PER Diem: April 17, 2009
Using the computers to determine the most likely outcome in each round of the playoffs
The playoffs start on Saturday, so it's time to set the computers spinning and get down to the nitty-gritty.
Thus, a bit of a double feature: I'm giving you who I'm picking and why, and will also tell you what our Playoff Odds tool thinks. We had it play out the postseason 5,000 times, and based on each team's final Hollinger Power Ranking and accounting for home-court advantage, it reported back to us.
Here are John Hollinger's top five NBA observations for Friday. Insider
- A glimmer of hope for Utah?
- Can the Jazz keep their core?
- Quite frankly, Nets should keep coach
- Flip a god fit as Wizards new coach
- T-Wolves need new blood in GM role
Before you see the percentages, let me offer a critique of my own system. This method works great for projecting regular-season results, but for the postseason it produces a more distributed set of outcomes than you're likely to see in real life. In other words, it's too friendly to underdogs. No. 8 seeds, for example, are 3-47 in the past quarter-century, but Detroit and Utah were given 9 percent and 20 percent chances, respectively, of pulling off what would be massive upsets.
That said, the relative percentages of each team give us a glimpse into which teams have the best chance of advancing in each round, not to mention which clubs have the best shot at winning the whole enchilada.
Speaking of which, the table below shows how the Power Rankings see this thing playing out. As suspected, the Cavs and Lakers have the best odds of holding the Larry O'Brien Trophy this June, followed by Orlando and Portland. Nobody else's odds eclipse 5 percent in this exercise, which means there are basically four teams worth paying attention to this postseason.
Chances of winning the title based on computer projection
|Team||Odds of winning title|
But before we get too deep, let's backtrack. I'm going to start with the first round and project through every playoff series until I have a winner, and we'll throw in the projected odds along the way:
Atlanta (53.9%) vs. Miami (46.1%): The computer sees this as the most evenly matched first-round series, and that matches the general perception that this one will go at least six games, if not the full seven.
Two key factors the computer can't account for will likely swing it. The first is the fact that Dwyane Wade can play a lot more minutes than he did in the regular season, which should give the Heat a big boost. The second, however, is how much success the Hawks have had defending Wade this season. In the three games before their meaningless final meeting Tuesday, Miami scored 73, 83 and 95 points against Atlanta.
I'm guessing the latter trend, combined with the Hawks' home-court advantage, will be the one that prevails, but not without some gnashing of teeth along the way.
My prediction: Hawks in 7
Boston (69.5%) vs. Chicago (30.5%): The absence of Garnett and the Bulls' strong play down the stretch combine to make the upset possible here, as Chicago has played much better since the midseason trade for John Salmons and Brad Miller. In fact, Chicago's 18-11 mark since the All-Star break isn't much different from Boston's 18-9.
I'm still taking Boston for a few reasons. First, Chicago's biggest advantage is its bench, and bench players become less important in the playoffs when the starters are hogging most of the minutes. Second, the Ben Gordon-Ray Allen matchup is a tough one for the Bulls defensively, and if they have to turn to Kirk Hinrich for extended burn to guard Allen, it will hurt them offensively.
Finally, if it comes down to a final game, Boston's passionate crowd can turn things in the Celtics' favor. But I'm not sure it will get to that point.
My prediction: Celtics in 6
Orlando (85.1%) vs. Philadelphia (14.9%): Other than Cleveland-Detroit, this is the series in which the computer sees the least possibility of an upset. Orlando won the regular-season series 3-0, and although Hedo Turkoglu and Rashard Lewis both enter the series banged up, the Magic still have presumptive defensive player of the year Dwight Howard and a brigade of 3-point shooters.
About all the 76ers can hang their hat on is that Turkoglu and Lewis might be subpar, or perhaps absent entirely, at the start of the series, and the fact that Orlando had an above-average turnover rate. Few teams are better at converting those miscues into transition buckets than the Sixers, and they'll need to do it repeatedly to avoid going up against Orlando's mighty D in the half court.
Perhaps the injuries will be enough to extend the series, but that's about as much optimism as I can summon for Philly.
My prediction: Magic in 5
Cleveland (91%) vs. Detroit (9%): Let's not waste everybody's time.
My prediction: Cavs in 4
Los Angeles (80%) vs. Utah (20%): The Power Rankings still see the Jazz as a decent team despite their late-season face-plant, which is why their odds look surprisingly strong here. But two huge factors mitigate against any hope of a Jazz upset, however.
First of all, as we saw a year ago, they have nobody who can guard Kobe Bryant, and their basic strategy of "foul first, ask questions later" should result in a parade of Lakers to the free-throw line.
Second, the Jazz don't appear physically or mentally to be in the same shape they were as recently as six weeks ago: Utah went only 7-11 since its 12-game winning streak was snapped by Atlanta on March 11.
The Jazz were able to beat L.A. once in Utah -- at the end of a long Lakers road trip, mind you -- and they'll probably be able to do it once again, but given their 1-20 mark in road games against teams with winning records, I see little hope of an upset.
My prediction: Lakers in 5
Denver (70%) vs. New Orleans (30%): Welcome to the mystery bracket. Three teams -- L.A., Cleveland and Orlando -- seem reasonably safe bets to make it to the conference finals, but who knows what's going to be happen in this quadrant?
The Hornets were 10-11 in their final 21 games and finished just 14th in the Power Rankings, so you can expect this system to be pessimistic about their chances.
On the other hand, the Hornets were 32-18 when Tyson Chandler played, and 17-19 when he didn't, and Chandler should be back -- albeit diminished -- for the opener. He's recovering from a sprained ankle and isn't at full strength yet, but the Hornets' frontcourt depth is so woeful that he should still make a big difference.
Of course, the Nuggets can make a similar case, having been 44-20 when Carmelo Anthony and Chauncey Billups both played, and those two are in far less dire straits health-wise than Chandler. A second pro-Denver wild card: Anthony shot much worse than his historic norms in the regular season and could bounce back in the playoffs.
Throw in the Nuggets' home-altitude advantage, and this could be over fairly quickly.
My prediction: Nuggets in 5
San Antonio (55%) vs. Dallas (45%): Here's another one the computer sees as a virtual toss-up.
The Spurs are in a compromised state right now, with Manu Ginobili out with an injury and Tim Duncan hobbled by bad knees, but they deserve a fair amount of respect based on track record alone. Additionally, Tony Parker is a matchup problem for the Mavs, with no Dallas guard having the speed to stay in front of him.
On the other hand, let me give you two salient sets of numbers: the Spurs' 22-16 record without Ginobili and Dallas's 33-19 record when Josh Howard plays. Since Howard will be present (though still troubled by a bad ankle) and Ginobili won't, one has to think Dallas might have the upper hand. Additionally, Dirk Nowitzki is nearly as bad a matchup for the Spurs as Parker is for Dallas.
Given how even they appear on paper, I expect a knock-down, drag-out series similar to their 2006 seven-game classic in intensity, if not in importance. And once again, I see the Mavs prevailing.
My prediction: Mavs in 6
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.
My prediction: Blazers in 6
Cleveland (83.6%) vs. Atlanta (16.4%): The Hawks match up pretty well against LeBron; at least, as well as you can reasonably hope to. So the guy who kills them might be Zydrunas Ilgauskas. Long big men give the Hawks fits because Al Horford and Josh Smith are too short to affect their shots.
And then there's the little matter of how they're going to score on the Cavs' vaunted defense. Joe Johnson will have to play James to something close to a draw for the Hawks to have any chance, but by this point Atlanta might have already won its Super Bowl by making it out of the first round.
The Hawks' raucous playoff crowds -- a far cry from the mausoleum conditions seen through most of the regular season -- might get them a game, much as it fueled them against Boston last year, but that's about it.
My prediction: Cavs in 5
Boston (35.3%) vs. Orlando (64.7%): This is the only series in which the computer prefers the lower-seeded team, as the Magic would likely be back at full strength by this point and the Celtics are undeniably weakened by the loss of Garnett (something I detailed in Thursday's column).
I've seen Orlando play Boston in person twice without Garnett (technically, Garnett played in the second game, but that wasn't the real KG), and while the Magic won both games, they hardly looked impressive in doing so. Even without Garnett, the Celtics are able to limit Dwight Howard's effectiveness thanks to the physical post D of Kendrick Perkins and their strong team concepts, and their struggling bench won't matter quite as much in postseason games if the starters play the bulk of the minutes.
All that said, I can't see the Celtics' beating the Magic four out of seven if Orlando is anywhere near full strength for this series. I expect the C's to struggle to score all series long, making their margin of error too small to beat Orlando consistently.
My prediction: Magic in 6
Denver (57.3%) vs. Dallas (42.7%): For those who are curious, the odds say the likelihood of winning two rounds to advance out of the mystery quadrant are 38.7 percent for Denver, 28.8 percent for San Antonio, 21.8 percent for Dallas and 10.8 percent for New Orleans.
I like Denver to emerge, thanks to its home court and the ability of Anthony to make big shots down the stretch. If you look at clutch shooting stats over the past few years, he's the unquestioned king, even better than LeBron and Kobe. While he can't touch those guys in the first 47 minutes, it's a nice crutch to have in the final seconds.
Denver can also put defensive ace Kenyon Martin on Nowitzki, while the Mavs don't have as strong a defender available for Anthony, especially with Howard's ankle remaining troublesome. I should also note that the Mavs lost all four times they faced the Nuggets in the regular season.
They'll fare a little better this time around, but they'll still come away with four losses.
My prediction: Nuggets in 7
Los Angeles (54.9%) vs. Portland (45.1%): Plenty of people, I'm sure, don't expect the Lakers to have any problem with the youthful Trail Blazers, if Portland can handle Houston.
So let me throw out these numbers (with the scoring margin adjusted for home and road games):
Team A: 22-8, +7.9 adjusted scoring margin
Team B: 23-7, +6.9 adjusted scoring margin
Team A, as you might have guessed, is Portland since the All-Star break, and Team B is L.A. Over the past two months the Blazers have played at least as well as the Lakers, including beating them twice by a combined 25 points. That comes with a 7-foot asterisk since Andrew Bynum missed much of that time with an injury, but it shows the margin between the teams isn't nearly as large as people might think.
Or try these: 20, 20, 21, 35, 3, 12, 8, 15, 30 and 28. Those were the Blazers' margins of victory while taking 10 of their final 11 games. They weren't just beating teams, folks, they were beating the holy hell out of them. And their road woes ceased late in the year; the Blazers won seven of their final nine road games, and one of the losses was at Cleveland in overtime.
I expect the Lakers to advance, thanks to their home-court advantage and the return of Bynum, but this is going to be a much, much tougher series than people think, and I think I might be the only person outside Portland who wouldn't be shocked if the Blazers ended up prevailing.
My prediction: Lakers in 7
Cavs (59.4%) vs. Magic (40.6%): The computer retains a healthy respect for the Magic despite their late-season struggles, and we shouldn't gloss over these facts: Orlando had two lopsided wins over Cleveland this season and its lone defeat against the Cavs was by four points. Additionally, if the Magic make it this far, we have to presume that Turkoglu and Lewis will be at or near full strength, so those injuries aren't likely to be a factor.
Chances of winning the East based on computer projections
|Team||Odds of winning title|
The bigger problem for Orlando is my perception of each team's upside. The Cavs can play LeBron nearly every minute of every game in the playoffs if they have to, which allows them to punch far above their regular-season weight. Meanwhile, I'm wondering if the Magic aren't regular-season overachievers who can't push the elevator to another floor in the postseason.
Perhaps they prove me wrong here; certainly I like how they match up against Cleveland well enough to think they'll win twice. But I can't see them pushing things to a seventh game, much less winning it.
My prediction: Cavs in 6
Western Conference Finals
Lakers (72.2%) vs. Nuggets (27.8%): Having personally witnessed the Lakers' four-game demolition of Denver a year ago, and noting that L.A. beat Denver three times out of four in the regular season this year, I think we can emphatically say that this is a terrible matchup for the Nuggets. They have nobody who can guard Kobe Bryant -- last year they resorted to the likes of Kenyon Martin and even Eduardo Najera -- and the length of Gasol and Bynum gives their front line fits.
Chances of winning the West based on computer projections
|Team||Odds of winning title|
In fact, the real conference finals matchup is going to be the second round, where either Portland or Houston has a much better shot at beating L.A. than the winner of the mystery bracket. And Denver's chances don't improve much if L.A. is out of the way. Even with home-court advantage, the Nuggets lose 57.0 percent of the pairings to Houston and 65.8 percent to Portland; the numbers for San Antonio (60.2 percent and 68.1 percent, respectively) are nearly identical.
In either of those matchups I'd expect the series to go at least six games. The same can't be said of Lakers-Nuggets: The altitude should get Denver one game, but that's as far as I'm willing to go.
My prediction: Lakers in 5
Cleveland (54.4%) vs. Los Angeles (45.6%): Call it David Stern's dream matchup. Kobe versus LeBron, a battle of the past two MVP winners (presuming LeBron wins this year), and one that matches the biggest stars in L.A. and New York (just kidding, Cavs fans). Let's just say there are going to be some deflated-looking people at the TV networks if we get Portland-Orlando instead.
Most common Finals matchups and results
|Pairing||% of sims||% won by East team|
The Lakers won both regular-season meetings, a positive indicator in their favor, but the Cavs will get the last two games on their home court, where they were 39-1 in games they tried to win. Yes, the Lakers were responsible for the "1", but that doesn't mean they're going to romp through the Q this time around.
Again, the Cavs' ability to play LeBron the whole game makes them a much more imposing team in the playoffs than they are in the regular season, as Boston found out a year ago, when a Cleveland team that gave up more points than it scored in the regular season took the mighty Celtics to the wire in a seventh game before succumbing.
This year, it's time to crown King James. I expect the Lakers to make this not just a competitive series, but a downright classic one. But in the end, the combination of the crowd at the Q and the best player in the world will be too much for the Lakers.
My prediction: Cavs in 7
John Hollinger writes for ESPN Insider. To e-mail him, click here.