Editor's note: Welcome to NBA Insider Daily! Devoted readers familiar with the "Per Diem" column can still look forward to both Kevin Pelton and Tom Haberstroh this season, and they are joined by fellow Insiders Bradford Doolittle and Amin Elhassan. Every day, one of them will tackle the big topics in the NBA from every perspective including advanced metrics, scouting and reporting.
On the eve of the 2014-15 regular season, it's our last chance to make projections for the upcoming campaign. Having dug into the results of my SCHOENE projection system, here are five surprising outcomes I expect to come to fruition.
1. The Cleveland Cavaliers will set the NBA record for offensive rating
As expected, the Cavaliers have demonstrated a unique ability to dominate the offensive glass while also stretching the floor. They rebounded 32.4 percent of their misses during the preseason; no other NBA team was above 30 percent.
2. The Detroit Pistons will have an above-average defense
The Pistons finished 26th a year ago. Surprisingly, SCHOENE projects an improvement all the way to 12th. In general, teams that change coaches regress to the mean defensively, and new head coach Stan Van Gundy led the Orlando Magic to five top-10 finishes in defensive rating over six seasons.
There also was evidence of a better defense in Detroit last season. When the Pistons played any two of their three frontcourt starters (Andre Drummond, Greg Monroe and Josh Smith) together, they allowed 108.9 points per 100 possesssions, per NBA.com/Stats. That defensive rating jumped to 115.1 when all three played together -- a mark that would have been worst in the NBA.
With Van Gundy limiting the minutes for the Drummond/Monroe/Smith frontline, Detroit ranked an impressive eighth in defensive rating during the preseason.
3. The Los Angeles Lakers will have the league's worst defensive rating
At this point, projecting the Lakers last might not qualify as bold. They finished 28th in defensive rating during the preseason, but the two teams behind them (the Los Angeles Clippers and the Oklahoma City Thunder) were contenders saving an extra gear for the regular season. Ditto the San Antonio Spurs, who finished 27th.
The Philadelphia 76ers, the one team SCHOENE projected to defend almost as badly as the Lakers, ranked 12th during the preseason. The Lakers look porous inside and out, and new head coach Byron Scott never finished better than 26th in defensive rating during his three seasons as head coach in Cleveland.
4. James Harden will lead the NBA in points
Harden
SCHOENE projects Kevin Durant to lead the league in points per game, but Durant's injury opened the door for Harden to outscore him -- and everyone else -- in terms of total points. Though Harden finished fifth in points per game last season (25.4), Durant isn't the only player ahead of him in that ranking likely to score less this season. Two of the other three (LeBron James and Kevin Love) teamed up in Cleveland, meaning fewer shots to go around. And the New York Knicks' move to the triangle could cut into Carmelo Anthony's scoring average.
During the preseason, Harden averaged 26.3 points per 36 minutes, good for seventh among players who saw at least 100 minutes of action. He and Anthony (26.4 points per 36 minutes) scored at nearly equal rates.
5. Stephen Curry and Klay Thompson will make more 3-pointers than the Lakers as a team
Curry
Thompson
OK, this is the only one SCHOENE doesn't specifically project. In fact, the Lakers are projected to make more 3s than five other teams. However, the projection system can't account for Scott's desire to avoid outside shots, a 180-degree turn from predecessor Mike D'Antoni.
The Lakers' 3-point numbers went up slightly after they were held without a triple in consecutive preseason games, attempting just eight total in that span. And the return of injured wing Nick Young to the lineup will help, because Young is the only Laker projected to make more than 100 3-pointers. Still, the Lakers attempted 3s on just 12.4 percent of their shot attempts in the preseason, a rate far lower than the 29th-ranked team (Minnesota at 17.4 percent). And though Curry and Thompson played fewer minutes during the preseason than they will in the regular campaign, they combined for 40 triples to 25 for the Lakers as a team.
News and notes
• The preseason leaderboard for wins above replacement player (WARP) by my system:
Preseason WARP Leaders
• The two rookie leaders in WARP were both members of the Miami Heat: 2013 second-round pick James Ennis and this year's first-round pick Shabazz Napier, both with 0.6 WARP. K.J. McDaniels of the Philadelphia 76ers and P.J. Hairston of the Charlotte Hornets also both had 0.6 WARP.
• Leaders in individual advanced stats (all minimum 100 minutes): true shooting percentage: John Jenkins, Atlanta Hawks (.826); usage rate: Kobe Bryant, L.A. Lakers (38.2 percent); rebound percentage: Reggie Evans, Sacramento Kings (28.3 percent); assist percentage: Chris Paul, L.A. Clippers (13.4 percent); steal percentage: Nikola Mirotic, Chicago Bulls (4.6 percent); block percentage: Samuel Dalembert, New York Knicks (10.0 percent).
• So just what does the preseason mean? I've gotten that question a few times in chats and on Twitter, and figured it was time to update aging studies on the issue. For players, I looked at last year's stats. Preseason performance explains about 28 percent of the variation in regular-season performance, as measured by the per-minute Win% stat (akin to PER). By comparison, SCHOENE projections explain about 62 percent of the variation in regular-season performance among players.
Taking it a step further, I used both SCHOENE and preseason ratings to project performance during the regular season. A regression suggests preseason performance should be about 13 percent of the projection. So it does matter, but not that much. The most any player's projection improves by incorporating preseason play (Mo Williams of the Minnesota Timberwolves had the biggest jump) is a little more than 10 percent.
• At the team level, over the past five full preseasons (not counting the post-lockout season), win-loss record explains about 12 percent of the variation in teams' regular-season records. Preseason point differential is slightly less predictive.
Several years ago, Roland Beech of 82games.com -- now with the Dallas Mavericks -- found that preseason records were less meaningful for teams coming off seasons with 50-plus wins than those that had won 30 games or fewer. For whatever reason, that trend didn't hold over the past five preseasons. There's been almost no correlation between preseason and regular-season record for teams coming off 30-win seasons or worse.
Strong preseasons should still improve our outlook for teams such as the Detroit Pistons and New Orleans Pelicans (5-2), as well as the Boston Celtics and Utah Jazz (5-3), but we're talking about upping expectations by a few wins rather than a major leap. And poor preseasons for the Clippers (2-6), Spurs (2-5) and Thunder (2-6) shouldn't be a concern.