In Defense of Hollinger

The purpose of this post is less to defend Hollinger (he is, after all, the one in discussion who is gainfully employed as a full time sportswriter...I don't think he needs any defending from myself)  than to attempt to correct some gross misconceptions observed on Celticsblog. The motivation is not so much due to how much or little I enjoy Hollinger's writing and ideas personally, but is rather due to a desire to continue the tradition of intelligent discussion on Celticsblog, rather than baseless babble.

As a caveat, Hollinger himself acknowledges the shortcomings of all of his rankings and statistics; no one thinks they are perfect.

And as one final, probably useless caveat (since too many people read what they want to read and not what is actually written) I do not think that Hollinger or his rankings are perfect or the be-all end-all. I think they are a valuable and often misunderstood and underrated part of basketball analysis. Nothing more, nothing less.


1. "All wins are equal"

This is simply not true. Yes, the BCS college system moved away from margin of victory in it's determining of school rankings. However, they did not make this change to create more accurate rankings, they did this because the public thought it was unsportsmanlike to watch Texas beat Northeastern South Dakota Tech by 70 points to help their BCS ranking. To appease the public and help remove the motivation for endlessly running up the score, the BCS removed margin of victory; however, were they able to still include margin of victory in addition to their other markers, they'd have a more accurate ranking.

The fact is, not all wins are equal. A blowout is more indicative of a superior team than a close fought victory. This is not only true logically (a truly great team does not need many games to hinge on one or two posessions down the stretch; a truly great team should have most games well in hand with time to spare) but has been proven time and time again by people researching the long-term overall relationship between margin of victory and win-loss record.


2. "He Put Them Too Low"

I don't know how many times this needs to be explained by posters here or Hollinger himself, but believe it or not, Hollinger has no rooting interest in placing teams high or low. Not only that, but he has no ability to place teams in any specific spot. Instead, what Hollinger has done is he has created a computer program incorporating data from all teams over the past 35 years or so that says: "Team A has 13-8 record with a 3.2 avg. margin of victory against opponents with a .549 winning percentage. In all other analyzed years, the average team with those statistics has finished at such-and-such a record with such-and-such playoff success, so maybe, logically, it would be most likely to expect a similar result from this team this year." Once Hollinger has written his prediction formula based only upon past results and data and not upon scouting or opinion, he has no ability to move teams up or down as he sees fit.


3. Specific Situations vs. Overall Odds

It seems like there is a huge misunderstanding of statistics around here. Statistics become more accurate with more data over a greater sample size and less accurate in more specific situations over less time and a smaller sample size. Thus Hollinger's rankings, while more accurate than most prognostications, can seem to be off-base when talking about a certain season, can be very awry when only looking at your favorite team, and can be completely askew when it comes down to 2 teams in one seven-game series. However, if Hollinger's predictions are wrong about your favorite team or a few playoff series go against his predictions, THIS DOES NOT MEAN HIS ANALYSIS IS INVALID! It simply means that these specific situations bucked their predicted trends. Which leads to...


4. What Is Statistically Significant?

Over long periods of time, shockingly small rates of success are statistically significant. All sports prognosticaters are wrong, and are wrong a lot of the time. Though some may have very good years, and very bad years, I'd be willing to bet that most prognosticaters are correct in their predictions about half the time over the course of their careers. Of course such talking heads love to ignore when they are wrong and endlessly point out when they are right to make it seem like they are right more often than not...but they are not. Take a look at Las Vegas. Over the course of history, I bet that the underdogs have covered very close to 50% of the time; that's how Vegas makes it's money. If you could be right even just 55-60% of the time, over a long period of time you'd make a lot of money. That's all we are talking about here, that's all Hollinger has to hit to be a well above average sports prognosticater. And that's over the course of the entire league, not just the Celtics. He doesn't need to be right about every team to be significantly better than average!


5. The Past Matters...Just Not How We Think It Should.

The past only matters in terms of determining and refining the formulas used to analyze current results, not in determining which team is better than others for the current year. Each year of data added makes predicting future seasons more accurate. It's very tempting and easy to say "well, Team B won it all last year, so they're the best team this year!" There are so many problems with such a statement, it's almost overwhelming. First, logically, how could this be true? At some point it would have to end, as Russell & Cous aren't winning titles anymore, sorry  to say. Additionally, when's the last time there was a repeat champion? It's been a while. Look at how bad Miami was the year after it's title and even worse the subsequent year; look at the dropoff Chicago had after exploding into the playoffs. So here's the question: when were those two teams bad again? By the logic of a lot of people on this board, Miami was the title favorite in '06-'07 until they were out of the playoffs, but this is not true; it was clear quite early that they were a much worse team than the year before. The same is true for Chicago; very early on it was evident that they needed to be re-evaluated and not judged on the past season's success.


6. "Our SOS is lower because we keep beating teams"

I don't really know what to say to this except that Hollinger clearly states that the strength of schedule of Team A is determined by finding the winning percentage of  all of Team A's opponents against all of their opponents except for against Team A. In other words, Team A cannot impact it's own strength of schedule.


7. Data Based on Past Results as a Predictor of the Future.

I really can't stress enough that Hollinger's predictions are not what he thinks will happen, it's an analysis of what has happened so far in relation to the history of the NBA. A team being ranked #1 does not mean that team is necessarily "better" than the #2 team, but that the #1 team has simply played better so far. However, analysis has shown that using such tools as predictors often yields more accurate predictions than predictions based upon other criteria; this is why Hollinger uses his method of analyzing results that are in the book to attempt to predict future outcomes.

8. The Current Cleveland Example

There are three teams playing phenomenally right now: Cleveland, LA, and Boston. Thus far, Cleveland has posted the most impressive results. I'm ok with that, and it has very little bearing on the actual outcome of a playoff series this year between the Cavs and Celts. Last year the Cavs outscored us in their series and we were able (fairly lucky) to win...the Cavs and Lebron are a very very scary playoff team. What Hollinger said in his PERdiem is basically that in any other year, a team playing as well as Cleveland has so far this season usually ends up winning the title. However, that is not the case this year because LA and Boston are also playing extremely well. But Cleveland has been a little more impressive thus far.  I'm not sure I see much wrong with that statement, as it's based upon the season so far and a lot can change. Besides, it's not exactly surprising. Lebron is pretty clearly the most unstoppable player in the NBA, and everyone has said that he needs to be surrounded by shooters and one scorer to help alleviate the pressure, and now he has that with Williams and a full season with West and Wally.


9. Above All, I Love the Celtics

And maybe tonight's +9 victory over the Jazz boosts us up a bit more.

I'll add more if/when I think of them.