back up his opinion...how conveinent.</p>
<div class="ucomment"> It’s easy to log complaints about his ratings when you don’t fully appreciate what Hollinger is measuring.</p> Part of that is his fault. He refuses to set forth the formulae he is using, stating only that we can find them in a book of his–so, it’s hard to take an objectively critical view of his methods. That said, there’s something intuitive about trying to measure a player’s “good” contributions (points, rebounds, assists, blocks, steals) minus his “negative” contributions (missed shots, turnovers). I assume that he weights each one of these measurement in some way--for example, I wouldn't be surprised if he discounts assists somewhat, because there isn't a strong correlation between number of assists and which team wins a game. The problem is, if you take and miss more shots, you’re just going to have a lower rating than someone who is a “black hole”–someone who only gets the ball in the post and never passes to a teammate. I'm not sure that the second guy is more valuable from a basketball sense. For that reason, I’d like to know whether there is a correlation between a teams’ overall PER and its record. It doesn’t seem that I’ll get that, though–maybe it is in the book?–so I’ll settle for pointing out that he’s not a trainied statistician, and many of these ideas have been lifted from baseball analysts. For example, do you realise that he is NOT ranking each player’s PER from last season–he is PROJECTING their PER for the upcoming year, by comparing each player to how some number of similar players from previous seasons did at the same age? This is actually a concept used by the PECOTA baseball player rating system, that appears to be slightly–very slightly–better than other projection systems. The idea is that you find other players with similar skills and similar heights (haha, obviously not in baseball) and see what their career arcs looked like. I see a couple of problems: first, PECOTA actually breaks up each player into categories based upon the SHAPE of their arc–some players have a sudden dropoff at a certain age, while others suffer a more gradual decline. Hollinger doesn’t seem to do this. Second, there is limited data for players with unusual skills. In baseball, this is less of a problem for a variety of reasons, but in basketball, how many analogs do you think he found for Jason Kidd? One? Three? And if that player or small group of players declined rapidly at age 35, he’ll predict that Kidd will, too, without looking closer at the numbers to see if they have validity. With regard to players with very common skill sets, who have already established their individual arc by being in the league between say, three and ten years, he probably gets pretty good results. For very young players who haven't really figured out how to play, or those with unusual talents, it wouldn't work as well, obviously. He also can't incorporate how each player in his system has been USED in generating those numbers--both the players he is rating, and their analogs. Maybe players have gone from the third option to the second option, who knows. Again, this is not a problem in baseball, where everyone has the SAME opportunities--a series of one-on-one interactions with the opposing players. That is, interaction with TEAMMATES is irrelevant, unlike in basketball. I think what he needs to do is calculate what PER each player actually earned last season, compare that number to the projections, and take a close look at the players with the largest variability to see if it is reasonable. If he’s projecting that Kidd will suffer a fall-off because, for instance, Michael Ray Richardson entered drug rehab as a 35-year-old, obviously it wouldn’t make too much sense.</p> Maybe he does that in his book.</p> I guess I'll never know.</div>
The PER formula is fine for reporting what has happened; however, Hollinger is a poor analyst and that filters into his projections.</p> I have zero respect for his projections.</p>
Hollinger said "The Nets are acting like and being treated like contenders, but it's been a long time since we've seen any evidence that they really are contenders. If folks are going to continue describing them this way, I don't think it's unreasonable to demand some tangible proof."</p> But last year said the Boston Celtics are going to win the Atlantic who have even less tangible proof it was going to miraculously improve. I mean everybody could have predicted the Celtics were nowhere near being contenders Celtics' fans included.</p> So how do you even believe what he writes?</p>
Objectively--that is, using past results to formulate an opinion--he's right about the Nets. However, players and teams aren't static; you can't measure improvement or decline just by roster changes. I'm convinced that the Nets ARE a contender, and have a balanced group of personnel, but you still need a little luck. Everyone has some injuries, but I'd wager that, of all the teams that lost two starters and their expected sixth man for a large part of the season, they're the only one that made the playoffs (except I guess for the Wizards, who limped their way in). The injuries to Cliff and Boone were pretty irrelevant, but when combined with Krstic's injury, we were forced to rely on some scrub who wasn't expected to play--who could have expected that the Nets would get what they did from Mikki?</p> The Nets have a contender-worthy core; the question is whether their supplemental parts are strong enough for the team to truly be a contender. All we can say is that they're likely better than last year; we have no idea whether it's enough, but I'm hopeful as I've said. DA and sophmore Marcus are certainly better than freshman Marcus and nothing. I still believe that Antoine will improve, and Boone should, too. Magloire is certainly better than Cliffy. Malik Allen/Rod Benson(?) vs. Mikki Moore . . . I'd be very happy with a push on that one. We need someone to take over that Eddie House role . . . can Hite contribute?</p> They've improved their defense through their roster changes--deleting House, Mikki, and especially Cliff, replacing them with Armstrong, Magloire, and Sean Williams. Both Boki andMarcus (the current weak links)made improving their defense a priority in the offseason. </p> And then you've got all the personnal stuff that hopefully has been remedied. </p> For many teams, the time to contend is short before time takes them apart. That's why the games are played.</p>
<div class='quotetop'>QUOTE (Dumpy)</div><div class='quotemain'></p> <div class="ucomment"> He refuses to set forth the formulae he is using, stating only that we can find them in a book of his–so, it’s hard to take an objectively critical view of his methods.</p> </div> </div></p> http://www.basketball-reference.com/about/per.html</p> </p> <pre>uPER = (1/MP)* [ 3P + (2/3)*AST + (2 - factor*(tmAST/tmFG))*FG + (FT*0.5*(1 + (1 - (tmAST/tmFG)) + (2/3)*(tmAST/tmFG))) - VOP*TO - VOP*DRBP*(FGA - FG) - VOP*0.44*(0.44 + (0.56*DRBP))*(FTA - FT) + VOP*(1 - DRBP)*(TRB - ORB) + VOP*DRBP*ORB + VOP*STL + VOP*DRBP*BLK - PF*((lgFT/lgPF) - 0.44*(lgFTA/lgPF)*VOP) ]</pre> <pre>factor = (2/3) - (0.5*(lgAST / lgFG)) / (2*(lgFG / lgFT)) VOP = lgPTS / (lgFGA - lgORB + lgTO + 0.44*lgFTA) DRBP = (lgTRB - lgORB) / lgTRB</pre> <pre>pace adjustment = lgPace / tmPace</pre> <pre>aPER = (pace adjustment)*uPER</pre> <pre>PER = aPER*(15/aPERlg)</pre> <pre></pre> I definitely agree with your post, though. I'm happy with Hollinger in comparison to most other basketball "experts" out there, but I definitely think he could do better work.</p>
<div class='quotetop'>QUOTE (peg182)</div><div class='quotemain'></p> Just stick with normal stats and a bunch of intelligent sports analysts to judge rankings of players, it's so much easier that way.</p> </div></p> </p> Slim pickins my friend. Slim pickins.</p>
