I agree with you on this but this hopefully, when RJ's % drops... Blair(27%), Hill(36%) will play better, Splitter will be in shape and more confortable with the team...
And any team on a hot streak has one or two players playing out of their minds (this is true of the Lakers with Gasol/Odom)
It is way too early in the season to put any real stock in Hollinger's rankings. I actually like his model, but not 8 games into the season. There's not enough comparative data to make any sense, especially when the strength of schedule is so skewed this early. Does anybody really expect Houston to play at a .167 clip all year or for Golden State to win 71 percent of their games? Hollinger's rankings will look a lot more realistic in another month or so.
These same Hollinger rankings had the Spurs #1 at the end of the regular season in 2007 when the Mavs finished 67-15 (and the Spurs were 58-24); who's laughing now?
I actually think his rankings are pretty good, though if there were some way to factor in injuries and put less weight on the very end of the regular season, it would be better. The Lakers were in the top 2 in Hollinger's rankings most of last season, but fell to #8 or so in the last 2 weeks of the regular season because they started playing their starters far fewer minutes.
Edit: Small sample size makes the current rankings pretty worthless anyway, but they give you a good idea of how a team is doing beyond their record later in the season. Hollinger did have Phoenix ahead of San Antonio (and both ahead of Dallas) at the end of the last regular season.
Certainly. The model/formula is what it is. I mentioned that two of the primary variables were scoring margin and recent performance, but that's not the entire story.
The scoring margin for home wins is slightly diminished relative to road wins by a fixed factor, for starters, because Hollinger is (likely) attempting to quantify the tangible advantage that teams have on their home floor. Playing on one's home floor offers a tangible, demonstrable advantage - I'll assume that you have no problem with this fact.
But, on to the bigger picture - you say the model is "non-objective" - I disagree, but I think I know what you are trying to say. The model is objective, by definition, but the variables that go into the model along with their relative contributions are arbitrary. That's perhaps the term you were looking for.
Here's Hollinger's formula -
RATING = (((SOS-0.5)/0.037)*0.67) + (((SOSL10-0.5)/0.037)*0.33) + 100 + (0.67*(MARG+(((ROAD-HOME)*3.5)/(GAMES))) + (0.33*(MARGL10+(((ROAD10-HOME10)*3.5)/(10)))))
SOS = Season win/loss percentage of team's opponents, expressed as a decimal (e.g., .500)
SOSL10 = Season win/loss percentage of team's last 10 opponents, expressed as a decimal (e.g., .500)
MARG = Team's average scoring margin
MARGL10 = Team's average scoring margin over the last 10 games
HOME = Team's home games
HOMEL10 = Team's home games over the last 10 games
ROAD = Team's road games
ROADL10 = Team's road games over the last 10 games
GAMES = Team's total games
The formula is both objective and arbitrary. But, the beauty of mathematical models is that this is O.K. There is no absolute truth to any of this. He created a model, explained his rationale, and the results are what they are.
Scoring margin is chosen to be the chief variable over win-loss record because historically-speaking scoring margin has been a far more reliable predictor of championship success. That's why he uses scoring margin.
Apart from that, though, the remaining factors and relative contributions that each variable has on the final rating are arbitrary and based mostly on reasonable, educated hypotheses.
Recent performance is more important than performance 4 months ago, according to Hollinger's formula. This point is certainly arguable. People are free to agree or disagree with either the premise or the arbitrary relative contributions that Hollinger has chosen to emphasize recent performance over past performance.
Teams have a demonstrable advantage on their home floor - this is a fact. But, the actual "deduction" from scoring margin that Hollinger applies to home wins (compared to team wins) is thoroughly arbitrary and is certainly open for debate. Do we deduct 2 points for home wins? 3 points? 4 points? The premise that teams have an advantage playing at home is an evidenced-based fact. But, the actual, numerical "penalty" on scoring margin that Hollinger gives to home wins is thoroughly arbitrary and most certainly open to argument, bearing in mind that there is no ultimate "right" answer.
Only the core feature of Hollinger's model (scoring margin over win-loss record) is immune to criticism if you want to subscribe to the idea of evidenced-based reasoning.
Hollinger's contention that recent wins should play more into the rating than past wins is merely a reasonable hypothesis; this isn't an evidenced-based fact and therefore this premise is certainly open to argument, although I'd be hard pressed to find anybody who actually disagrees with the premise.
Assuming you agree with the above premise (recent wins should "count more" than past wins), the actual numerical factors that Hollinger chose to emphasize recent wins over past wins is arbitrary and most certainly open to argument.
No model ever claims absolute knowledge. Hollinger's model is certainly no exception. Unfortunately, I find that people often disagree or criticize Hollinger's model for the wrong reasons.
The primary feature of the model is that it emphasizes scoring margin over win-loss record. I have no problem with that. Strictly speaking, if the model just used win-loss record, it would be thoroughly redundant and useless. I'd be happy to consider other models that used both scoring margin and win loss record. Such models would be compatible with the fact that scoring margin better predicts championship success than win-loss record as long as the relative weight of scoring margin compared to win-loss record was greater than .500.
Lastly, regarding the "meat" of your contention with my post when you made this statement
I agree completely. The model does not take into consideration the potential of injured players, and therefore I was not justified in using that to rationalize the rankings. I agree 100%. My rationalization of Houston ranking over the spurs was nonsensical because the model does not take into account those factors to begin with.It's really nice that this fancy, mathematical, non-objective formula requires how a team seems or gives credit for the potential of injured players that don't actually play.
I agree. That's a much better way to put it. There's a caveat to follow.
Again, I agree with the premise, and I agree with the logic. However, the Spurs are ahead of both Houston and Phoenix in both win-loss record, scoring margin, and they have a head to head win over each team to boot. That's really odd. Assuming the math is correct (I won't even attempt it), there's some anomaly there which we may have to chalk up to insufficient data. On the surface, I think I'd guess that you need a minimum of 20 games in order for it to start to flesh out.
Ah, I just took your word that it did. It seemed the only thing that might explain the anomaly. I think it's probably just some function of the math. Someone will probably check the formulas to see if they're massaging the numbers to keep Heat fans happy.
in other words: Why do things the easy way if there is a hard way?
Hollingers formular is like claiming that running backwards is faster than running forwards. which is A: not true and B: gives you a good chance to fall on your a**.
I agree. That's a much better way to put it. There's a caveat to follow.
Again, I agree with the premise, and I agree with the logic. However, the Spurs are ahead of both Houston and Phoenix in both win-loss record, scoring margin, and they have a head to head win over each team to boot. That's really odd. Assuming the math is correct (I won't even attempt it), there's some anomaly there which we may have to chalk up to insufficient data. On the surface, I think I'd guess that you need a minimum of 20 games in order for it to start to flesh out.
Ah, I just took your word that it did. It seemed the only thing that might explain the anomaly. I think it's probably just some function of the math. Someone will probably check the formulas to see if they're massaging the numbers to keep Heat fans happy.
Agreed. I'd say it's reasonable to assume that what you described cons utes one of the major limitations of the model early in the season.
Might be a better idea for him to tell ESPN not to make him start releasing them until a bit later in the season. It costs him quite a bit of credibility to the unwashed if this is what they come up with. The more I think about it, the more I simply can't conceive that LA is in second place.
Yeah no kidding. The state of his rankings right now simply do not jive with my subjective view of the field (I, too, see the Lakers on top at this moment in time).
I think the narrow 2 point victory vs. Houston at home for the Lakers (W 112 - 110) is probably one of the major data points that's holding them back from the #1 ranking. That, and a (somewhat) narrow 5 point win at home for the Lakers against the Toronto Raptors (W 108 - 103)
Miami, on the other hand, lost narrowly to NOH (by 3) and that was on the Hornet's home floor. Their second loss to Boston also came on Boston's home floor, so you've got NOH and BOS home court advantage taking away slightly from their scoring margins of victory over the Heat for these two games.
Miami's 5 wins, on the other hand, have all been by scoring margins no less than 10, including a decisive win over Orlando by 26 points (high scoring margin against a difficult to beat Orlando bodes well for their overall rating).
I think the principal explanation for any anomalies in the statistical power rankings offered by Hollinger is the sparse amount of data. Usually (and I suspect for that very reason) Hollinger's rankings aren't made public until a reasonable amount of the season has passed. I was surprised to see them out this morning.
I have a feeling a number of discussions much like ours are linked to an "I told you so" email to ESPN headquarters.
Well, Miami just losing at home to the Jazz should certainly put a dent in their hollinger ratings.
Jazz came back being down 22 at Miami to win by 2. Familiar face francisco elson made 2 clutch free throws with .4 seconds left in OT to put the Jazz up 2, though he was actually trying to miss the second one on purpose (he failed and accidentally banked the free throw in).
pretty funny discussion.
I wonder how many fans try hard to find something, when there isn't anything at all.
I try it simple again:
Hollinger's thesis, that if the scoring margin is factored in like he does, you get a better prediction for overall success is plain FALSE.
it's The Emperor's New Clothes for half informed stats believer.
Hollinger uses an empirical analysis to create a formula. which is an acceptable scientific method, if the number of observations is sufficient.
but then he does something illegitimate.
he takes out some abnormalities (ironically it seems to have been mainly the Spurs numbers he used as his prime example) and instead of trying to equate or eliminate the extreme cases of an observation sample, he uses it to create a thesis and then adjusts his formula that way, that it backs up this thesis.
that's not science, that's charlatanry.
it's the Hwang Woo-Suk method.
for whatever reason such stuff can be pretty persistent, especially if it's apparently sophisticated and that way somehow elitist. "oh, you still just watch the game and try to see the good and the bad of a team. I use the Hollinger fomular. yes, it's a bit elaborate indeed."
maybe it would make sense if people invest some of the time they spend to figure out the formula, to just double check how often it works and how often it produces wrong results. and then think about it.
if I key in 2 plus 2 for 10 times in my calculator and 5 times the result is 3, it's acceptable to assume the calculator is buggy, while it's probably not acceptable to claim to have discovered a new mathematic insight.
No, scoring margin is, quite simply, the best predictor of any single stat. Go check the history. The question is how, after a formula that gives it so much weight, teams with a bad scoring margin (Houston, Phoenix) are ahead of teams with a decent scoring margin (San Antonio, Dallas), and I'm waiting for a math wonk to figure out the anomaly in the formula.
I did.
that's why I'm confident enough to do this claim.
Hollinger either didn't go back 30-40 years, or he just ignored those years, when his theory didn't work.
(and I guess we don't need to discuss, that the teams with the better records tendentially always have a better scoring margin. that's within the nature of the whole thing.)
Clearly you haven't, or you wouldn't continue to discount the value of scoring margin as a statistical predictor.
Just looking at it superficially, there's a pretty clear correlation between margin and ultimate success, at least since 2002. In that time, here are the champions, their scoring margins, and their placement among all NBA teams:
2002 -- LAL + 7.1 (2nd)
2003 -- SAS + 5.4 (3rd)
2004 -- DET +5.8 (tied 2nd)
2005 -- SAS +7.8 (1st)
2006 -- MIA +3.9 (5th)
2007 -- SAS +8.4 (1st)
2008 -- BOS +10.3 (1st)
2009 -- LAL +7.7 (2nd)
2010 -- LAL +4.7 (7th)
So, if you just picked the team with the best scoring margin at the start of each of those playoffs, you'd have correctly picked the ultimate champion on 3 of 9 occasions and if you'd picked the 1st or 2nd team in that category, you'd have gotten 6 of 9.
I suspect that if you dug deeper, you'd find that when the team with the best scoring margin didn't win, it was more often than not at least a conference finalist -- for example, Cleveland had the best margin in 2009 and lost in the ECF; Dallas had the best scoring margin in 2003 and lost in the WCF; and Sacramento had the best margin in 2002 and lost in the WCF.
You're so laughably ignorant it's embarrassing. There's an entire academic field dealing with this sort of thing, it's called APBRmetrics, look it up. Our sources are from peer reviewed literature (http://www.bepress.com/jqas/)
You're literally pulling things out of your ass at this moment in time. You claimed to have investigated this issue; it wouldn't surprise me one bit if your "investigation" involved nothing more than amateurish google searches or word of mouth from your "bros" at the bar.
I'm not even going to put in the effort to explain why you're wrong, it's not worth it.
What the kind of analogy is that, and what does it have to do with evidenced-based sports statistical models?maybe it would make sense if people invest some of the time they spend to figure out the formula, to just double check how often it works and how often it produces wrong results. and then think about it.
if I key in 2 plus 2 for 10 times in my calculator and 5 times the result is 3, it's acceptable to assume the calculator is buggy, while it's probably not acceptable to claim to have discovered a new mathematic insight.
Actually, don't answer that. You're so far behind on fundamental, 4th grade ideas of science, theories, hypotheses, and so forth that it would take me far too long to bring you up to speed.
jesus. you really need to take a break.
(and I'm not going into this: you are the biggest &%$%&$§*(/ whatsover in the world contest. can't see you being very successful with this style with your academic peers)
so, back to my starting point and my position:
I deeply doubt Hollingers fomulas (either PER as well as his team ranking formular) and I'm absolutely convinced that it produces multiple wrong conclusions.
(to many to ever accept it as a reliable tool)
end of story.
did I pull this out of my ass?
well, a bit it's a gut feeling and another bit is what you call the "entire academic field dealing with this sort of thing"
(some might have noticed my position about the accuracy of Hollinger over the years and my critics. well maybe not. however)
assuming that whatever I use as an argument won't count much at this point of our "discussion", I can only try it this way:
the "academic field" (which I highly respect) is NOT accepting Hollingers theories and formulas without doubts. I don't even think that he has many fans in this department. (see for example the critics on PER from David Berri)
just a quote (if you've got the time, read the whole article. best would be to read the wages of wins at all):
let's put it in simple words.http://dberri.wordpress.com/2006/11/...ciency-rating/
Having noted the importance of offensive and defensive efficiency, Hollinger proceeds to discuss a variety of measures of performance which serve as building blocks for PERs. These building blocks include Points per Shot Attempt, Pure Point Rating, Assist Ratio, Turnover Ratio, Rebound Rate, and Usage Rate. He defends these measures as “improvements” over existing metrics, often noting that the rankings that result evaluate players in a fashion consistent with what NBA observers would believe. In other words, his metrics fit what he believed about the players before he started.
Unfortunately, this is not the way science works. We do not begin with our beliefs, play with the numbers until our beliefs are confirmed, and then call it a day. Models are not evaluated in terms of whether they are consistent with what we believe, but in terms of their ability to explain what we purport to explain (and furthermore, provide predictive power).
Hollingers formula don't work. (they work sometimes, that's not enough to be labeled "accurate").
the major problems are the arbitrary weights he uses for several stats.
(obstructed asked how can scoring margin be that much overrated in the maths, that it produces obviously absurd results. O_V. you gave the answer to yourself, which is pretty much what I tried to tell: it's overrated and therefore the result of the Hollinger formula is damn wrong. as you see, as everyone else can see.)
if you guys like Hollingers maths. that's ok.
I don't. and I tell the reasons.
I hope that's ok with you.
Double lol's for this one. Loved the analogies too, made my day.
So you didn't actually check the history, you just took an article someone wrote that questions the scientific reliability of Hollinger's formulas and made a bunch of unrelated conclusions based on that article without understanding what it was talking about.
jesus. you really need to take a break.
(and I'm not going into this: you are the biggest &%$%&$§*(/ whatsover in the world contest. can't see you being very successful with this style with your academic peers)
so, back to my starting point and my position:
I deeply doubt Hollingers fomulas (either PER as well as his team ranking formular) and I'm absolutely convinced that it produces multiple wrong conclusions.
(to many to ever accept it as a reliable tool)
end of story.
did I pull this out of my ass?
well, a bit it's a gut feeling and another bit is what you call the "entire academic field dealing with this sort of thing"
(some might have noticed my position about the accuracy of Hollinger over the years and my critics. well maybe not. however)
assuming that whatever I use as an argument won't count much at this point of our "discussion", I can only try it this way:
the "academic field" (which I highly respect) is NOT accepting Hollingers theories and formulas without doubts. I don't even think that he has many fans in this department. (see for example the critics on PER from David Berri)
just a quote (if you've got the time, read the whole article. best would be to read the wages of wins at all):
let's put it in simple words.
Hollingers formula don't work. (they work sometimes, that's not enough to be labeled "accurate").
the major problems are the arbitrary weights he uses for several stats.
(obstructed asked how can scoring margin be that much overrated in the maths, that it produces obviously absurd results. O_V. you gave the answer to yourself, which is pretty much what I tried to tell: it's overrated and therefore the result of the Hollinger formula is damn wrong. as you see, as everyone else can see.)
if you guys like Hollingers maths. that's ok.
I don't. and I tell the reasons.
I hope that's ok with you.
Fact: Hollinger's formulas have absolutely NO INTEREST in appealing to scientists, they are intended to try to point to strength of an NBA player or an NBA team.
I'm guessing that this was sarcasm?Only way the Spurs beat the Lakers is if the Lakers suffer an injury or Parker finally develops a consistent 3 point shot....
Both are rather slim...
Correction: I'm HOPING that this was sarcasm, it wasn't though was it? <FACEPALM>
Whether or not the Spurs beat LA has everything to do with developing consistent defense and nothing to do with Parkers ability to hit a 3......
WTF is "scoring margin" anyway? Is that the same as point diff, or yet another useless derived metric?
When I used the term, I was referring to point differential. If that's the incorrect terminology, then I simply made a mistake.
Good News: After beating the clips, we moved up from 13 to 12.
Weird News: We're still ranked behind the 1-6 Rockets.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
