Hollinger stats sucks
by JustAnotherFan on Jul 5, 2008 9:40 PM CDT
29 comments
Lately I read lots of posts quoting Hollinger's PER. Personally I think that is crap made by a frustrated man who could not put his statistics skills for better use.
I don't get why people care to give attention to this. Unless you agree that Carl Landry is more efficient than Steve Nash or Deron Williams, Brandon Haywood would go over Tyson Chandler and that JR Smith and Rip Hamilton are basically the same thing...
With that excuse for analysis Hollinger keeps posting his fallacies disguised as mathematical stuff to say the things he believed anyway despite the numbers
As someone once said: statistics is the art of torturing the numbers until they speak your truth...
FanPosts are user-created posts from the BlogABull community, and are to be treated as the opinions and views of that particular user, not that of the blogger or blog community as a whole.
0 recs |
29 comments
Email
Print
Comments
Hollinger's numbers aren't the end all be all for determing who is a good player
So, I don’t see the point in bashing them. That’s just like saying who has the highest points per game average every season is the best player.
by Ozzie Montana on Jul 5, 2008 10:20 PM CDT reply actions 0 recs
You mean the guy
that had Dwyane Wade’s 2006 playoff performance as the best ever? No way…..
by Jbonelli on Jul 5, 2008 10:26 PM CDT reply actions 0 recs
yeah what a moron
ive seen hollinger write some dumb things in the past but ever since he wrote that article that put dwade’s 2006 finals performance over ALL SIX of jordan’s i decided never to read his crap again
by sin on Jul 6, 2008 1:04 AM CDT up reply actions 0 recs
I hope you boycott every columnist who has made stupid proclamations
by Ozzie Montana on Jul 6, 2008 1:31 AM CDT up reply actions 0 recs
nah
just the ones that continue to do it over and over
by sin on Jul 6, 2008 1:40 AM CDT up reply actions 0 recs
Wade being number 1
That was just horrible. I think his stats are ok. Its a good TOOL. Not a system. It should be looked at what it is, a section of a scouting report on somebody you can use or as another point of view in statistical comparison. But common sense at some point kicks in. I live in LA and this year when the Lakers were playing Celtics, my friends all were kind of nervous. I am like arent you the same guys that were arguing he is better than MJ and you are nervous? They are like what do you mean/? I told them basically when MJ was winning his, every time we went to the Finals, it was more of a celebration and a show as to what MJ will do to win it this time. I was never ever nervous or unsure of if we will win. That was almost a given to me. I knew we had MJ and we werent losing and he never let us down. So back to Hollinger, sometimes you have to actually watch some tapes too and use your common sense. I bet if he actually sat down and watched the MJ finals, any of them, and then watched Wade’s, he would find a formula to make sure MJ was listed top 6.
by NamingRightsOnSale on Jul 6, 2008 6:45 AM CDT up reply actions 0 recs
the other reason
it was so ridiculous was he actually picked that performance subjectively – while wade did rank #1 by PER, he then “ranked” the performances subjectively and still put it #1.
by Jaina on Jul 6, 2008 7:39 PM CDT up reply actions 0 recs
Nobody says any of those things
PER and per-minute stats in general should be taken with a huge grain of salt for players who don’t play much, like Carl Landry. There’s just not enough of a sample size there to tell us much about those players (except maybe that they should be playing more). PER also doesn’t measure defense, which is why Tyson Chandler is better than his PER and J.R. Smith is worse. If somebody claimed that PER is the only way to evaluate players, yeah, I’d agree that that’s bullshit. But nobody actually claims that; all people claim is that it’s a useful way of comparing players who don’t play similar minutes. Of course it’s got its flaws and limitations, but so do all stats. Should we not pay attention to any stats at all?
by Big D on Jul 5, 2008 10:46 PM CDT reply actions 0 recs
To a great extent I feel like reminding myself
To not feed the trolls. But on I go anyhow…I really should know better.
The issue here seems to be a misuse of PER based upon a lack of understanding of what it is trying to provide. Hollinger’s PER is a tool that, combined with a lot of additional information, is extremely useful in determining a player’s overall value and contribution to their team. A car is a tool; a car can be used to drive to the store for groceries, and a car can be used to drive into a tree. The grammatically incorrect and incoherent anti-Hollinger ramblings here (“Hollinger sucks!!!! LOLZ”) are the equivalent of standing on the street and watching a car driven into a tree, and then rambling on the stupidity of anyone ever using a car.
There are some limitations to Hollinger’s PER formula – for one, it does a poor job valuing defense and certain other measurable aspects of team play – but I find it to be a very, very valuable tool which accurately measures what it is attempting to measure. Taken in conjunction with plus/minus data and scouting observations, I think you can develop a very accurate picture of a player’s overall worth. Probably my only issue with the valuation (outside of it’s obvious limitations at measuring defense) is that it slightly overvalues “Usage” rate; while the ability to use possessions and create offense is a valuable skill, I think he overrates it a little bit. (However, the imperfections in his formula are far, far less significant and meaningful than the fatally flawed Wages of Wins formula).
There’s more than a little bit of a philosophical divide here, of empiricists versus intuitionists, which sides will never see eye-to-eye on these types of matters. I don’t wish to eliminate the intuitionists from this site, however I’d really like to see an improvement in the level of discourse. I’m not holding my breath.
Parental Advisory - Explicit Content
by Jivas on Jul 5, 2008 11:14 PM CDT reply actions 0 recs
Perfectly said almost
I was going to post what I thought in detail, then I read this. Thank you for saving me some time. I do have one comment though about it and its simply that I think a players usage rate should be used differently. I think the lower your usage rate on offense, the better it is usually if your true shooting percentage is high. That makes you a role player that doesnt need touches to be good. These are the perfect pieces to a championship team. I bet Rodman had a usage rate on offense that was laughably low, but he was more productive than anybody because of his D and team play and screen setting. Maybe Hollinger can put in something that makes defensive value show as well. Maybe points per pos against when a certain player is on the floor is a good start. Any how back to the main point, PER is a tool, use it as a tool.
by NamingRightsOnSale on Jul 6, 2008 6:51 AM CDT up reply actions 0 recs
I don't disagree with you
Both stats and “intuition/empirism” are valuable tools. I just think that some people (including Hollinger himself) use too much of it. Every week we read Hollinger write that player X is better than Y because his PER is better. And that is absurd! I don’t care if Wade’s or Jordan’s PER was better. They played with different team mates, against different opponents… You will often see that the Bulls should sign player Z instead of W because of his PER. I disagree. It is not a tool good enough to determine alone who a team should hire.
by JustAnotherFan on Jul 29, 2008 12:27 PM CDT up reply actions 0 recs
Well said
PER is one measurement of a player; it’s a good one. The only problem is when it’s given the weight of canon to support an argument, such as best player ever, etc. Right or wrong, that approach is a conversation killer, and convinvces noone. The simple reality is that stats are intuitively wieghted until they pass Dean Oliver’s laughability test; so much for their empiricism. We will all continue to have our biases of the value of an assist vs a rebound,eg, and these disagreements are what make a blog worth the while, and BaB one of the very best for the Bulls-centric fan.
by California Al on Jul 6, 2008 4:25 AM CDT reply actions 0 recs
"Lately I read"
I find that hard to believe.
by hscs on Jul 6, 2008 8:42 AM CDT reply actions 0 recs
PER is just one measure
How many advisors does the President of the United States have? A dozen? Each one by themselves would come up with some really good ideas and some not so good ideas. But when all twelve come together, the good ideas of each stay while the bad ideas get filtered out, creating a wholly effective advisory board (for the sake of metaphor, disregard the current presidential term).
So of course PER by itself will have the odd conclusions like you noted above – it’s just one measure among many. But when you balance it out with minutes played, Landry falls off. When you balance it out with your own judgment of defensive ability, Tyson leaps over Haywood. And I think J.R. Smith and Rip Hamilton are closer than you do, but going along the same lines, when you balance it out with character, J.R. Smith falls off.
Don’t rip stats just because they can’t stand on their own. I would argue that nothing can stand on its own, not even gut feeling from watching a guy play. After all, Larry Hughes, Steve Francis, and Stephon Marbury have three of the biggest contracts in the game… but if the GMs that signed them had really looked at their Hollinger stats beforehand, they would’ve realized immediately it wouldn’t turn out well.
by YaoPau on Jul 6, 2008 10:33 AM CDT reply actions 0 recs
there's a lot of misconceptions about PER in your comment
It doesn’t need to be “balanced out”. It’s just like baseball’s OPS+; a more or less made up, yet useful number that adjusts for factors outside of the individual’s control, and defines an average level of production. It shouldn’t be arbitrarily balanced out (character? really?), because using it properly means putting it in context anyway. The best thing about PER is that it adjusts for pace and minutes, and players with low minutes and good production actually get better when given more playing time.
by hscs on Jul 6, 2008 11:01 AM CDT up reply actions 0 recs
Hmm...
players with low minutes and good production actually get better when given more playing time.
You sure about that? In my experience, it’s the exact opposite. You wouldn’t expect Landry to post a 25 PER given 40 minutes a game.
And everything has to be balanced out, be it points per game, +/-, gut feeling on defense, character, whatever. OPS has to be balanced out by at-bats, defense, speed, batting average, and dozens of other factors to get a decent sense of how good a player actually is.
Just because PER is in context – adjusted per minute, and for pace – doesn’t mean it can stand alone without minutes and pace. A 20 PER in 40mpg is more valuable than a 20 PER in 20mpg, and we’ve seen so many players whose games are defined by the pace they play at (Nash, Marion, Stephen Jackson).
by YaoPau on Jul 6, 2008 11:27 AM CDT up reply actions 0 recs
You're wrong
Increased minutes generally lead to an increased PER.
I’m not sure what you’re understanding of PER is. It’s mostly a measurement of offensive production, and it’s a great first step for an in-depth comparison. MPG and pace are factors outside of a player’s control (Stephen Jackson’s best PER was when he was playing on a team 29th in pace factor, there isn’t a big difference for Marion, Nash is chicken and the egg), so adjusting for them puts everyone on a level playing field. Sample size is certainly an issue, but you can’t discount production if it comes from 20 MPG. It should only raise questions about why a productive player only received 20 MPG. PER doesn’t really measure overall value or impact, so don’t use it like it does. It should only give you a good idea about a player’s overall value or impact.
OPS (my remark was about adjusted OPS) is OBP plus SLG. Adjusted OPS (OPS+) adjusts for the park and league, and like PER, is only based on what’s considered most important. They’re both made-up numbers that don’t measure everything, but measure enough to be considered valuable.
by hscs on Jul 6, 2008 12:17 PM CDT up reply actions 0 recs
Interesting read
but the author leaves out one hugely important factor that would lead to a higher PER: players can improve their game from season to season. There’s a reason why some guys play 20mpg year after year, while others start at 20mpg and move to 30-40mpg: they improve and earn more minutes. I’d like to see the database behind the graph before making any judgments. The fact that they used Boris Diaw seems to indicate they used young players who had a chance to improve and earn the extra minutes.
The “Increased Minutes Leads to Increased PER” theory unfortunately doesn’t make any logical sense on a single year basis because of the increase in competition level and fatigue. I’m familiar with OPS+, and you wouldn’t expect a bench player with a 180 OPS+ in 150 at-bats to duplicate that number given 600 at-bats.
by YaoPau on Jul 6, 2008 12:49 PM CDT up reply actions 0 recs
so what?
Ziller used vets too. Players can improve, and coaches can be morons. It doesn’t mean more minutes will result in decreased production (Ziller’s conlusion, if you read it).
I have no idea why you’re using my OPS+ analogy as a straw man.
by hscs on Jul 6, 2008 2:17 PM CDT up reply actions 0 recs
So where do you think we're at presently?
My take is that our deficiency is just that we’re a little below the mean across the board, and the question is where and how do we either improve a little across the board, or get a lot, lot better at a couple of positions. Do you think there’s enough improvement coming from our 2nd, 3rd and 4th years players to eventually put a balanced team of above average players on the floor? Or do you think the numbers point towards the need for a bigger upgrade than realistic growth probablities indicates?
by California Al on Jul 6, 2008 4:57 PM CDT up reply actions 0 recs
Rose is a pretty big upgrade
and the Bulls dropped so far defensively last season that there really isn’t a clear need. Thomas and Noah haven’t played enough, and it will be really hard to upgrade on Deng and Gordon for the amount of time the Bulls could lock them up for.
by hscs on Jul 6, 2008 9:14 PM CDT up reply actions 0 recs
Because it was your analogy?
We’re going nowhere with this, and I think it’s mostly because it’s so hard to prove either way that PER increases or decreases with minutes. The author of the piece concluded by saying there is no correlation, and I’m fine with leaving it at that.
But I think that conclusion helps us settle your original disagreement to my post: that PER does not need to be balanced out by other factors. By definition, because we have no idea how PER will translate given a change in minutes, we have to balance it out with other factors.
by YaoPau on Jul 6, 2008 8:16 PM CDT up reply actions 0 recs
straw men suck
And arbitrarily ‘balancing out’ PER isn’t the right way to use it, or any other stat. Use some common sense.
by hscs on Jul 6, 2008 8:58 PM CDT up reply actions 0 recs
You, me, and every GM balance out stats arbitrarily
I assume you were pro-Derrick Rose, and partially because you saw Michael Beasley’s gaudy stats and arbitrarily balanced them out with questions about his height and character.
‘Arbitrary’ isn’t really accurate. It’s more guess and check, and I think both of us have seen enough players develop over the years for our balancing out to be fairly accurate.
by YaoPau on Jul 6, 2008 9:45 PM CDT up reply actions 0 recs
you know what else sucks?....
the bulls starting line up
by PippenAintEasy on Jul 7, 2008 11:29 PM CDT reply actions 0 recs
Hollinger's PER stats and the value of IQ scores
American philosopher, John Dewey, was asked what he thought about IQs and he told the following story:
An IQ is like a New England farmer who wants to know the weight of a pig. He places
a board through a fence and places the pig on one end. He then places rocks on the
other end. He does this until the two balance out. When they do, he guesses the
weight of the rocks!
I love this story.
For one it is funny but also behind the humor is a truth. And that truth is that a New England farmer has a sense of the weight of rocks.
So with basketball statistics, they can point the way. They can highlight. Let you know what you already know. But you have to remember,
They aren’t the truth!
You also have to remember GIGO. Garbabe in, garbage out. You can move a whole bunch of numbers across the page to assault the sensibilities of readers until you have numbed them to reality. Basketball, one of the fastest sports, has the least statistics. Baseball, the slooooooooooooooowest sport, has tons more statistics.
I’d trust someone with baseball stats way before I’d trust anyone with basketball stats.
And it isn’t just fans, I hear sportscasters marvelling at all the rebounds this guy or that guy got way long ago. And they are indeed valid stats. The problem is in the particulars. The game has changed and so have the rules. At one time you didn’t wait until the 5th foul to start shooting non-shooting fouls. Imagine how that would add rebounds to your average. Even in the same year, someone who HAS to take shots at the end (bad shots) simply because they’re behind, won’t have a great shooting average. Also, what most people around realize is that there are some people who just go for stats all the time. Sometimes people go for stats at the end of the year like when Michael Cage and Oakley were fighting for a rebound title.
I obviously feel strongly about this and am sorry to go on like this but
Heh
I’m a way old guy.
"To a man with a hammer, every solution looks like a nail." SLC
by WayOldGuy on Jul 9, 2008 9:20 AM CDT reply actions 0 recs
Comments For This Post Are Closed
Secondary Sidebar
FanPosts
Community blog posts and discussion.
Certain photos copyright © 2009 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
