How The Warriors Are Wrecking The Rest Of The NBA — And Our Forecast Model


After two straight down years by the standards of their previous dynasty, the Golden State Warriors are making up for lost time early this season. Golden State has the best record in the league so far, with 18 wins in their first 20 contests, and they are tied for the second-best start in franchise history — trailing only the 20-0 record posted by the 2015-16 Warriors, who eventually ran their season-opening streak to 24 wins (and ultimately broke the all-time single-season NBA wins record).

The Warriors are off to a vintage start

Seasons with the most wins in the first 20 games for the Warriors franchise, 1946-2022

Per Game
Season Wins Losses Pts Scored Pts Allowed Diff.
2016 20 0 115.3 99.9 +15.4
2022 18 2 114.0 100.4 +13.7
2015 18 2 106.9 95.7 +11.2
2017 17 3 119.1 106.5 +12.7
2018 15 5 117.4 106.2 +11.2
1976 14 6 105.5 98.7 +6.8
1972 14 6 104.8 99.6 +5.2
1975 14 6 106.8 103.9 +2.9
2019 13 7 115.9 110.4 +5.5
1974 13 7 110.6 105.4 +5.2

Includes the Golden State, San Francisco and Philadelphia Warriors.

So everything old is new again. But how have the Warriors been able to bounce back from a 54-83 record over the past two seasons — missing the playoffs both years22 — to be so dominant from the opening tip in 2021-22? And why were statistical projection systems like our RAPTOR-based model so wrong about their chances headed into the season? Let’s break down a trio of factors that are driving Golden State’s renaissance.

Steph Curry is breaking basketball, again.

It’s no coincidence that the Warriors went 15-50 in 2018-19, when Stephen Curry missed all but five games with a broken hand, nor that they improved to finish above .500 with Curry suiting up for 63 of 72 games last season. Curry was outstanding a year ago, finishing third in MVP voting while ranking 10th in our RAPTOR wins above replacement metric. But the improvement-obsessed Curry has raised his game even more this year. He’s currently making 5.5 3-pointers per game, which would break his own all-time record — and give him a fourth season of five or more threes per contest, something nobody else in NBA history has even done once. With an offensive RAPTOR rating of +9.8 per 100 possessions, Curry is also vying for just the second individual season of a double-digit offensive RAPTOR in NBA history,23 which would join (you guessed it) his own 2015-16 campaign in that exclusive club.

Curry’s improvement has extended to defense as well, and that might be the more unexpected development. While RAPTOR has at times been higher on Curry’s defense than his conventional reputation at that end of the court (see his +2.5 mark across the 2014-15 and 2015-16 seasons), it has generally tended to rate Curry as a poor defender in recent years: From 2016-17 through 2020-21, Curry had a below-average -0.7 defensive RAPTOR, which ranked 256th out of the 386 players who logged at least 2,500 minutes over that span. But this season, Curry’s defensive value is back up to +2.6 per 100 possessions, 35th-highest of any player in the league and the second-best single-season mark of his career (again, trailing only 2015-16). Not only are the Warriors playing significantly better defense with Curry on the court than off, but Curry’s individual numbers are strong (as indicated by his +1.3 “box score” defensive RAPTOR), including good shot defense and one of the lowest rates of fouling for any player with a steal rate of at least 2.5 percent.

Taken together, Curry’s two-way impact has made him worth over 12 points per 100 possessions more than an average player so far this season — a leap of nearly 6 points per 100 from even Curry’s MVP-caliber 2020-21 performance. 

A strong defense has gotten even better.

It’s not just Curry who’s been driving the Warriors’ early success. His teammates have also made more than their share of contributions — and that starts on defense. This is nothing new: According to Basketball-Reference.com, the Warriors owned the fifth-best D in the league last season, allowing 2.2 fewer points per 100 than the average team. But Golden State has clamped down even harder on opposing offenses so far this year, leading the league in defensive efficiency with a mark 8.1 points better than average per 100 (which is particularly impressive considering how much better defensively the league as a whole has been in the early part of this season).

Some of the usual suspects are anchoring Golden State’s defensive effort. Draymond Green has held opponents to a field goal percentage 4 points lower than their usual averages, one of the best rates in the league, and though Andre Iguodala has been battling injuries, he has an outstanding +7.1 defensive RAPTOR in his return to the Bay Area after two seasons away. Warrior mainstay Kevon Looney continues to be one of the most underappreciated defensive big men in the game, putting up a better-than-average RAPTOR at that end for the sixth consecutive season. But Golden State’s newcomers are also playing a huge role. Bolstered by Otto Porter Jr. (+2.9 defensive RAPTOR) and Nemanja Bjelica (+3.1) in the frontcourt, the Warriors’ newcomers24 have combined for a collective +3.2 defensive RAPTOR, which ranks No. 1 among all teams’ new additions this season. (Meanwhile, our preseason forecast had projected both Porter and Bjelica to be below-average defenders.) And even depth players who aren’t new to the team but have taken on expanded roles, such as guard Gary Payton II — who has the best defensive RAPTOR (+9.0) of any player with at least 250 minutes this season — have excelled at the defensive end.

This ensemble cast of defenders recently hampered the great ex-Warrior Kevin Durant, an example of the strong D that can be the backbone of what this team accomplishes by season’s end.

An illustration of a slot machine depicting NBA trades.

Related: Build An NBA Contender With Our Roster-Shuffling Machine Read more. »

This offense has improved a lot.

For a team led by Curry playing at an elite level, it was surprising that the 2020-21 Warriors weren’t better on offense. Though Golden State ranked ninth in 3-point percentage, seventh in effective field-goal percentage and seventh in true shooting percentage, the team was only 19th in offensive rating with an efficiency mark 1.2 points per 100 possessions below league average. How was that possible? Simply put, outside of shooting, they were below-average (if not among the NBA’s very worst) in each of the other Four Factors: turnovers, rebounding and drawing fouls.

Some of that hasn’t changed. The Warriors remain among the most turnover-prone teams in the league. But they are now above-average at getting to the line, and their offensive rebounding is among the most improved of any team, with Looney ranking second in the league in rebound percentage at that end of the court. (Plus Andrew Wiggins may be leading the league in magical rebounding levitations.) Those two changes have helped the Warriors do a much better job of maximizing the value of every possession — which has had a multiplicative effect because their shooting has also been so extremely good.

To illustrate that point, Golden State is leading the NBA in effective field goal percentage, true shooting percentage and threes made per 100 possessions, a deadly combination that harks back to the Durant-era Warriors offenses that were among the most unstoppable ever. Again, some of that owes to Curry returning to his peak 2016-era form: When your best player is scoring 28.6 points per game with a .645 true shooting percentage, your offense is almost certainly going to be dominant. But Curry’s teammates are also doing more this year, shooting for an effective field-goal percentage (55.6 percent) more than 2 points higher than they did last year (53.5), including a scorching 63.5 percent off of direct passes from Curry himself, according to Second Spectrum data.25 And in part due to the additions of Porter and Bjelica, who can both stretch the floor with 3-point shooting on offense, Golden State now ranks among the top five in share of shots taken at the rim and from three, and the bottom five in share of shots taken from midrange, according to data from Cleaning the Glass. That formula has added up to a lot of offensive success for the Warriors in the early going.


So why didn’t our stats see this coming? It’s a good lesson in the whole becoming more than the sum of its parts, as well as the power of a singular NBA superstar like Curry. In the latter case, Curry outplaying projections by 7.0 points per 100 possessions of RAPTOR has had a massive effect on his team’s bottom line; we think he’s already been worth 5.1 more wins than a replacement-level player in just the first 20 Warriors games of the season. At age 33, most of Curry’s most similar players were on a declining trajectory, but there’s a reason why he is a one-of-a-kind superstar — Curry just keeps defying expectations.

How the Warriors have exceeded expectations

Preseason predicted versus actual overall RAPTOR ratings for the 2021-22 Golden State Warriors

Forecast Actual Diff.
Player Poss. Off. Def. Off. Def. Off. Def.
Stephen Curry 1,399 +6.6 -1.1 +9.8 +2.6 +3.2 +3.7
Andrew Wiggins 1,280 -0.5 +0.1 +0.1 -0.1 +0.6 -0.2
Jordan Poole 1,279 +0.2 -1.9 +0.7 -2.5 +0.5 -0.6
Draymond Green 1,220 -0.8 +2.9 -1.1 +1.7 -0.3 -1.2
Kevon Looney 785 -2.5 +1.5 -1.0 +2.0 +1.5 +0.5
Otto Porter 686 +0.5 -0.5 +2.3 +2.9 +1.8 +3.4
Damion Lee 651 -0.2 +0.7 -1.8 +0.8 -1.6 +0.1
Nemanja Bjelica 608 -0.7 -0.6 -0.2 +3.1 +0.5 +3.7
Juan Toscano-Anderson 604 -1.8 +1.3 -0.5 -0.3 +1.3 -1.6
Gary Payton II 575 -0.2 +0.7 +3.4 +9.0 +3.6 +8.3
Andre Iguodala 525 -1.8 -0.1 +0.4 +7.1 +2.2 +7.2
Jonathan Kuminga 222 -2.3 -1.0 -4.2 +2.1 -1.9 +3.1
Chris Chiozza 197 +1.4 -0.9 -4.4 +0.6 -5.8 +1.5
Moses Moody 194 -1.9 -1.5 -7.0 -2.2 -5.1 -0.7
Jeff Dowtin 13 -1.7 -0.8 -9.2 -5.3 -7.5 -4.5
Team total +1.1 +0.4 +4.5 +6.8 +3.3 +6.3

Team ratings are based on a possession-weighted composite of the individual ratings.

Source: NBA Advanced Stats

And fit-wise, his fellow Warriors — particularly the team’s new acquisitions — have found roles that accentuate their strengths. Among Golden State regulars, only Green, Damion Lee and Juan Toscano-Anderson have undershot their preseason RAPTOR forecasts (and those misses have been far from bad). While, say, Payton probably won’t continue to have a +12.4 RAPTOR going forward, NBA stats take a relatively short period of time to stabilize, so there isn’t a good reason to think these Warriors won’t be a team to be reckoned with all season long.

In fact, the impending return of Klay Thompson, who hasn’t played a minute in the NBA since Game 6 of the 2019 Finals, will add another wrinkle to the 2021-22 Warriors’ story. Our projection model, which heavily penalizes aging players with a similar history of major injuries, doesn’t think highly of Thompson at all, which means Golden State’s team rating might actually go down when he rejoins the rotation. But given how the Warriors have defied their projections so far — and given how Thompson might add even more shooting to that dimension of their offense, even if he is not the same player he was before the injuries — it would not be the least bit surprising to see his return teach the model yet another lesson in a Warriors season that has been full of them.

Check out our latest NBA predictions.

Watch: https://abcnews.go.com/fivethirtyeight/video/makes-lebron-james-great-fivethirtyeight-74851925