2011 NBA Playoffs Predictions: Conference Semifinals

posted in NBA

The first round of the 2011 NBA Playoffs is over, and our survival odds have been updated. Check our 2011 NBA Playoff Odds page to see how much the Western Conference teams’ title chances have increased with the elimination of the Spurs, and to see which team still has an under 1% chance of winning the whole thing. (Hint: rhymes with Matlanta Rocks.)

As we mentioned in our round one predictions post, we’re unofficially entering Henry Abbot’s TrueHoop Stat Geek Smackdown 2011 via our picks here on the blog. Though we missed two first round series (Memphis over San Antonio, and Atlanta over Orlando), we’re still doing great — because literally every other competitor missed both of those series as well.

We correctly picked the length of two series (Miami in 5, and Chicago in 5), which puts us tied for second place, a mere 2 points behind first-place Benjamin Morris of Skeptical Sports. In fact, if Dallas hadn’t blown a twenty-three-point lead in game 4 against the Blazers, and as a result had closed out their series one game earlier, we’d be alone in first place.

Hopefully these conference semifinal picks will keep us near the top of the standings:

#1 Chicago vs. #5 Atlanta: Chicago in 5 (32% chance)
#8 Memphis vs. #4 Okla City: Oklahoma City in 7 (20% chance) … very close to OKC in 5 (19%)
#2 LA Lakers vs. #3 Dallas: Los Angeles in 7 (20% chance) … close to Dallas in 6 (17%) or LA in 5 (17%)
#2 Miami vs. #3 Boston: Miami in 5 (21% chance) … essentially a toss-up with Miami in 7 (20%)

As you can see, the choices in this round were much tougher than in the last. Our odds predict a 75% chance that at least one of these matchups will go 7 games, and there’s even a slim chance (7%) we may see three of them stretch to full length.

Boston versus Miami could be the most intriguing pairing, as our odds give Miami a 2 in 3 shot of winning, despite Boston taking 3 of 4 from Miami in the regular season. The predicted distribution of series length according to our power ratings shows almost an equal chance of Miami in 5 (21%) and Miami in 7 (20%). Despite Boston winning the season series, Miami outscored the Celtics overall and the Heat seemed to get better as the season wore on. Therefore, we’re going to downplay the possibility that there are specific matchup issues that our ratings may be missing, and stick with the straight numbers: Miami in 5.

  • http://twitter.com/skepticalsports Benjamin Morris

    I have Chicago in 5, and OKC, LA, and Miami in 6. Empirically, I’ve found 6 almost always trumps 7, no matter what the sims say. Good luck.

  • David (TeamRankings)


    Thanks for the comment. I had seen your recent post, in which you mentioned that 5 or 6 games is almost always the best choice. I had the exact same concern that you dealt with in yesterday’s update: that due to the Smackdown format, pegging the mode correctly is more important than minimizing overall error. But it looks like you have some (unpublished) evidence that picking 7 as the mode is still a bad choice. I’d love to see your reasoning, as it’s not an issue I personally have looked too deeply into.

    We did take a quick look at historical data this morning, just to confirm picking 7 isn’t TOTALLY crazy. I’ll be posting what we found as new blog entry shortly.

    Good luck, and nice job in round one!