Note: This assumes the league will not veto the trade, as the rumors indicate.
Cleveland’s management is hard to read.
In a world where information is leaked at a high rate, they shocked the world by drafting Anthony Bennett first overall. I wasn’t sure what they would do with Andrew Wiggins and Kevin Love, but the wait is finally over. We don’t need to discuss team strategy and philosophy about Wiggins and Love anymore — bank on an apparent star from the stats who’s been on a series of mediocre to lousy teams or bet on a kid ten years younger when you have LeBron James at the tail end of his prime?
Although the Cavaliers have preserved their core of Kyrie Irving, Dion Waiters, Tristan Thompson and Anderson Varejao, it’s a new team, entirely. It’s essentially their main players from last season, plus Love and LeBron and his friends with a new coach. Thus, it’s a fun exercise in predicting an NBA team’s season.
There are two core elements in how most NBA predictions are formed. Who will play and how much, and how valuable will each player be? The minutes distribution is an underrated facet and it can be particularly tricky. Derrick Rose is the obvious example, but for every player there are numerous considerations like off-season knee surgeries, age, injury history, height and weight, coach preference, position drifting, etc. Given some conservative assumptions — based on past season totals and an apparent rotation pecking order — however, you have a reasonable shot at prediction coupled with a good player rater. (There are also other tweaks available aside from minutes and player “goodness,” some of which I already incorporate and some of which I’m experimenting on.)
Using my prediction machine from last season, which beat a few professionals — and placed well even compared to Vegas — I’ve compiled an initial assessment of the team after Love’s trade.
This includes Mike Miller, Ray Allen, Shawn Marion, James Jones, Brendan Haywood, and a few extra minutes thrown to “replacement level” guys. Since Ray Allen has not been announced, I hedged my bet and gave him a mere 500 minutes. I’m expecting Marion to play with them, and if he doesn’t it’s a hit on their defense, but would improve their offense with more minutes to Miller.
2015 Cleveland Cavaliers: historic offense?
Wins: 58
Offensive efficiency relative to league average: +8.6 (historic)
Defensive efficiency relative to league average: +1.8 (below average)
In case you don’t know relative offensive efficiency well, last season the Clippers led the league with a +5.4 mark, and the greatest ever was +9.2 back in Dallas in 2004 when they had both Steve Nash and Dirk Nowitzki (via basketball-reference.) This is an offense at the same level as the Phoenix-Nash Suns, at their apex. If you think I’m crazy, Nate Silver predicts 65 wins for Cleveland next season, and it’s certainly not on defense, because they’re bringing in guys like Love and Miller (given the system he’s using, LeBron would not rate higher than Luol Deng on defense.)
The main reason Silver’s projection is awfully optimistic is his minutes distribution, that monkey wrench I mentioned earlier. Oddly, he lists minutes per games without totals or games played, but if we multiply that column by 82 we get a total of 19,680 minutes, which is the total a team would have for a season without overtime. Thus, it’s reasonable these can be translated to total minutes. He has Kyrie Irving at 2870 minutes, which is strange because that would land him near the league leaders in minutes played and he has yet to reach that total in his career — prorating the lockout season, he’s been near 2000 minutes twice and 2500 last season. He has Varejao at 2132 minutes; Varejao hasn’t done that since LeBron was last on the team and now he’s entering his mid-30’s.
If you do a simple weighted average of Nate Silver’s minutes and his individual net points metric, you do indeed get +9.6 for the team as a whole. However, substituting my own minutes (giving the extra minutes to more lowly rated players), you get something closer to 60 wins, which feels more reasonable as mid-60’s win teams are rare and Cleveland has glaring problems on defense. (For a technical note, my own system is more conservative and has a regression to the mean built in.) Yet even at 60 wins, based on the players involved, the metrics estimate their offense to be world-shattering. How can any decent projection system spit out something so extreme?
Prediction folly and diminishing returns
Firstly, LeBron’s offensive impact is astronomical. He’s a one-man offense, and is able to do three things at once, at a highly elite level, that almost no one in history has been able to replicate: shot volume, efficiency, and playmaking. And they’re adding one of the best offensive power forwards of the post-Jordan, three-point era. Given that the Cavaliers have lost no one of significant value on offense — Jarrett Jack is a middling shooter and his name is apt, Deng is lukewarm on offense, and other parts like Anthony Bennett and Andrew Bynum were near disastrous — we should expect one of the greatest turnarounds on offense ever.
Yet, will these parts fit together? This is where we ran into trouble with the 2011 Heat. John Hollinger, whose NBA predictions were consistently amazing as he routinely beat Vegas and other analysts, saw the worse-case scenario for the Heat as 61 wins — worse-case being they surrounded them only with poor players you’d see at the end of benches. Neil Paine, who works with Nate Silver now and is using similar methodology, gave conservative estimates by reducing each star’s rating and also came up with 61 wins. Miami was a disappointing their first year, but that’s not entirely accurate. Looking at their roster in 2011, Hollinger’s estimate of the Heat being surrounded with a 10 PER cast was a fairly astute guess. They won “only” 58 games, but their expected win total, based on their point differential, was 61 games.
This player value was translated decently well even though each player went from being the number one option to a smaller role. LeBron’s usage dropped by 3.8 points, Dwyane Wade’s by 2.4, and Chris Bosh’s by 5.2. But players usually have transferable skills. Bosh shot less often, but spotted up on the perimeter more and spaced the floor. Wade and LeBron learned to be deadly off-ball players, either through cutting or three-point spot-up shooting (for LeBron). A drop in usage typically translates to a greater efficiency, which was true to a small extent for Wade and to a large extent for James. The average player, for example, would have less “bail-out” shots because those could be shared by the other stars, and the pressured induced on the defense by all three would create easier shots elsewhere. This balance is hard to find, and hard to estimate, in the NBA, and it wasn’t a perfect fit in Miami because of Wade’s poor shooting. Elite defenses could shade off Wade and clog the lane whenever he didn’t have the ball.
A key mechanism here is the idea of diminishing returns. If you start five new players, who are all estimated at +5 because of their offense, you are not going to be a +25 team when those five players are on the court. Offensive players typically derive their value from having the ball in their hands, and with only one ball on the court, a star player who’s shooting less often will become less valuable. The whole is less than the sum of its parts, hence Team USA losing to less talented teams and having close calls in recent years. This was part of the problem with the Miami Heat, as Wade and James were two of the most ball dominant players ever. It took some additions like Miller, Chris Andersen, and Allen to reach the level that people expected out of the team.
Returning to another example, the 2008 Celtics were not assumed to be title favorites, as strange as that sounds now. Hollinger pegged them as a 51-win team, and Paine at 48 wins. No one at ESPN voted for Boston. The disparity is partly due to a new defensive coaching style championed by Tom Thibodeau, and better performances from young guys Rajon Rondo and Kendrick Perkins, but it was also the beautiful fit of the players. Kevin Garnett would take over the defense and play a supporting role on defense, spacing the floor and passing; Paul Pierce was the one-on-one scorer, could defend big wings, and could also space the floor; and Allen was their shooter and could work well off Garnett’s screens. In retrospect, the team made perfect sense. Diminishing returns had little effect on the team because the players had different roles and could focus their efforts on their own strengths.
Bringing this back to the Cavaliers, LeBron has just finished a four-year stint with a ball dominant guard in Wade and should feel no ill-effects here, as Irving is a better shooter (people like to cite his low SportVU catch-and-shoot 3PT percentage, but his Synergy spot-up stats were better in previous seasons; shooting percentages are volatile year-to-year). Love is fine spotting up more or crashing the boards, and he’s a wonderful passer — the Love-to-LeBron outlet passes will be awe-inspiring. Irving is a question mark, as he hasn’t made much progress since entering the league and hasn’t had a significant dent on Cleveland’s overall offense, but at least he’s a good shooter and very young. With defensive role players and shooters around them, you can maximize their talent pretty well.
With regard to diminishing returns and the Cavs, perhaps their efforts can partly be redistributed to defense. I would expect Cleveland to hit league average on defense as long as Varejao is healthy and for their offense to hit something mortal, like +6 on offense or less. (One note is that Miami’s offense was spectacular if you ignored offensive rebounding, but Cleveland has the rare gift of being able to kill teams on offensive rebounding with Love and Varejao and on shot efficiency, which is the same way Portland’s offense surprised everyone.)
The Irving and LeBron dynamic will be fascinating. Since LeBron played next to Wade and is coming from a near league-leading offense, I wouldn’t normally expect his usage rate to significantly fall, but I’m not sure how often Irving will be handling the ball. Based on last season’s stats, the team as a whole has a 2.5 higher average usage rate. That’s a large change, but it’s not crippling. I’d expect Waiters to take the largest hit — no way with Irving/LeBron/Love he’s going to be a 27 usage player again, unless they reduce his minutes and stick him on the bench.
I don’t think Love will be completely Bosh-ified, and people don’t quite understand how good of a scorer he is (he outscored Stephen Curry, and he did this next to a point guard who can’t shoot and a low-post center). Plus he has real three-point range, while Bosh relied on mid-range shots and worked hard to stretch his range. In any case, Love gets to run pick-and-rolls/pops with two of the best slashers and scorers. For reference, Dragic and Frye were the deadliest combo last year and a surprise team. Interaction is tricky to gauge in the NBA, but there should be positive effects on Cleveland given some of their skills. His defensive deficiencies are overstated — Minnesota’s defensive rating was nearly tied with Miami’s and that was in a tougher conference. Love’s oft-cited opponent field-goal percentage at the rim is misleading, because it doesn’t include free throws, which is a gross oversight given how common free throws are near the rim and the fact that Minnesota ranked first in free throws allowed per field goal.
League’s best offense and a title: mutually exclusive?
While the last team that won a title while leading the league in offensive efficiency was the 1997 Chicago Bulls, a team led by Michael Jordan that was still elite defensively, there have been several close calls since then. The Los Angeles Lakers have won titles while ranking second on offense multiple times, and Miami was a hair from leading the league in 2013, according to basketball-reference. LeBron has actually proven in Miami you can be a superior team on offense and win titles. All that matters is that you outscore your opponents. And flanked by a rising point guard with shooting skills, and a unique power forward who can crash the boards, pass, and stretch the floor, the Cavaliers are poised to become a special team on offense. To win a title in Cleveland, uniqueness is befitting.
