Skip to main content
Next Season’s NBA Heavyweights: Warriors, Cavs, Spurs, Rockets … Timberwolves?

UPDATE (June 30, 5:38 p.m.): Just as we were publishing this story, it was reported that Minnesota Timberwolves’ point guard Ricky Rubio will be traded to the Utah Jazz for a first-round draft pick. The story has been updated to reflect the trade.

It’s a dangerous time of year to be an NBA fan. With free agency officially getting underway on Saturday, and players such as Paul George available via the trade market, you can talk yourself into any number of far-fetched scenarios wherein your favorite team puts just the right pieces together and suddenly becomes a contender. (What if the Spurs added Blake Griffin? What if the Celtics brought in both George and Gordon Hayward?) Sometimes dreams really do come true — like when the Rockets landed Chris Paul this week — but most of the time, you’ll wind up disappointed instead.

At FiveThirtyEight, we sometimes play this dangerous game with spreadsheets — specifically, with a spreadsheet that projects team records based on our CARMELO player projections. And there’s one team that really caught our spreadsheet’s eye: the Minnesota Timberwolves. The Wolves already made their big move of the summer, acquiring the Bulls’ Jimmy Butler for Zach LaVine, Kris Dunn and an exchange of first-round draft picks. When we plugged the Wolves’ CARMELO projections into the spreadsheet,1 it came up with a projected record of 50-33.

The Timberwolves look like contenders

CARMELO projections for the 2017-18 Minnesota Timberwolves

Jimmy Butler 33 +3.8 +1.0
Andrew Wiggins 32 +1.5 -1.9
Gorgui Dieng 30 -1.0 +2.6
Karl-Anthony Towns 37 +3.7 +0.3
Shabazz Muhammad 16 -0.1 -3.1
Nemanja Bjelica 16 -0.7 +0.7
Tyus Jones 15 -0.3 -0.8
Cole Aldrich 10 -2.2 +2.6
Justin Patton 8 -2.6 +0.4
Repacement-level players 43 -1.7 -0.3
Team total 240 +5.4 +1.1
Timberwolves’ projected record 49.5 32.5

But that doesn’t account for the significant cap space cleared by the Rubio deal. If Minnesota added free agent point Jeff Teague, for example, their projected record would improve to 53-29. If they signed Kyle Lowry instead, they’d project to finish at 58-24. They could also use the extra cap room to sign a frontcourt player.

Projecting the Timberwolves to win 50-something games seems awfully daring, especially for a team that’s burned CARMELO in the past. (CARMELO boldly projected the Wolves to win 46 games last season. Instead, they won 31.) But let me walk you through what the system is “thinking.” The projection reflects a combination of three factors: Butler, the Timberwolves’ youth, and their bad luck last season.

Jimmy Butler is really good, and he’s replacing players who were really bad

CARMELO expects Butler to be worth about 10 wins next season, as compared to a replacement-level player. Oftentimes, replacement level is too low a bar when it comes to assessing an NBA acquisition. If the Celtics added players such as George and Hayward, their minutes would partly come at the expense of other pretty good players such as Avery Bradley and Jae Crowder.2 Thus, their net gain might not be as large as you’d think.

But the players the Wolves gave up for Butler weren’t making positive contributions at all, at least according to advanced statistics such as Real Plus-Minus and Box Plus/Minus. (CARMELO uses a combination of these stats to make its projections, weighting RPM more heavily.) LaVine is a good athlete who can create shots but who was woefully inadequate on defense; thus, he was no better than replacement level last season, these metrics figure. And Dunn, like many rookies, was overmatched, playing at a below-replacement-level clip. Thus, Butler is a true 10- or 11-win upgrade, relative to the players Minnesota gave up for him.

We should note, however, that where Butler falls on the spectrum between “really good” and “superstar” is a matter of some debate. According to RPM, Butler was the seventh-best player in the NBA last season on a per-possession basis and the third most valuable by wins added above replacement level when also considering his playing time. By a more subjective measure — the views of sportswriters voting for the All-NBA teams — he was somewhere between the 11th- and the 15th-best player in the league, by contrast.

Karl-Anthony Towns and Andrew Wiggins should continue to improve

The Wolves’ two former No. 1 overall picks are young — Karl-Anthony Towns turns 22 in November, while Andrew Wiggins will turn 23 in February — and both still have plenty of room to grow, especially on defense. Towns already has a well-rounded offensive game, having developed into a dangerous outside shooter last year (37 percent from 3-point range). But the advanced metrics are somewhat split on his defense, with RPM viewing it as below-average — unusual for a 7-footer3 — while stats based on opponents’ field goal percentages suggest that he does a respectable job of rim protection. Towns’s defense tended to fall apart in the fourth quarter last season, and overwork could have been an issue — he was second in the NBA in minutes played, behind Wiggins.

Wiggins’s indifferent defense has been a subject of frequent critique at FiveThirtyEight. But the advanced metrics are uniformly in agreement that it’s poor. He allowed an effective field goal percentage of 56 percent last season on shots where he was the nearest defender.4 NBA shooters also have an effective field goal percentage of 56 percent on uncontested shots, so it’s as though he wasn’t playing defense at all. Because Wiggins is a good athlete with a long wingspan — factors that usually predict good defense — the problems mostly boil down to technique and effort, and those things can sometimes be improved.

The Timberwolves were unlucky

Minnesota was outscored by only 1.2 points per game last season, and yet they went 31-51. If that seems like a mismatch, it is. A team with that point differential would typically expect to go about 38-44, according to the Pythagorean record as calculated at Thus, the Wolves underperformed by seven wins last year, relative to their number of points scored and allowed. That’s because they didn’t play well in crunch time and went 10-18 in games decided by 6 points or fewer.

It’s easy to come up with hypotheses for why they played so poorly in these situations. Towns and Wiggins played too many minutes; Wiggins and LaVine took poor shots; Rubio isn’t a scorer, which limited their options in the half-court; they were bad on defense overall, and those differences are magnified in crunch time.

The fact is, however, that teams who underperform their Pythagorean records by as much as the Wolves did last season usually don’t have the same problem the next time around, or at least not to the same extent. There had been 19 previous cases since the NBA-ABA merger where a team underperformed its Pythagorean record by seven or more wins. On average, they fell only one win short of their Pythagorean record in the following season. There’s certainly some skill in which teams fare best in crunch time — and Butler, who’s both a good defender and a versatile scorer, can help the Wolves with that — but losing so many games in the clutch is usually partly a matter of bad luck.

Teams like the Timberwolves usually improved their luck

Difference between actual and Pythagorean wins for teams that underperformed their Pythagorean record by 7 or more wins, 1976-2017

2013-14 Timberwolves -8 -3
2011-12 76ers -8 +3
2010-11 Timberwolves -7 -2
2007-08 Raptors -8 0
2006-07 Celtics -7 -1
2002-03 Nets -7 -2
1999-2000 Nets -7 0
1997-98 Pistons -9 -3
1996-97 Celtics -7 +3
1994-95 Bulls -7 +2
1994-95 Trail Blazers -8 -4
1992-93 Kings -8 +2
1991-92 Timberwolves -8 -2
1989-90 Timberwolves -7 -1
1985-86 SuperSonics -10 -3
1984-85 Trail Blazers -7 -4
1982-83 Pacers -7 -2
1978-79 Bucks -9 -2
1976-77 Suns -9 -2
Average -8 -1


What could go wrong — or very, very right

In addition to all the bad things that could happen to the Wolves from a basketball standpoint — injuries, poor chemistry, etc. — they’re also a challenging team to forecast. For the past two seasons, the Wolves have unquestionably had a lot of talent on their roster but have also unquestionably been bad. It isn’t quite as clear why this disconnect occurred. Towns, Wiggins, Rubio and LaVine are all somewhat unusual players, and they each engender disagreements both between the various statistical systems and between stats and “eye test” evaluations. The way RPM and CARMELO looked at the Wolves, Wiggins and especially LaVine were part of the problem last season, while Towns and Rubio were part of the solution. If that assessment was wrong, then jettisoning LaVine could be more costly than the system assumes. And as I mentioned, RPM and CARMELO view Butler as a borderline-superstar player and not “merely” an All-Star; that’s another source of uncertainty.

On the flip side, the Timberwolves do have some additional cap space and an opportunity to round out their roster via players such as Taj Gibson, J.J. Redick or Danilo Gallinari. Even modest improvements could go a long way because they don’t have a deep rotation as currently constructed.

Or the Wolves could go really bold and package Wiggins for another star. Before landing Butler, the Timberwolves were reportedly in the market for George, for example. But a straight-up trade of Wiggins for George would work under the NBA’s salary cap rules given the Wolves’ extra cap space. It would be a hugely risky move — George will be a free agent next summer and has said he wants to play for the Lakers — but a core of George, Butler and Towns could make the Timberwolves legitimate title contenders. Or at least, the spreadsheet says so.


  1. Assuming the Wolves re-sign restricted free agent Shabazz Muhammad but make no other changes.

  2. And they also might have to sacrifice players such as Bradley and Crowder as part of trades, or to clear cap room.

  3. RPM almost always rates players that tall as net-positive defenders.

  4. And a maximum of 6 feet from the shooter; we consider shots where no defender was within 6 feet to have been uncontested.

Nate Silver is the founder and editor in chief of FiveThirtyEight.