Those variables all exist with regular season games, the idea behind having a large sample size is that you will eliminate the minor variables (injuries, momentum, etc.) and focus on only the relevant ones. Once again, it comes down to a larger N.
All I want to know is: Did he, or did he not pick Blazers to win before the series, like so many ESPN experts did? You can't seriously tell me that if anyone doesn't know any stats and just simply watching Laker and Cavs series would arrive a different conclusion than his elaborated garbage. If anything, he would have said 100%, which is more accurate than Hollinger at the end of the day.
I just checked his prediction, Blazers in 6, from the link in other thread. What does it say about your garbage stats, Mr. Bball-stat-genius?
Why do people go ape over Hollinger's articles? He did researched, presented historical records and trends, and showed which teams fell into which category. He never said 'they shouldn't bother playing the games because of this article.' He never said he was a genius or that the outcomes of games are predicated on his findings. Clutch had a similar post of historical Rockets stats going into game 4 and everyone liked it.
The difference is, he presented it as more than playoff aficionados, he presented in the context of using his stats approach to pick winners. And he writes so damn many elaborate pieces about them. I have no problem of occasionally breaking out aficionados. "Thus, pencil in the Lakers, Cavs and Nuggets for Round 2. Heck, use a Sharpie if you want...As for the other five spots in the second round? Those are pretty much up for grabs" There are two approaches to the prediction. One is to analyze the two teams and games played. The other is to use small sample of historical stats, which may or may not have relevance because teams, even rules, played in are vastly different. So, what additional info his elaborated analysis tells me? Nothing.
You know, if you actually read what he writes, you'll find that he doesn't merely go off what his computer tells him. He actually does analyze the teams as well. http://sports.espn.go.com/nba/playo...?columnist=hollinger_john&page=PERDiem-090417 [rquoter] Portland (64.3%) vs. Houston (35.7%): In many ways this is the most interesting first-round series. First of all, if somebody is going to beat the Lakers in the West, it will be one of these two teams. Second, things look fairly even on paper, especially after we account for the Rockets' improved play since Tracy McGrady went out: They went 22-8 minus the hobbled T-Mac. But most of all I like this series because it will help us with a big question: What matters more, head-to-head matchups or overall regular-season performance? The Blazers undoubtedly have the edge in the latter category, winning 54 games with the league's fifth-best scoring margin and coming on like gangbusters down the stretch. Not only did the Blazers match Houston's 22-8 mark in their final 30 games, they did it without a loss to a single sub-.500 team. However, Houston won two of the three head-to-head meetings, and the only one they lost came courtesy of a miraculous last-second shot by Portland's Brandon Roy. That has to give the Rockets confidence that they can end their string of first-round defeats, and if you believe experience matters, the fact that this is the Blazers' first rodeo also works to Houston's advantage. That said, I like Portland here. The young Blazers' increasingly steady play over the second half of the season bodes well for their postseason, as does their playing style: The league's slowest-paced regular-season team isn't going to be thrown off by a slower postseason pace. [/rquoter]
Oh, please. At least quote something in the same article. Plus, when he does the team analysis, he got it all wrong. Tells you how useful those analysis are. Just for fun, I'll take a stab at his "Based on NBA history, there's about a 90 percent chance that, at worst, your team will be playing a seventh game in Portland" claim. Let's take a naive prior of the odds of either team winning is 50-50, for the Rockets to not at least reach game 7, Portland need to win 3 out of 4, with 50% odds, Rockets has at least 7/8 chance of win one in game 3, 4 or 5, thus forcing game 7. I don't know how folks feel about my naive estimate of 87.5% versus his estimate of 90% using obscure historical stats, probably equally nothing, but there, at least mine's much easier.
BTW, Durv, seeing how he based his prediction largely on his computer numbers (e.g. Atlanta (53.9%) vs. Miami (46.1%): Hawks in 7. Portland (64.3%) vs. Houston (35.7%):Blazers in 6), he basically forecast games to be played with odds ratio. His team analysis is actually a sideshow. Although he did mention Houston-Portland head to head, but it did nothing in changing his forecast.
What I quoted was from the article I linked to -- the article where he actually gives his picks for each round and briefly explains why he made them. If you're going to complain about his pre-playoff picks, you should attempt to actually read what he wrote. And when you say "he got it all wrong", this is amusing to me. What do you expect -- he's a fortune teller? The guy isn't going to pick every series correctly. Why is he obligated to go with "head to head" instead of "vs entire league"? He at least made it clear that its open question which is more relevant. He wasn't being dogmatic about it, as you're portraying it. He has his views on what matters -- he think how teams individually matchup means less than simply which team was better overall heading into the playoffs. It's a valid, if boring, perspective.
I didn't say he is obligated, I just point out he doesn't adjust his numbers with team analysis, which your "he doesn't merely go off what his computer tells him" seem to suggest. I do read it.
well, it just illustrates how useful (or useless, rather) his pure computer based number is. Most people who actually pay attention to team details (Houston Media, Portland Media, Houston fans, Portland Fans). wouldn't predict a 65% (which for him translates to 6 games) Portland win. It's not that I take him as fortune teller, rather, he's selling like it, with his constant updates of odds. If his all so precise percentage numbers don't mean squat (here, the technical term is uninformative, the odds ratio doesn't converge with the mean), what good does constant update of those odds do besides the obvious?
He picked Dallas to beat San Antonio in 6, even though the projections said Dallas was an underdog. Clearly there was some adjustment there. Again, he's obviously of the opinion that the best indicator of how well a team will do in the playoffs is how well they were playing heading into the playoffs. We acknowledges and discusses other factors like player matchups, it's not like he's oblivious to it, but he evidently sees those as secondary. I think you make way too much over these forecasts. At best, these are guesses. It's just a game. You can take a supposed basketball expert, and I bet his pre-playoff picks won't be any more accurate on average than what a computer would spit out. To me, that's not the important thing, because there are too many unknowns anyways to make a really accurate projection of every series. What I care about is the actual discussion of the matchups, and the basketball arguments that are made. Hollinger does a very good job breaking that down. He understands and explains the details that matter better than most.
Further, here's some thoughts on usefulness of information: Suppose you are a bet-man. You sit there trying to figure out your odds. You know the teams records and scores of games. You watched fare amount of games since the games affect your trade. Does the additional information Hollinger offers give me edge in improving my odds estimates? Hardly any. His "90 percent" estimate hardly beat my naive estimate. Anyone who know the scores of the first two games would reach the conclusion of "slightly better than even" odds on his own.
Epic fail today, I was due for a fall But that is why I don't try to generate income from such things At least Hollinger made some predictions and rationel. Of course he is going to miss some. Like any of us making predictions public before sports events.
We are not talking about Dallas series are we? Dallas series, the only series he adjusted, is obvious: it's 55%, 45% the whole season, pretty even, and consider TD is hurt and Gino is out (season numbers don't fully reflect that). It's not that hard to see why he allows flexibility. See my earlier post for why I consider those useless. In fact, like I said, I'll take most Houston and Portland media or fans' estimates - homer aside - over him. PS: I don't know why are we arguing about odds, at this point, who cares. hmm, this is a thread about his odds predictions, right?
I don't know what you're basing that on. And the updating of playoff odds serves one obvious purpose -- it helps give a more refined idea of what the likely matchups in the next round will be. I bet every playoff team tries to anticipate what their most likely opponent would be in the following round, and that projection would have to change on a daily basis.
I certainly didn't hear many predicting a 3:1 odds ratio of Portland beating us around here, Nor is Blazeredge that confident before the series begins. Nor was Portland media thinking they have good odds. (many rather play Lakers) Is that enough?
This is an interesting question, and I'm not sure what the answer is. I very much doubt Hollinger would tell me something significant about the Rockets I don't already know. You being an informed fan, he probably wouldn't tell you anything you don't already know as well. For the more casual fans, or ones that tend to get swayed by blind homerism, his information could be useful.
