MY NBA UPSET RANKINGS

WHICH TOP SEEDS MIGHT FALL IN ROUND 1?

Which first-round NBA playoff series are most likely to feature a win by the lower-seeded team?

Though the playoffs might be different from the regular season, the outcome of playoff series can be predicted in large part from three factors: the regular-season point differentials of the two teams, the results of the head-to-head series and who has home-court advantage.

Based on how well those criteria have predicted best-of-seven playoff series since the year 2000, we can assess the likelihood of the lower-seeded team emerging victorious. (We'll call that an upset, even though in one notable case this year the lower-seeded team is the market favorite. More on that in a second.) I've used those probabilities to rank the series from most to least likely to produce an upset along with an explanation for why each series ranks where it does.

1. PORTLAND TRAIL BLAZERS (3) VS. OKLAHOMA CITY THUNDER (6)

Upset probability: 52 percent

Although this model doesn't specifically consider the absence of Portland center Jusuf Nurkic, who suffered a season-ending compound leg fracture in late March, the Thunder are still favored over the Blazers on the basis of their 4-0 sweep of the season series. No wonder Portland, which played only six reserves in Wednesday's season finale against the Sacramento Kings, seemed willing to lose that game and slide down to the No. 4 seed in the Western Conference rather than play Oklahoma City. (As it turned out, the Blazers won anyway.)

The lower seed sweeping the head-to-head series is relatively rare, having occurred only three times in the same conference since 2000: the 2007 "We Believe" Golden State Warriors knocked off the top-seeded Dallas Mavericks in the first round after sweeping them in the regular season, but the other two lower-seeded teams (the Brooklyn Nets over the Miami Heat in 2014 and the Thunder over the San Antonio Spurs in the conference finals that same year) ended up losing.

More generally, the season series is a strong playoff predictor. Lower seeds in the same conference that won the head-to-head series have gone on to pull off a playoff upset 43 percent of the time in best-of-seven series since 2000, as compared to only 18 percent when they lost the season series and 26 percent when they split it 2-2.

Add in injuries to Nurkic and guard CJ McCollum (still working back to full strength from a knee injury suffered in mid-March) and Oklahoma City's massive edge in playoff experience (5,930 more career playoff minutes than Portland, the largest advantage for a lower seed in this year's first round) and it's no surprise that oddsmakers favor the Thunder despite the Blazers' home-court advantage. If Paul George misses time with the right shoulder soreness that forced him out of Wednesday's needed win in Milwaukee that could even things up.

2. HOUSTON ROCKETS (4) VS. UTAH JAZZ (5)

Upset probability: 39 percent

In terms of regular-season performance, this is the closest first-round series despite Houston's three-game lead in the standings. Utah actually had the West's second-best point differential (plus-5.3 ppg, with the Rockets third at plus-4.8) but an 0-7 record in games decided by three points or fewer cost the Jazz home-court advantage in the first round. (The next-worst record in games decided by three or fewer this season was by the New York Knicks, who went 4-7.)

The head-to-head series was close too, with the teams splitting four games and each winning one in blowout in fashion. Assuming Utah's point guards can stay healthy, everything about this matchup points to a tight series likely to go six or seven games, in which case home-court advantage could make the difference for Houston.

3. DENVER NUGGETS (2) VS. SAN ANTONIO SPURS (7)

Upset probability: 28 percent

While 2-7 matchups have produced only one upset since the first round went to best-of-seven in 2003 (the Spurs over Dallas in 2010), this is an unusually close pair of teams. Of the 34 series in that span, Denver's advantage in terms of point differential over San Antonio is seventh smallest, and it's one of 11 times the teams split the season series.

Playoff experience is also an edge for the Spurs, the second lower-seeded team with more playoff experience on their roster than their higher-seeded opponent. Lower seeds with more playoff experience have won 37 percent of best-of-seven series since 2000, as compared to 18 percent for lower seeds with less playoff experience.

These factors shouldn't be overstated; the Nuggets were the better team in the regular season, and crushed San Antonio by 28 points in the most recent head-to-head meeting last week. (The other three games were all decided by eight points or fewer.) So Denver should be favored, but the makings for an upset are here.

4. PHILADELPHIA 76ERS (3) VS. BROOKLYN NETS (6)

Upset probability: 26 percent

There's a Joel Embiid-sized asterisk on any projection of this series, as the Sixers' All-Star center told reporters Friday he doesn't know whether he'll be able to play in Saturday's Game 1 because of increasing pain in his left knee. (He's currently listed as "doubtful.")

Embiid played in all four games against the Nets, which the two teams split, though only the most recent meeting -- a 13-point Philadelphia home win -- came after the 76ers acquired Tobias Harris ahead of the trade deadline.

If the Sixers are healthy, they're clearly the better team compared to a young Brooklyn squad that had a negative point differential during the regular season. But given Philadelphia was outscored by 2.0 points per 100 possessions with Embiid on the bench after the All-Star break -- an 819-minute sample, per NBA Advanced Stats -- his absence could change everything.

5. BOSTON CELTICS (4) VS. INDIANA PACERS (5)

Upset probability: 24 percent

The Celtics and Pacers were separated by just a game in the standings, and the gap in their point differentials (plus-4.4 for Boston, plus-3.4 for Indiana) is third smallest of the eight series.

However, full-season stats include the 36 games played by Victor Oladipo, and the 3-1 season series win by the Celtics -- highlighted by wins home and away over the past two weeks of the regular season to secure home court in this series -- seem to offer a more accurate picture of how things might go. Since Oladipo suffered a season-ending quadriceps rupture, the Pacers' point differential is dead even. Use that instead of their season-long mark and their chances of an upset drop by half.

6. TORONTO RAPTORS (2) VS. ORLANDO MAGIC (7)

Upset probability: 19 percent

Though there's reason to believe the Magic's late-season improvement is real by virtue of replacing unproductive reserves Jerian Grant and Mo Bamba, be careful reading too much into Orlando's strong finish. As Tom Haberstroh of NBCSports.com and Zach Kram of the Ringer have reiterated this week, the late part of the season doesn't tend to be a good predictor of playoff performance.

The Magic splitting the season series was, however, perhaps even more impressive than it appears. The Raptors' two wins came by a combined 14 points, and Orlando won by 15 (in a game Kawhi Leonard missed) and 29 (without Kyle Lowry).

7. GOLDEN STATE WARRIORS (1) VS. LA CLIPPERS (8)

Upset probability: 8 percent

By virtue of beating the Warriors in overtime at Staples Center in November -- a game better remembered for Draymond Green's postgame tirade against teammate Kevin Durant -- the Clippers are given a slim chance of pulling off an upset. The Clippers also lost by only two at Oracle Arena in December, but the past two meetings were Golden State wins by a combined 55 points.

8. MILWAUKEE BUCKS (1) VS. DETROIT PISTONS (8)

Upset probability: <1 percent

The Bucks finished with the eighth-best point differential of the 2000s, per Basketball-Reference.com, and the Pistons were outscored during the regular season, making this the fifth-largest first-round mismatch in the best-of-seven era.

Milwaukee swept the season series, winning three of the four games by double-digits. And Detroit star Blake Griffin is day-to-day because of a knee injury that sidelined him for three of the season's final six games. So this series looks unlikely to be competitive.